0.8.0

* Removing some unsafe

* Removing some unsafe (2)

* Remove murmurhash

CHANGELOG.md

@@ -1,3 +1,15 @@
+Tantivy 0.8.0
+=====================
+*No change in the index format*
+- API Breaking change in the collector API. (@jwolfe, @fulmicoton)
+- Multithreaded search (@jwolfe, @fulmicoton) 
+
+
+Tantivy 0.7.1
+=====================
+*No change in the index format*
+- Bugfix: NGramTokenizer panics on non ascii chars
+- Added a space usage API
 
 Tantivy 0.7
 =====================

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "tantivy"
-version = "0.7.0"
+version = "0.8.0"
 authors = ["Paul Masurel <paul.masurel@gmail.com>"]
 license = "MIT"
 categories = ["database-implementations", "data-structures"]
@@ -12,12 +12,12 @@ readme = "README.md"
 keywords = ["search", "information", "retrieval"]
 
 [dependencies]
-base64 = "0.9.1"
+base64 = "0.10.0"
 byteorder = "1.0"
 lazy_static = "1"
 regex = "1.0"
 fst = {version="0.3", default-features=false}
-fst-regex = { version="0.2", optional=true}
+fst-regex = { version="0.2" }
 lz4 = {version="1.20", optional=true}
 snap = {version="0.2"}
 atomicwrites = {version="0.2.2", optional=true}
@@ -29,12 +29,11 @@ serde = "1.0"
 serde_derive = "1.0"
 serde_json = "1.0"
 num_cpus = "1.2"
-itertools = "0.7"
+itertools = "0.8"
 levenshtein_automata = {version="0.1", features=["fst_automaton"]}
 bit-set = "0.5"
 uuid = { version = "0.7", features = ["v4", "serde"] }
-crossbeam = "0.4"
-crossbeam-channel = "0.2"
+crossbeam = "0.5"
 futures = "0.1"
 futures-cpupool = "0.1"
 owning_ref = "0.4"
@@ -49,12 +48,14 @@ owned-read = "0.4"
 failure = "0.1"
 htmlescape = "0.3.1"
 fail = "0.2"
+scoped-pool = "1.0"
+murmurhash32 = "0.2"
 
 [target.'cfg(windows)'.dependencies]
 winapi = "0.2"
 
 [dev-dependencies]
-rand = "0.5"
+rand = "0.6"
 maplit = "1"
 
 [profile.release]
@@ -68,11 +69,11 @@ overflow-checks = true
 
 [features]
 # by default no-fail is disabled. We manually enable it when running test.
-default = ["mmap", "no_fail", "regex_query"]
+default = ["mmap", "no_fail"]
 mmap = ["fst/mmap", "atomicwrites"]
-regex_query = ["fst-regex"]
 lz4-compression = ["lz4"]
 no_fail = ["fail/no_fail"]
+unstable = [] # useful for benches.
 
 [badges]
 travis-ci = { repository = "tantivy-search/tantivy" }

README.md

@@ -21,7 +21,7 @@
 
 **Tantivy** is a **full text search engine library** written in rust.
 
-It is closer to Lucene than to Elastic Search and Solr in the sense it is not
+It is closer to [Apache Lucene](https://lucene.apache.org/) than to [Elastic Search](https://www.elastic.co/products/elasticsearch) and [Apache Solr](https://lucene.apache.org/solr/) in the sense it is not
 an off-the-shelf search engine server, but rather a crate that can be used
 to build such a search engine.
 

examples/basic_search.rs

@@ -16,10 +16,11 @@ extern crate tempdir;
 // Importing tantivy...
 #[macro_use]
 extern crate tantivy;
-use tantivy::collector::TopCollector;
+use tantivy::collector::TopDocs;
 use tantivy::query::QueryParser;
 use tantivy::schema::*;
 use tantivy::Index;
+use tempdir::TempDir;
 
 fn main() -> tantivy::Result<()> {
     // Let's create a temporary directory for the
@@ -34,7 +35,7 @@ fn main() -> tantivy::Result<()> {
     // be indexed".
 
     // first we need to define a schema ...
-    let mut schema_builder = SchemaBuilder::default();
+    let mut schema_builder = Schema::builder();
 
     // Our first field is title.
     // We want full-text search for it, and we also want
@@ -105,37 +106,37 @@ fn main() -> tantivy::Result<()> {
     // For convenience, tantivy also comes with a macro to
     // reduce the boilerplate above.
     index_writer.add_document(doc!(
-        title => "Of Mice and Men",
-        body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
-                bank and runs deep and green. The water is warm too, for it has slipped twinkling \
-                over the yellow sands in the sunlight before reaching the narrow pool. On one \
-                side of the river the golden foothill slopes curve up to the strong and rocky \
-                Gabilan Mountains, but on the valley side the water is lined with trees—willows \
-                fresh and green with every spring, carrying in their lower leaf junctures the \
-                debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
-                limbs and branches that arch over the pool"
+    title => "Of Mice and Men",
+    body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
+            bank and runs deep and green. The water is warm too, for it has slipped twinkling \
+            over the yellow sands in the sunlight before reaching the narrow pool. On one \
+            side of the river the golden foothill slopes curve up to the strong and rocky \
+            Gabilan Mountains, but on the valley side the water is lined with trees—willows \
+            fresh and green with every spring, carrying in their lower leaf junctures the \
+            debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
+            limbs and branches that arch over the pool"
     ));
 
     index_writer.add_document(doc!(
-        title => "Of Mice and Men",
-        body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
-                bank and runs deep and green. The water is warm too, for it has slipped twinkling \
-                over the yellow sands in the sunlight before reaching the narrow pool. On one \
-                side of the river the golden foothill slopes curve up to the strong and rocky \
-                Gabilan Mountains, but on the valley side the water is lined with trees—willows \
-                fresh and green with every spring, carrying in their lower leaf junctures the \
-                debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
-                limbs and branches that arch over the pool"
+    title => "Of Mice and Men",
+    body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
+            bank and runs deep and green. The water is warm too, for it has slipped twinkling \
+            over the yellow sands in the sunlight before reaching the narrow pool. On one \
+            side of the river the golden foothill slopes curve up to the strong and rocky \
+            Gabilan Mountains, but on the valley side the water is lined with trees—willows \
+            fresh and green with every spring, carrying in their lower leaf junctures the \
+            debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
+            limbs and branches that arch over the pool"
     ));
 
     // Multivalued field just need to be repeated.
     index_writer.add_document(doc!(
-       title => "Frankenstein",
-       title => "The Modern Prometheus",
-       body => "You will rejoice to hear that no disaster has accompanied the commencement of an \
-                enterprise which you have regarded with such evil forebodings.  I arrived here \
-                yesterday, and my first task is to assure my dear sister of my welfare and \
-                increasing confidence in the success of my undertaking."
+    title => "Frankenstein",
+    title => "The Modern Prometheus",
+    body => "You will rejoice to hear that no disaster has accompanied the commencement of an \
+             enterprise which you have regarded with such evil forebodings.  I arrived here \
+             yesterday, and my first task is to assure my dear sister of my welfare and \
+             increasing confidence in the success of my undertaking."
     ));
 
     // This is an example, so we will only index 3 documents
@@ -212,15 +213,10 @@ fn main() -> tantivy::Result<()> {
     //
     // We are not interested in all of the documents but
     // only in the top 10. Keeping track of our top 10 best documents
-    // is the role of the TopCollector.
-    let mut top_collector = TopCollector::with_limit(10);
+    // is the role of the TopDocs.
 
     // We can now perform our query.
-    searcher.search(&*query, &mut top_collector)?;
-
-    // Our top collector now contains the 10
-    // most relevant doc ids...
-    let doc_addresses = top_collector.docs();
+    let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;
 
     // The actual documents still need to be
     // retrieved from Tantivy's store.
@@ -229,12 +225,10 @@ fn main() -> tantivy::Result<()> {
     // the document returned will only contain
     // a title.
 
-    for doc_address in doc_addresses {
+    for (_score, doc_address) in top_docs {
         let retrieved_doc = searcher.doc(doc_address)?;
         println!("{}", schema.to_json(&retrieved_doc));
     }
 
     Ok(())
 }
-
-use tempdir::TempDir;

examples/custom_collector.rs

@@ -0,0 +1,187 @@
+// # Custom collector example
+//
+// This example shows how you can implement your own
+// collector. As an example, we will compute a collector
+// that computes the standard deviation of a given fast field.
+//
+// Of course, you can have a look at the tantivy's built-in collectors
+// such as the `CountCollector` for more examples.
+
+extern crate tempdir;
+
+// ---
+// Importing tantivy...
+#[macro_use]
+extern crate tantivy;
+use tantivy::collector::{Collector, SegmentCollector};
+use tantivy::fastfield::FastFieldReader;
+use tantivy::query::QueryParser;
+use tantivy::schema::Field;
+use tantivy::schema::{Schema, FAST, INT_INDEXED, TEXT};
+use tantivy::Index;
+use tantivy::SegmentReader;
+
+#[derive(Default)]
+struct Stats {
+    count: usize,
+    sum: f64,
+    squared_sum: f64,
+}
+
+impl Stats {
+    pub fn count(&self) -> usize {
+        self.count
+    }
+
+    pub fn mean(&self) -> f64 {
+        self.sum / (self.count as f64)
+    }
+
+    fn square_mean(&self) -> f64 {
+        self.squared_sum / (self.count as f64)
+    }
+
+    pub fn standard_deviation(&self) -> f64 {
+        let mean = self.mean();
+        (self.square_mean() - mean * mean).sqrt()
+    }
+
+    fn non_zero_count(self) -> Option<Stats> {
+        if self.count == 0 {
+            None
+        } else {
+            Some(self)
+        }
+    }
+}
+
+struct StatsCollector {
+    field: Field,
+}
+
+impl StatsCollector {
+    fn with_field(field: Field) -> StatsCollector {
+        StatsCollector { field }
+    }
+}
+
+impl Collector for StatsCollector {
+    // That's the type of our result.
+    // Our standard deviation will be a float.
+    type Fruit = Option<Stats>;
+
+    type Child = StatsSegmentCollector;
+
+    fn for_segment(
+        &self,
+        _segment_local_id: u32,
+        segment: &SegmentReader,
+    ) -> tantivy::Result<StatsSegmentCollector> {
+        let fast_field_reader = segment.fast_field_reader(self.field)?;
+        Ok(StatsSegmentCollector {
+            fast_field_reader,
+            stats: Stats::default(),
+        })
+    }
+
+    fn requires_scoring(&self) -> bool {
+        // this collector does not care about score.
+        false
+    }
+
+    fn merge_fruits(&self, segment_stats: Vec<Option<Stats>>) -> tantivy::Result<Option<Stats>> {
+        let mut stats = Stats::default();
+        for segment_stats_opt in segment_stats {
+            if let Some(segment_stats) = segment_stats_opt {
+                stats.count += segment_stats.count;
+                stats.sum += segment_stats.sum;
+                stats.squared_sum += segment_stats.squared_sum;
+            }
+        }
+        Ok(stats.non_zero_count())
+    }
+}
+
+struct StatsSegmentCollector {
+    fast_field_reader: FastFieldReader<u64>,
+    stats: Stats,
+}
+
+impl SegmentCollector for StatsSegmentCollector {
+    type Fruit = Option<Stats>;
+
+    fn collect(&mut self, doc: u32, _score: f32) {
+        let value = self.fast_field_reader.get(doc) as f64;
+        self.stats.count += 1;
+        self.stats.sum += value;
+        self.stats.squared_sum += value * value;
+    }
+
+    fn harvest(self) -> <Self as SegmentCollector>::Fruit {
+        self.stats.non_zero_count()
+    }
+}
+
+fn main() -> tantivy::Result<()> {
+    // # Defining the schema
+    //
+    // The Tantivy index requires a very strict schema.
+    // The schema declares which fields are in the index,
+    // and for each field, its type and "the way it should
+    // be indexed".
+
+    // first we need to define a schema ...
+    let mut schema_builder = Schema::builder();
+
+    // We'll assume a fictional index containing
+    // products, and with a name, a description, and a price.
+    let product_name = schema_builder.add_text_field("name", TEXT);
+    let product_description = schema_builder.add_text_field("description", TEXT);
+    let price = schema_builder.add_u64_field("price", INT_INDEXED | FAST);
+    let schema = schema_builder.build();
+
+    // # Indexing documents
+    //
+    // Lets index a bunch of fake documents for the sake of
+    // this example.
+    let index = Index::create_in_ram(schema.clone());
+
+    let mut index_writer = index.writer(50_000_000)?;
+    index_writer.add_document(doc!(
+        product_name => "Super Broom 2000",
+        product_description => "While it is ok for short distance travel, this broom \
+        was designed quiditch. It will up your game.",
+        price => 30_200u64
+    ));
+    index_writer.add_document(doc!(
+        product_name => "Turbulobroom",
+        product_description => "You might have heard of this broom before : it is the sponsor of the Wales team.\
+            You'll enjoy its sharp turns, and rapid acceleration",
+        price => 29_240u64
+    ));
+    index_writer.add_document(doc!(
+        product_name => "Broomio",
+        product_description => "Great value for the price. This broom is a market favorite",
+        price => 21_240u64
+    ));
+    index_writer.add_document(doc!(
+        product_name => "Whack a Mole",
+        product_description => "Prime quality bat.",
+        price => 5_200u64
+    ));
+    index_writer.commit()?;
+    index.load_searchers()?;
+
+    let searcher = index.searcher();
+    let query_parser = QueryParser::for_index(&index, vec![product_name, product_description]);
+
+    // here we want to get a hit on the 'ken' in Frankenstein
+    let query = query_parser.parse_query("broom")?;
+    if let Some(stats) = searcher.search(&query, &StatsCollector::with_field(price))? {
+        println!("count: {}", stats.count());
+        println!("mean: {}", stats.mean());
+        println!("standard deviation: {}", stats.standard_deviation());
+    }
+
+    Ok(())
+}

examples/custom_tokenizer.rs

@@ -5,7 +5,7 @@
 
 #[macro_use]
 extern crate tantivy;
-use tantivy::collector::TopCollector;
+use tantivy::collector::TopDocs;
 use tantivy::query::QueryParser;
 use tantivy::schema::*;
 use tantivy::tokenizer::NgramTokenizer;
@@ -20,7 +20,7 @@ fn main() -> tantivy::Result<()> {
     // be indexed".
 
     // first we need to define a schema ...
-    let mut schema_builder = SchemaBuilder::default();
+    let mut schema_builder = Schema::builder();
 
     // Our first field is title.
     // In this example we want to use NGram searching
@@ -68,12 +68,12 @@ fn main() -> tantivy::Result<()> {
     // heap for the indexer can increase its throughput.
     let mut index_writer = index.writer(50_000_000)?;
     index_writer.add_document(doc!(
-        title => "The Old Man and the Sea",
-        body => "He was an old man who fished alone in a skiff in the Gulf Stream and \
-         he had gone eighty-four days now without taking a fish."
+    title => "The Old Man and the Sea",
+    body => "He was an old man who fished alone in a skiff in the Gulf Stream and \
+     he had gone eighty-four days now without taking a fish."
     ));
     index_writer.add_document(doc!(
-       title => "Of Mice and Men",
+    title => "Of Mice and Men",
        body => r#"A few miles south of Soledad, the Salinas River drops in close to the hillside
                 bank and runs deep and green. The water is warm too, for it has slipped twinkling
                 over the yellow sands in the sunlight before reaching the narrow pool. On one
@@ -84,7 +84,7 @@ fn main() -> tantivy::Result<()> {
                 limbs and branches that arch over the pool"#
     ));
     index_writer.add_document(doc!(
-        title => "Frankenstein",
+    title => "Frankenstein",
         body => r#"You will rejoice to hear that no disaster has accompanied the commencement of an
                 enterprise which you have regarded with such evil forebodings.  I arrived here
                 yesterday, and my first task is to assure my dear sister of my welfare and
@@ -104,11 +104,9 @@ fn main() -> tantivy::Result<()> {
     // here we want to get a hit on the 'ken' in Frankenstein
     let query = query_parser.parse_query("ken")?;
 
-    let mut top_collector = TopCollector::with_limit(10);
-    searcher.search(&*query, &mut top_collector)?;
+    let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;
 
-    let doc_addresses = top_collector.docs();
-    for doc_address in doc_addresses {
+    for (_, doc_address) in top_docs {
         let retrieved_doc = searcher.doc(doc_address)?;
         println!("{}", schema.to_json(&retrieved_doc));
     }

examples/deleting_updating_documents.rs

@@ -10,7 +10,7 @@
 // Importing tantivy...
 #[macro_use]
 extern crate tantivy;
-use tantivy::collector::TopCollector;
+use tantivy::collector::TopDocs;
 use tantivy::query::TermQuery;
 use tantivy::schema::*;
 use tantivy::Index;
@@ -27,10 +27,9 @@ fn extract_doc_given_isbn(index: &Index, isbn_term: &Term) -> tantivy::Result<Op
     // The second argument is here to tell we don't care about decoding positions,
     // or term frequencies.
     let term_query = TermQuery::new(isbn_term.clone(), IndexRecordOption::Basic);
-    let mut top_collector = TopCollector::with_limit(1);
-    searcher.search(&term_query, &mut top_collector)?;
+    let top_docs = searcher.search(&term_query, &TopDocs::with_limit(1))?;
 
-    if let Some(doc_address) = top_collector.docs().first() {
+    if let Some((_score, doc_address)) = top_docs.first() {
         let doc = searcher.doc(*doc_address)?;
         Ok(Some(doc))
     } else {
@@ -44,7 +43,7 @@ fn main() -> tantivy::Result<()> {
     //
     // Check out the *basic_search* example if this makes
     // small sense to you.
-    let mut schema_builder = SchemaBuilder::default();
+    let mut schema_builder = Schema::builder();
 
     // Tantivy does not really have a notion of primary id.
     // This may change in the future.

examples/faceted_search.rs

@@ -25,7 +25,7 @@ fn main() -> tantivy::Result<()> {
     // Let's create a temporary directory for the
     // sake of this example
     let index_path = TempDir::new("tantivy_facet_example_dir")?;
-    let mut schema_builder = SchemaBuilder::default();
+    let mut schema_builder = Schema::builder();
 
     schema_builder.add_text_field("name", TEXT | STORED);
 
@@ -62,11 +62,10 @@ fn main() -> tantivy::Result<()> {
     let mut facet_collector = FacetCollector::for_field(tags);
     facet_collector.add_facet("/pools");
 
-    searcher.search(&AllQuery, &mut facet_collector).unwrap();
+    let facet_counts = searcher.search(&AllQuery, &facet_collector).unwrap();
 
-    let counts = facet_collector.harvest();
     // This lists all of the facet counts
-    let facets: Vec<(&Facet, u64)> = counts.get("/pools").collect();
+    let facets: Vec<(&Facet, u64)> = facet_counts.get("/pools").collect();
     assert_eq!(
         facets,
         vec![

examples/iterating_docs_and_positions.rs

@@ -18,7 +18,7 @@ use tantivy::{DocId, DocSet, Postings};
 fn main() -> tantivy::Result<()> {
     // We first create a schema for the sake of the
     // example. Check the `basic_search` example for more information.
-    let mut schema_builder = SchemaBuilder::default();
+    let mut schema_builder = Schema::builder();
 
     // For this example, we need to make sure to index positions for our title
     // field. `TEXT` precisely does this.

examples/snippet.rs

@@ -10,11 +10,11 @@ extern crate tempdir;
 // Importing tantivy...
 #[macro_use]
 extern crate tantivy;
-use tantivy::collector::TopCollector;
+use tantivy::collector::TopDocs;
 use tantivy::query::QueryParser;
 use tantivy::schema::*;
 use tantivy::Index;
-use tantivy::SnippetGenerator;
+use tantivy::{Snippet, SnippetGenerator};
 use tempdir::TempDir;
 
 fn main() -> tantivy::Result<()> {
@@ -23,7 +23,7 @@ fn main() -> tantivy::Result<()> {
     let index_path = TempDir::new("tantivy_example_dir")?;
 
     // # Defining the schema
-    let mut schema_builder = SchemaBuilder::default();
+    let mut schema_builder = Schema::builder();
     let title = schema_builder.add_text_field("title", TEXT | STORED);
     let body = schema_builder.add_text_field("body", TEXT | STORED);
     let schema = schema_builder.build();
@@ -35,15 +35,15 @@ fn main() -> tantivy::Result<()> {
 
     // we'll only need one doc for this example.
     index_writer.add_document(doc!(
-        title => "Of Mice and Men",
-        body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
-                bank and runs deep and green. The water is warm too, for it has slipped twinkling \
-                over the yellow sands in the sunlight before reaching the narrow pool. On one \
-                side of the river the golden foothill slopes curve up to the strong and rocky \
-                Gabilan Mountains, but on the valley side the water is lined with trees—willows \
-                fresh and green with every spring, carrying in their lower leaf junctures the \
-                debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
-                limbs and branches that arch over the pool"
+    title => "Of Mice and Men",
+    body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
+            bank and runs deep and green. The water is warm too, for it has slipped twinkling \
+            over the yellow sands in the sunlight before reaching the narrow pool. On one \
+            side of the river the golden foothill slopes curve up to the strong and rocky \
+            Gabilan Mountains, but on the valley side the water is lined with trees—willows \
+            fresh and green with every spring, carrying in their lower leaf junctures the \
+            debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
+            limbs and branches that arch over the pool"
     ));
     // ...
     index_writer.commit()?;
@@ -54,18 +54,34 @@ fn main() -> tantivy::Result<()> {
     let query_parser = QueryParser::for_index(&index, vec![title, body]);
     let query = query_parser.parse_query("sycamore spring")?;
 
-    let mut top_collector = TopCollector::with_limit(10);
-    searcher.search(&*query, &mut top_collector)?;
+    let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;
 
-    let snippet_generator = SnippetGenerator::new(&searcher, &*query, body)?;
+    let snippet_generator = SnippetGenerator::create(&searcher, &*query, body)?;
 
-    let doc_addresses = top_collector.docs();
-    for doc_address in doc_addresses {
+    for (score, doc_address) in top_docs {
         let doc = searcher.doc(doc_address)?;
         let snippet = snippet_generator.snippet_from_doc(&doc);
+        println!("Document score {}:", score);
         println!("title: {}", doc.get_first(title).unwrap().text().unwrap());
         println!("snippet: {}", snippet.to_html());
+        println!("custom highlighting: {}", highlight(snippet));
     }
 
     Ok(())
 }
+
+fn highlight(snippet: Snippet) -> String {
+    let mut result = String::new();
+    let mut start_from = 0;
+
+    for (start, end) in snippet.highlighted().iter().map(|h| h.bounds()) {
+        result.push_str(&snippet.fragments()[start_from..start]);
+        result.push_str(" --> ");
+        result.push_str(&snippet.fragments()[start..end]);
+        result.push_str(" <-- ");
+        start_from = end;
+    }
+
+    result.push_str(&snippet.fragments()[start_from..]);
+    result
+}

examples/stop_words.rs

@@ -15,7 +15,7 @@ extern crate tempdir;
 // Importing tantivy...
 #[macro_use]
 extern crate tantivy;
-use tantivy::collector::TopCollector;
+use tantivy::collector::TopDocs;
 use tantivy::query::QueryParser;
 use tantivy::schema::*;
 use tantivy::tokenizer::*;
@@ -23,7 +23,7 @@ use tantivy::Index;
 
 fn main() -> tantivy::Result<()> {
     // this example assumes you understand the content in `basic_search`
-    let mut schema_builder = SchemaBuilder::default();
+    let mut schema_builder = Schema::builder();
 
     // This configures your custom options for how tantivy will
     // store and process your content in the index; The key
@@ -72,26 +72,26 @@ fn main() -> tantivy::Result<()> {
     title => "The Old Man and the Sea",
     body => "He was an old man who fished alone in a skiff in the Gulf Stream and \
      he had gone eighty-four days now without taking a fish."
-  ));
+    ));
 
     index_writer.add_document(doc!(
-      title => "Of Mice and Men",
-      body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
-              bank and runs deep and green. The water is warm too, for it has slipped twinkling \
-              over the yellow sands in the sunlight before reaching the narrow pool. On one \
-              side of the river the golden foothill slopes curve up to the strong and rocky \
-              Gabilan Mountains, but on the valley side the water is lined with trees—willows \
-              fresh and green with every spring, carrying in their lower leaf junctures the \
-              debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
-              limbs and branches that arch over the pool"
-  ));
+    title => "Of Mice and Men",
+    body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
+            bank and runs deep and green. The water is warm too, for it has slipped twinkling \
+            over the yellow sands in the sunlight before reaching the narrow pool. On one \
+            side of the river the golden foothill slopes curve up to the strong and rocky \
+            Gabilan Mountains, but on the valley side the water is lined with trees—willows \
+            fresh and green with every spring, carrying in their lower leaf junctures the \
+            debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
+            limbs and branches that arch over the pool"
+    ));
 
     index_writer.add_document(doc!(
-       title => "Frankenstein",
-       body => "You will rejoice to hear that no disaster has accompanied the commencement of an \
-                enterprise which you have regarded with such evil forebodings.  I arrived here \
-                yesterday, and my first task is to assure my dear sister of my welfare and \
-                increasing confidence in the success of my undertaking."
+    title => "Frankenstein",
+    body => "You will rejoice to hear that no disaster has accompanied the commencement of an \
+             enterprise which you have regarded with such evil forebodings.  I arrived here \
+             yesterday, and my first task is to assure my dear sister of my welfare and \
+             increasing confidence in the success of my undertaking."
     ));
 
     index_writer.commit()?;
@@ -105,15 +105,11 @@ fn main() -> tantivy::Result<()> {
     // stop words are applied on the query as well.
     // The following will be equivalent to `title:frankenstein`
     let query = query_parser.parse_query("title:\"the Frankenstein\"")?;
+    let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;
 
-    let mut top_collector = TopCollector::with_limit(10);
-
-    searcher.search(&*query, &mut top_collector)?;
-
-    let doc_addresses = top_collector.docs();
-
-    for doc_address in doc_addresses {
+    for (score, doc_address) in top_docs {
         let retrieved_doc = searcher.doc(doc_address)?;
+        println!("\n==\nDocument score {}:", score);
         println!("{}", schema.to_json(&retrieved_doc));
     }
 

examples/working_with_json.rs

@@ -9,7 +9,7 @@ fn main() -> tantivy::Result<()> {
     // Check out the basic example if this is confusing to you.
     //
     // first we need to define a schema ...
-    let mut schema_builder = SchemaBuilder::default();
+    let mut schema_builder = Schema::builder();
     schema_builder.add_text_field("title", TEXT | STORED);
     schema_builder.add_text_field("body", TEXT);
     schema_builder.add_u64_field("year", INT_INDEXED);

src/collector/chained_collector.rs

@@ -1,142 +0,0 @@
-use collector::Collector;
-use DocId;
-use Result;
-use Score;
-use SegmentLocalId;
-use SegmentReader;
-
-/// Collector that does nothing.
-/// This is used in the chain Collector and will hopefully
-/// be optimized away by the compiler.
-pub struct DoNothingCollector;
-impl Collector for DoNothingCollector {
-    #[inline]
-    fn set_segment(&mut self, _: SegmentLocalId, _: &SegmentReader) -> Result<()> {
-        Ok(())
-    }
-    #[inline]
-    fn collect(&mut self, _doc: DocId, _score: Score) {}
-    #[inline]
-    fn requires_scoring(&self) -> bool {
-        false
-    }
-}
-
-/// Zero-cost abstraction used to collect on multiple collectors.
-/// This contraption is only usable if the type of your collectors
-/// are known at compile time.
-///
-/// ```rust
-/// #[macro_use]
-/// extern crate tantivy;
-/// use tantivy::schema::{SchemaBuilder, TEXT};
-/// use tantivy::{Index, Result};
-/// use tantivy::collector::{CountCollector, TopCollector, chain};
-/// use tantivy::query::QueryParser;
-///
-/// # fn main() { example().unwrap(); }
-/// fn example() -> Result<()> {
-///     let mut schema_builder = SchemaBuilder::new();
-///     let title = schema_builder.add_text_field("title", TEXT);
-///     let schema = schema_builder.build();
-///     let index = Index::create_in_ram(schema);
-///     {
-///         let mut index_writer = index.writer(3_000_000)?;
-///         index_writer.add_document(doc!(
-///             title => "The Name of the Wind",
-///         ));
-///         index_writer.add_document(doc!(
-///             title => "The Diary of Muadib",
-///         ));
-///         index_writer.add_document(doc!(
-///             title => "A Dairy Cow",
-///         ));
-///         index_writer.add_document(doc!(
-///             title => "The Diary of a Young Girl",
-///         ));
-///         index_writer.commit().unwrap();
-///     }
-///
-///     index.load_searchers()?;
-///     let searcher = index.searcher();
-///
-///     {
-///         let mut top_collector = TopCollector::with_limit(2);
-///         let mut count_collector = CountCollector::default();
-///         {
-///             let mut collectors = chain().push(&mut top_collector).push(&mut count_collector);
-///             let query_parser = QueryParser::for_index(&index, vec![title]);
-///             let query = query_parser.parse_query("diary")?;
-///             searcher.search(&*query, &mut collectors).unwrap();
-///         }
-///         assert_eq!(count_collector.count(), 2);
-///         assert!(top_collector.at_capacity());
-///     }
-///
-///     Ok(())
-/// }
-/// ```
-pub struct ChainedCollector<Left: Collector, Right: Collector> {
-    left: Left,
-    right: Right,
-}
-
-impl<Left: Collector, Right: Collector> ChainedCollector<Left, Right> {
-    /// Adds a collector
-    pub fn push<C: Collector>(self, new_collector: &mut C) -> ChainedCollector<Self, &mut C> {
-        ChainedCollector {
-            left: self,
-            right: new_collector,
-        }
-    }
-}
-
-impl<Left: Collector, Right: Collector> Collector for ChainedCollector<Left, Right> {
-    fn set_segment(
-        &mut self,
-        segment_local_id: SegmentLocalId,
-        segment: &SegmentReader,
-    ) -> Result<()> {
-        self.left.set_segment(segment_local_id, segment)?;
-        self.right.set_segment(segment_local_id, segment)?;
-        Ok(())
-    }
-
-    fn collect(&mut self, doc: DocId, score: Score) {
-        self.left.collect(doc, score);
-        self.right.collect(doc, score);
-    }
-
-    fn requires_scoring(&self) -> bool {
-        self.left.requires_scoring() || self.right.requires_scoring()
-    }
-}
-
-/// Creates a `ChainedCollector`
-pub fn chain() -> ChainedCollector<DoNothingCollector, DoNothingCollector> {
-    ChainedCollector {
-        left: DoNothingCollector,
-        right: DoNothingCollector,
-    }
-}
-
-#[cfg(test)]
-mod tests {
-
-    use super::*;
-    use collector::{Collector, CountCollector, TopCollector};
-
-    #[test]
-    fn test_chained_collector() {
-        let mut top_collector = TopCollector::with_limit(2);
-        let mut count_collector = CountCollector::default();
-        {
-            let mut collectors = chain().push(&mut top_collector).push(&mut count_collector);
-            collectors.collect(1, 0.2);
-            collectors.collect(2, 0.1);
-            collectors.collect(3, 0.5);
-        }
-        assert_eq!(count_collector.count(), 3);
-        assert!(top_collector.at_capacity());
-    }
-}

src/collector/count_collector.rs

@@ -1,4 +1,5 @@
 use super::Collector;
+use collector::SegmentCollector;
 use DocId;
 use Result;
 use Score;
@@ -11,14 +12,14 @@ use SegmentReader;
 /// ```rust
 /// #[macro_use]
 /// extern crate tantivy;
-/// use tantivy::schema::{SchemaBuilder, TEXT};
+/// use tantivy::schema::{Schema, TEXT};
 /// use tantivy::{Index, Result};
-/// use tantivy::collector::CountCollector;
+/// use tantivy::collector::Count;
 /// use tantivy::query::QueryParser;
 ///
 /// # fn main() { example().unwrap(); }
 /// fn example() -> Result<()> {
-///     let mut schema_builder = SchemaBuilder::new();
+///     let mut schema_builder = Schema::builder();
 ///     let title = schema_builder.add_text_field("title", TEXT);
 ///     let schema = schema_builder.build();
 ///     let index = Index::create_in_ram(schema);
@@ -43,59 +44,86 @@ use SegmentReader;
 ///     let searcher = index.searcher();
 ///
 ///     {
-///	        let mut count_collector = CountCollector::default();
 ///         let query_parser = QueryParser::for_index(&index, vec![title]);
 ///         let query = query_parser.parse_query("diary")?;
-///         searcher.search(&*query, &mut count_collector).unwrap();
+///         let count = searcher.search(&query, &Count).unwrap();
 ///
-///         assert_eq!(count_collector.count(), 2);
+///         assert_eq!(count, 2);
 ///     }
 ///
 ///     Ok(())
 /// }
 /// ```
-#[derive(Default)]
-pub struct CountCollector {
-    count: usize,
-}
+pub struct Count;
 
-impl CountCollector {
-    /// Returns the count of documents that were
-    /// collected.
-    pub fn count(&self) -> usize {
-        self.count
-    }
-}
+impl Collector for Count {
+    type Fruit = usize;
 
-impl Collector for CountCollector {
-    fn set_segment(&mut self, _: SegmentLocalId, _: &SegmentReader) -> Result<()> {
-        Ok(())
-    }
+    type Child = SegmentCountCollector;
 
-    fn collect(&mut self, _: DocId, _: Score) {
-        self.count += 1;
+    fn for_segment(&self, _: SegmentLocalId, _: &SegmentReader) -> Result<SegmentCountCollector> {
+        Ok(SegmentCountCollector::default())
     }
 
     fn requires_scoring(&self) -> bool {
         false
     }
+
+    fn merge_fruits(&self, segment_counts: Vec<usize>) -> Result<usize> {
+        Ok(segment_counts.into_iter().sum())
+    }
+}
+
+#[derive(Default)]
+pub struct SegmentCountCollector {
+    count: usize,
+}
+
+impl SegmentCollector for SegmentCountCollector {
+    type Fruit = usize;
+
+    fn collect(&mut self, _: DocId, _: Score) {
+        self.count += 1;
+    }
+
+    fn harvest(self) -> usize {
+        self.count
+    }
 }
 
 #[cfg(test)]
 mod tests {
-
-    use collector::{Collector, CountCollector};
+    use super::{Count, SegmentCountCollector};
+    use collector::Collector;
+    use collector::SegmentCollector;
 
     #[test]
-    fn test_count_collector() {
-        let mut count_collector = CountCollector::default();
-        assert_eq!(count_collector.count(), 0);
-        count_collector.collect(0u32, 1f32);
-        assert_eq!(count_collector.count(), 1);
-        assert_eq!(count_collector.count(), 1);
-        count_collector.collect(1u32, 1f32);
-        assert_eq!(count_collector.count(), 2);
-        assert!(!count_collector.requires_scoring());
+    fn test_count_collect_does_not_requires_scoring() {
+        assert!(!Count.requires_scoring());
+    }
+
+    #[test]
+    fn test_segment_count_collector() {
+        {
+            let count_collector = SegmentCountCollector::default();
+            assert_eq!(count_collector.harvest(), 0);
+        }
+        {
+            let mut count_collector = SegmentCountCollector::default();
+            count_collector.collect(0u32, 1f32);
+            assert_eq!(count_collector.harvest(), 1);
+        }
+        {
+            let mut count_collector = SegmentCountCollector::default();
+            count_collector.collect(0u32, 1f32);
+            assert_eq!(count_collector.harvest(), 1);
+        }
+        {
+            let mut count_collector = SegmentCountCollector::default();
+            count_collector.collect(0u32, 1f32);
+            count_collector.collect(1u32, 1f32);
+            assert_eq!(count_collector.harvest(), 2);
+        }
     }
 
 }

src/collector/facet_collector.rs

@@ -1,20 +1,17 @@
 use collector::Collector;
+use collector::SegmentCollector;
 use docset::SkipResult;
 use fastfield::FacetReader;
 use schema::Facet;
 use schema::Field;
-use std::cell::UnsafeCell;
+use std::cmp::Ordering;
 use std::collections::btree_map;
 use std::collections::BTreeMap;
 use std::collections::BTreeSet;
 use std::collections::BinaryHeap;
 use std::collections::Bound;
 use std::iter::Peekable;
-use std::mem;
 use std::{u64, usize};
-use termdict::TermMerger;
-
-use std::cmp::Ordering;
 use DocId;
 use Result;
 use Score;
@@ -46,12 +43,6 @@ impl<'a> Ord for Hit<'a> {
     }
 }
 
-struct SegmentFacetCounter {
-    pub facet_reader: FacetReader,
-    pub facet_ords: Vec<u64>,
-    pub facet_counts: Vec<u64>,
-}
-
 fn facet_depth(facet_bytes: &[u8]) -> usize {
     if facet_bytes.is_empty() {
         0
@@ -91,14 +82,14 @@ fn facet_depth(facet_bytes: &[u8]) -> usize {
 /// ```rust
 /// #[macro_use]
 /// extern crate tantivy;
-/// use tantivy::schema::{Facet, SchemaBuilder, TEXT};
+/// use tantivy::schema::{Facet, Schema, TEXT};
 /// use tantivy::{Index, Result};
 /// use tantivy::collector::FacetCollector;
 /// use tantivy::query::AllQuery;
 ///
 /// # fn main() { example().unwrap(); }
 /// fn example() -> Result<()> {
-///     let mut schema_builder = SchemaBuilder::new();
+///     let mut schema_builder = Schema::builder();
 ///
 ///     // Facet have their own specific type.
 ///     // It is not a bad practise to put all of your
@@ -141,13 +132,10 @@ fn facet_depth(facet_bytes: &[u8]) -> usize {
 ///			let mut facet_collector = FacetCollector::for_field(facet);
 ///         facet_collector.add_facet("/lang");
 ///         facet_collector.add_facet("/category");
-///         searcher.search(&AllQuery, &mut facet_collector).unwrap();
-///
-///         // this object contains count aggregate for all of the facets.
-///         let counts = facet_collector.harvest();
+///         let facet_counts = searcher.search(&AllQuery, &facet_collector).unwrap();
 ///
 ///         // This lists all of the facet counts
-///         let facets: Vec<(&Facet, u64)> = counts
+///         let facets: Vec<(&Facet, u64)> = facet_counts
 ///             .get("/category")
 ///             .collect();
 ///         assert_eq!(facets, vec![
@@ -159,13 +147,10 @@ fn facet_depth(facet_bytes: &[u8]) -> usize {
 ///     {
 ///			let mut facet_collector = FacetCollector::for_field(facet);
 ///         facet_collector.add_facet("/category/fiction");
-///         searcher.search(&AllQuery, &mut facet_collector).unwrap();
-///
-///         // this object contains count aggregate for all of the facets.
-///         let counts = facet_collector.harvest();
+///         let facet_counts = searcher.search(&AllQuery, &facet_collector).unwrap();
 ///
 ///         // This lists all of the facet counts
-///         let facets: Vec<(&Facet, u64)> = counts
+///         let facets: Vec<(&Facet, u64)> = facet_counts
 ///             .get("/category/fiction")
 ///             .collect();
 ///         assert_eq!(facets, vec![
@@ -178,13 +163,10 @@ fn facet_depth(facet_bytes: &[u8]) -> usize {
 ///    {
 ///			let mut facet_collector = FacetCollector::for_field(facet);
 ///         facet_collector.add_facet("/category/fiction");
-///         searcher.search(&AllQuery, &mut facet_collector).unwrap();
-///
-///         // this object contains count aggregate for all of the facets.
-///         let counts = facet_collector.harvest();
+///         let facet_counts = searcher.search(&AllQuery, &facet_collector).unwrap();
 ///
 ///         // This lists all of the facet counts
-///         let facets: Vec<(&Facet, u64)> = counts.top_k("/category/fiction", 1);
+///         let facets: Vec<(&Facet, u64)> = facet_counts.top_k("/category/fiction", 1);
 ///         assert_eq!(facets, vec![
 ///             (&Facet::from("/category/fiction/fantasy"), 2)
 ///         ]);
@@ -194,28 +176,28 @@ fn facet_depth(facet_bytes: &[u8]) -> usize {
 /// }
 /// ```
 pub struct FacetCollector {
-    facet_ords: Vec<u64>,
     field: Field,
-    ff_reader: Option<UnsafeCell<FacetReader>>,
-    segment_counters: Vec<SegmentFacetCounter>,
-
-    // facet_ord -> collapse facet_id
-    current_segment_collapse_mapping: Vec<usize>,
-    // collapse facet_id -> count
-    current_segment_counts: Vec<u64>,
-    // collapse facet_id -> facet_ord
-    current_collapse_facet_ords: Vec<u64>,
-
     facets: BTreeSet<Facet>,
 }
 
+pub struct FacetSegmentCollector {
+    reader: FacetReader,
+    facet_ords_buf: Vec<u64>,
+    // facet_ord -> collapse facet_id
+    collapse_mapping: Vec<usize>,
+    // collapse facet_id -> count
+    counts: Vec<u64>,
+    // collapse facet_id -> facet_ord
+    collapse_facet_ords: Vec<u64>,
+}
+
 fn skip<'a, I: Iterator<Item = &'a Facet>>(
     target: &[u8],
     collapse_it: &mut Peekable<I>,
 ) -> SkipResult {
     loop {
         match collapse_it.peek() {
-            Some(facet_bytes) => match facet_bytes.encoded_bytes().cmp(target) {
+            Some(facet_bytes) => match facet_bytes.encoded_str().as_bytes().cmp(target) {
                 Ordering::Less => {}
                 Ordering::Greater => {
                     return SkipResult::OverStep;
@@ -240,15 +222,8 @@ impl FacetCollector {
     /// is of the proper type.
     pub fn for_field(field: Field) -> FacetCollector {
         FacetCollector {
-            facet_ords: Vec::with_capacity(255),
-            segment_counters: Vec::new(),
             field,
-            ff_reader: None,
-            facets: BTreeSet::new(),
-
-            current_segment_collapse_mapping: Vec::new(),
-            current_collapse_facet_ords: Vec::new(),
-            current_segment_counts: Vec::new(),
+            facets: BTreeSet::default(),
         }
     }
 
@@ -278,143 +253,100 @@ impl FacetCollector {
         }
         self.facets.insert(facet);
     }
-
-    fn set_collapse_mapping(&mut self, facet_reader: &FacetReader) {
-        self.current_segment_collapse_mapping.clear();
-        self.current_collapse_facet_ords.clear();
-        self.current_segment_counts.clear();
-        let mut collapse_facet_it = self.facets.iter().peekable();
-        self.current_collapse_facet_ords.push(0);
-        let mut facet_streamer = facet_reader.facet_dict().range().into_stream();
-        if !facet_streamer.advance() {
-            return;
-        }
-        'outer: loop {
-            // at the begining of this loop, facet_streamer
-            // is positionned on a term that has not been processed yet.
-            let skip_result = skip(facet_streamer.key(), &mut collapse_facet_it);
-            match skip_result {
-                SkipResult::Reached => {
-                    // we reach a facet we decided to collapse.
-                    let collapse_depth = facet_depth(facet_streamer.key());
-                    let mut collapsed_id = 0;
-                    self.current_segment_collapse_mapping.push(0);
-                    while facet_streamer.advance() {
-                        let depth = facet_depth(facet_streamer.key());
-                        if depth <= collapse_depth {
-                            continue 'outer;
-                        }
-                        if depth == collapse_depth + 1 {
-                            collapsed_id = self.current_collapse_facet_ords.len();
-                            self.current_collapse_facet_ords
-                                .push(facet_streamer.term_ord());
-                            self.current_segment_collapse_mapping.push(collapsed_id);
-                        } else {
-                            self.current_segment_collapse_mapping.push(collapsed_id);
-                        }
-                    }
-                    break;
-                }
-                SkipResult::End | SkipResult::OverStep => {
-                    self.current_segment_collapse_mapping.push(0);
-                    if !facet_streamer.advance() {
-                        break;
-                    }
-                }
-            }
-        }
-    }
-
-    fn finalize_segment(&mut self) {
-        if self.ff_reader.is_some() {
-            self.segment_counters.push(SegmentFacetCounter {
-                facet_reader: self.ff_reader.take().unwrap().into_inner(),
-                facet_ords: mem::replace(&mut self.current_collapse_facet_ords, Vec::new()),
-                facet_counts: mem::replace(&mut self.current_segment_counts, Vec::new()),
-            });
-        }
-    }
-
-    /// Returns the results of the collection.
-    ///
-    /// This method does not just return the counters,
-    /// it also translates the facet ordinals of the last segment.
-    pub fn harvest(mut self) -> FacetCounts {
-        self.finalize_segment();
-
-        let collapsed_facet_ords: Vec<&[u64]> = self
-            .segment_counters
-            .iter()
-            .map(|segment_counter| &segment_counter.facet_ords[..])
-            .collect();
-        let collapsed_facet_counts: Vec<&[u64]> = self
-            .segment_counters
-            .iter()
-            .map(|segment_counter| &segment_counter.facet_counts[..])
-            .collect();
-
-        let facet_streams = self
-            .segment_counters
-            .iter()
-            .map(|seg_counts| seg_counts.facet_reader.facet_dict().range().into_stream())
-            .collect::<Vec<_>>();
-
-        let mut facet_merger = TermMerger::new(facet_streams);
-        let mut facet_counts = BTreeMap::new();
-
-        while facet_merger.advance() {
-            let count = facet_merger
-                .current_kvs()
-                .iter()
-                .map(|it| {
-                    let seg_ord = it.segment_ord;
-                    let term_ord = it.streamer.term_ord();
-                    collapsed_facet_ords[seg_ord]
-                        .binary_search(&term_ord)
-                        .map(|collapsed_term_id| {
-                            if collapsed_term_id == 0 {
-                                0
-                            } else {
-                                collapsed_facet_counts[seg_ord][collapsed_term_id]
-                            }
-                        }).unwrap_or(0)
-                }).sum();
-            if count > 0u64 {
-                let bytes: Vec<u8> = facet_merger.key().to_owned();
-                // may create an corrupted facet if the term dicitonary is corrupted
-                let facet = unsafe { Facet::from_encoded(bytes) };
-                facet_counts.insert(facet, count);
-            }
-        }
-        FacetCounts { facet_counts }
-    }
 }
 
 impl Collector for FacetCollector {
-    fn set_segment(&mut self, _: SegmentLocalId, reader: &SegmentReader) -> Result<()> {
-        self.finalize_segment();
+    type Fruit = FacetCounts;
+
+    type Child = FacetSegmentCollector;
+
+    fn for_segment(
+        &self,
+        _: SegmentLocalId,
+        reader: &SegmentReader,
+    ) -> Result<FacetSegmentCollector> {
         let facet_reader = reader.facet_reader(self.field)?;
-        self.set_collapse_mapping(&facet_reader);
-        self.current_segment_counts
-            .resize(self.current_collapse_facet_ords.len(), 0);
-        self.ff_reader = Some(UnsafeCell::new(facet_reader));
-        Ok(())
+
+        let mut collapse_mapping = Vec::new();
+        let mut counts = Vec::new();
+        let mut collapse_facet_ords = Vec::new();
+
+        let mut collapse_facet_it = self.facets.iter().peekable();
+        collapse_facet_ords.push(0);
+        {
+            let mut facet_streamer = facet_reader.facet_dict().range().into_stream();
+            if facet_streamer.advance() {
+                'outer: loop {
+                    // at the begining of this loop, facet_streamer
+                    // is positionned on a term that has not been processed yet.
+                    let skip_result = skip(facet_streamer.key(), &mut collapse_facet_it);
+                    match skip_result {
+                        SkipResult::Reached => {
+                            // we reach a facet we decided to collapse.
+                            let collapse_depth = facet_depth(facet_streamer.key());
+                            let mut collapsed_id = 0;
+                            collapse_mapping.push(0);
+                            while facet_streamer.advance() {
+                                let depth = facet_depth(facet_streamer.key());
+                                if depth <= collapse_depth {
+                                    continue 'outer;
+                                }
+                                if depth == collapse_depth + 1 {
+                                    collapsed_id = collapse_facet_ords.len();
+                                    collapse_facet_ords.push(facet_streamer.term_ord());
+                                    collapse_mapping.push(collapsed_id);
+                                } else {
+                                    collapse_mapping.push(collapsed_id);
+                                }
+                            }
+                            break;
+                        }
+                        SkipResult::End | SkipResult::OverStep => {
+                            collapse_mapping.push(0);
+                            if !facet_streamer.advance() {
+                                break;
+                            }
+                        }
+                    }
+                }
+            }
+        }
+
+        counts.resize(collapse_facet_ords.len(), 0);
+
+        Ok(FacetSegmentCollector {
+            reader: facet_reader,
+            facet_ords_buf: Vec::with_capacity(255),
+            collapse_mapping,
+            counts,
+            collapse_facet_ords,
+        })
     }
 
+    fn requires_scoring(&self) -> bool {
+        false
+    }
+
+    fn merge_fruits(&self, segments_facet_counts: Vec<FacetCounts>) -> Result<FacetCounts> {
+        let mut facet_counts: BTreeMap<Facet, u64> = BTreeMap::new();
+        for segment_facet_counts in segments_facet_counts {
+            for (facet, count) in segment_facet_counts.facet_counts {
+                *(facet_counts.entry(facet).or_insert(0)) += count;
+            }
+        }
+        Ok(FacetCounts { facet_counts })
+    }
+}
+
+impl SegmentCollector for FacetSegmentCollector {
+    type Fruit = FacetCounts;
+
     fn collect(&mut self, doc: DocId, _: Score) {
-        let facet_reader: &mut FacetReader = unsafe {
-            &mut *self
-                .ff_reader
-                .as_ref()
-                .expect("collect() was called before set_segment. This should never happen.")
-                .get()
-        };
-        facet_reader.facet_ords(doc, &mut self.facet_ords);
+        self.reader.facet_ords(doc, &mut self.facet_ords_buf);
         let mut previous_collapsed_ord: usize = usize::MAX;
-        for &facet_ord in &self.facet_ords {
-            let collapsed_ord = self.current_segment_collapse_mapping[facet_ord as usize];
-            self.current_segment_counts[collapsed_ord] += if collapsed_ord == previous_collapsed_ord
-            {
+        for &facet_ord in &self.facet_ords_buf {
+            let collapsed_ord = self.collapse_mapping[facet_ord as usize];
+            self.counts[collapsed_ord] += if collapsed_ord == previous_collapsed_ord {
                 0
             } else {
                 1
@@ -423,8 +355,24 @@ impl Collector for FacetCollector {
         }
     }
 
-    fn requires_scoring(&self) -> bool {
-        false
+    /// Returns the results of the collection.
+    ///
+    /// This method does not just return the counters,
+    /// it also translates the facet ordinals of the last segment.
+    fn harvest(self) -> FacetCounts {
+        let mut facet_counts = BTreeMap::new();
+        let facet_dict = self.reader.facet_dict();
+        for (collapsed_facet_ord, count) in self.counts.iter().cloned().enumerate() {
+            if count == 0 {
+                continue;
+            }
+            let mut facet = vec![];
+            let facet_ord = self.collapse_facet_ords[collapsed_facet_ord];
+            facet_dict.ord_to_term(facet_ord as u64, &mut facet);
+            // TODO
+            facet_counts.insert(Facet::from_encoded(facet).unwrap(), count);
+        }
+        FacetCounts { facet_counts }
     }
 }
 
@@ -456,9 +404,9 @@ impl FacetCounts {
         let right_bound = if facet.is_root() {
             Bound::Unbounded
         } else {
-            let mut facet_after_bytes: Vec<u8> = facet.encoded_bytes().to_owned();
-            facet_after_bytes.push(1u8);
-            let facet_after = unsafe { Facet::from_encoded(facet_after_bytes) }; // ok logic
+            let mut facet_after_bytes: String = facet.encoded_str().to_owned();
+            facet_after_bytes.push('\u{1}');
+            let facet_after = Facet::from_encoded_string(facet_after_bytes);
             Bound::Excluded(facet_after)
         };
         let underlying: btree_map::Range<_, _> = self.facet_counts.range((left_bound, right_bound));
@@ -505,14 +453,14 @@ mod tests {
     use core::Index;
     use query::AllQuery;
     use rand::distributions::Uniform;
+    use rand::prelude::SliceRandom;
     use rand::{thread_rng, Rng};
-    use schema::Field;
-    use schema::{Document, Facet, SchemaBuilder};
+    use schema::{Document, Facet, Field, Schema};
     use std::iter;
 
     #[test]
     fn test_facet_collector_drilldown() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let facet_field = schema_builder.add_facet_field("facet");
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -527,7 +475,8 @@ mod tests {
                 n /= 4;
                 let leaf = n % 5;
                 Facet::from(&format!("/top{}/mid{}/leaf{}", top, mid, leaf))
-            }).collect();
+            })
+            .collect();
         for i in 0..num_facets * 10 {
             let mut doc = Document::new();
             doc.add_facet(facet_field, facets[i % num_facets].clone());
@@ -536,12 +485,10 @@ mod tests {
         index_writer.commit().unwrap();
         index.load_searchers().unwrap();
         let searcher = index.searcher();
-
         let mut facet_collector = FacetCollector::for_field(facet_field);
         facet_collector.add_facet(Facet::from("/top1"));
-        searcher.search(&AllQuery, &mut facet_collector).unwrap();
+        let counts = searcher.search(&AllQuery, &facet_collector).unwrap();
 
-        let counts: FacetCounts = facet_collector.harvest();
         {
             let facets: Vec<(String, u64)> = counts
                 .get("/top1")
@@ -555,18 +502,16 @@ mod tests {
                     ("/top1/mid2", 50),
                     ("/top1/mid3", 50),
                 ]
-                    .iter()
-                    .map(|&(facet_str, count)| (String::from(facet_str), count))
-                    .collect::<Vec<_>>()
+                .iter()
+                .map(|&(facet_str, count)| (String::from(facet_str), count))
+                .collect::<Vec<_>>()
             );
         }
     }
 
     #[test]
-    #[should_panic(
-        expected = "Tried to add a facet which is a descendant of \
-                    an already added facet."
-    )]
+    #[should_panic(expected = "Tried to add a facet which is a descendant of \
+                               an already added facet.")]
     fn test_misused_facet_collector() {
         let mut facet_collector = FacetCollector::for_field(Field(0));
         facet_collector.add_facet(Facet::from("/country"));
@@ -575,7 +520,7 @@ mod tests {
 
     #[test]
     fn test_doc_unsorted_multifacet() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let facet_field = schema_builder.add_facet_field("facets");
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -592,8 +537,7 @@ mod tests {
         assert_eq!(searcher.num_docs(), 1);
         let mut facet_collector = FacetCollector::for_field(facet_field);
         facet_collector.add_facet("/subjects");
-        searcher.search(&AllQuery, &mut facet_collector).unwrap();
-        let counts = facet_collector.harvest();
+        let counts = searcher.search(&AllQuery, &facet_collector).unwrap();
         let facets: Vec<(&Facet, u64)> = counts.get("/subjects").collect();
         assert_eq!(facets[0].1, 1);
     }
@@ -607,7 +551,7 @@ mod tests {
 
     #[test]
     fn test_facet_collector_topk() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let facet_field = schema_builder.add_facet_field("facet");
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -619,14 +563,16 @@ mod tests {
                 let facet = Facet::from(&format!("/facet/{}", c));
                 let doc = doc!(facet_field => facet);
                 iter::repeat(doc).take(count)
-            }).map(|mut doc| {
+            })
+            .map(|mut doc| {
                 doc.add_facet(
                     facet_field,
                     &format!("/facet/{}", thread_rng().sample(&uniform)),
                 );
                 doc
-            }).collect();
-        thread_rng().shuffle(&mut docs[..]);
+            })
+            .collect();
+        docs[..].shuffle(&mut thread_rng());
 
         let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
         for doc in docs {
@@ -639,9 +585,8 @@ mod tests {
 
         let mut facet_collector = FacetCollector::for_field(facet_field);
         facet_collector.add_facet("/facet");
-        searcher.search(&AllQuery, &mut facet_collector).unwrap();
+        let counts: FacetCounts = searcher.search(&AllQuery, &facet_collector).unwrap();
 
-        let counts: FacetCounts = facet_collector.harvest();
         {
             let facets: Vec<(&Facet, u64)> = counts.top_k("/facet", 3);
             assert_eq!(
@@ -664,13 +609,13 @@ mod bench {
     use query::AllQuery;
     use rand::{thread_rng, Rng};
     use schema::Facet;
-    use schema::SchemaBuilder;
+    use schema::Schema;
     use test::Bencher;
     use Index;
 
     #[bench]
     fn bench_facet_collector(b: &mut Bencher) {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let facet_field = schema_builder.add_facet_field("facet");
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -694,8 +639,8 @@ mod bench {
 
         b.iter(|| {
             let searcher = index.searcher();
-            let mut facet_collector = FacetCollector::for_field(facet_field);
-            searcher.search(&AllQuery, &mut facet_collector).unwrap();
+            let facet_collector = FacetCollector::for_field(facet_field);
+            searcher.search(&AllQuery, &facet_collector).unwrap();
         });
     }
 }

src/collector/int_facet_collector.rs

@@ -79,7 +79,7 @@ mod tests {
     // make sure we have facet counters correctly filled
     fn test_facet_collector_results() {
 
-        let mut schema_builder = schema::SchemaBuilder::new();
+        let mut schema_builder = schema::Schema::builder();
         let num_field_i64 = schema_builder.add_i64_field("num_i64", FAST);
         let num_field_u64 = schema_builder.add_u64_field("num_u64", FAST);
         let text_field = schema_builder.add_text_field("text", STRING);

src/collector/mod.rs

@@ -1,7 +1,91 @@
 /*!
-Defines how the documents matching a search query should be processed.
+
+# Collectors
+
+Collectors define the information you want to extract from the documents matching the queries.
+In tantivy jargon, we call this information your search "fruit".
+
+Your fruit could for instance be :
+- [the count of matching documents](./struct.Count.html)
+- [the top 10 documents, by relevancy or by a fast field](./struct.TopDocs.html)
+- [facet counts](./struct.FacetCollector.html)
+
+At one point in your code, you will trigger the actual search operation by calling
+[the `search(...)` method of your `Searcher` object](../struct.Searcher.html#method.search).
+This call will look like this.
+
+```verbatim
+let fruit = searcher.search(&query, &collector)?;
+```
+
+Here the type of fruit is actually determined as an associated type of the collector (`Collector::Fruit`).
+
+
+# Combining several collectors
+
+A rich search experience often requires to run several collectors on your search query.
+For instance,
+- selecting the top-K products matching your query
+- counting the matching documents
+- computing several facets
+- computing statistics about the matching product prices
+
+A simple and efficient way to do that is to pass your collectors as one tuple.
+The resulting `Fruit` will then be a typed tuple with each collector's original fruits
+in their respective position.
+
+```rust
+# extern crate tantivy;
+# use tantivy::schema::*;
+# use tantivy::*;
+# use tantivy::query::*;
+use tantivy::collector::{Count, TopDocs};
+#
+# fn main() -> tantivy::Result<()> {
+# let mut schema_builder = Schema::builder();
+#     let title = schema_builder.add_text_field("title", TEXT);
+#     let schema = schema_builder.build();
+#     let index = Index::create_in_ram(schema);
+#     let mut index_writer = index.writer(3_000_000)?;
+#       index_writer.add_document(doc!(
+#       title => "The Name of the Wind",
+#      ));
+#     index_writer.add_document(doc!(
+#        title => "The Diary of Muadib",
+#     ));
+#     index_writer.commit().unwrap();
+#     index.load_searchers()?;
+#     let searcher = index.searcher();
+#     let query_parser = QueryParser::for_index(&index, vec![title]);
+#     let query = query_parser.parse_query("diary")?;
+let (doc_count, top_docs): (usize, Vec<(Score, DocAddress)>) =
+    searcher.search(&query, &(Count, TopDocs::with_limit(2)))?;
+#     Ok(())
+# }
+```
+
+The `Collector` trait is implemented for up to 4 collectors.
+If you have more than 4 collectors, you can either group them into
+tuples of tuples `(a,(b,(c,d)))`, or rely on `MultiCollector`'s.
+
+# Combining several collectors dynamically
+
+Combining collectors into a tuple is a zero-cost abstraction: everything
+happens as if you had manually implemented a single collector
+combining all of our features.
+
+Unfortunately it requires you to know at compile time your collector types.
+If on the other hand, the collectors depend on some query parameter,
+you can rely on `MultiCollector`'s.
+
+
+# Implementing your own collectors.
+
+See the `custom_collector` example.
+
 */
 
+use downcast;
 use DocId;
 use Result;
 use Score;
@@ -9,7 +93,7 @@ use SegmentLocalId;
 use SegmentReader;
 
 mod count_collector;
-pub use self::count_collector::CountCollector;
+pub use self::count_collector::Count;
 
 mod multi_collector;
 pub use self::multi_collector::MultiCollector;
@@ -17,237 +101,267 @@ pub use self::multi_collector::MultiCollector;
 mod top_collector;
 
 mod top_score_collector;
-pub use self::top_score_collector::TopScoreCollector;
-#[deprecated]
-pub use self::top_score_collector::TopScoreCollector as TopCollector;
+pub use self::top_score_collector::TopDocs;
 
 mod top_field_collector;
-pub use self::top_field_collector::TopFieldCollector;
+pub use self::top_field_collector::TopDocsByField;
 
 mod facet_collector;
 pub use self::facet_collector::FacetCollector;
 
-mod chained_collector;
-pub use self::chained_collector::{chain, ChainedCollector};
+/// `Fruit` is the type for the result of our collection.
+/// e.g. `usize` for the `Count` collector.
+pub trait Fruit: Send + downcast::Any {}
+
+impl<T> Fruit for T where T: Send + downcast::Any {}
 
 /// Collectors are in charge of collecting and retaining relevant
 /// information from the document found and scored by the query.
 ///
-///
 /// For instance,
 ///
 /// - keeping track of the top 10 best documents
 /// - computing a breakdown over a fast field
 /// - computing the number of documents matching the query
 ///
-/// Queries are in charge of pushing the `DocSet` to the collector.
+/// Our search index is in fact a collection of segments, so
+/// a `Collector` trait is actually more of a factory to instance
+/// `SegmentCollector`s for each segments.
 ///
-/// As they work on multiple segments, they first inform
-/// the collector of a change in a segment and then
-/// call the `collect` method to push the document to the collector.
-///
-/// Temporally, our collector will receive calls
-/// - `.set_segment(0, segment_reader_0)`
-/// - `.collect(doc0_of_segment_0)`
-/// - `.collect(...)`
-/// - `.collect(last_doc_of_segment_0)`
-/// - `.set_segment(1, segment_reader_1)`
-/// - `.collect(doc0_of_segment_1)`
-/// - `.collect(...)`
-/// - `.collect(last_doc_of_segment_1)`
-/// - `...`
-/// - `.collect(last_doc_of_last_segment)`
+/// The collection logic itself is in the `SegmentCollector`.
 ///
 /// Segments are not guaranteed to be visited in any specific order.
-pub trait Collector {
+pub trait Collector: Sync {
+    /// `Fruit` is the type for the result of our collection.
+    /// e.g. `usize` for the `Count` collector.
+    type Fruit: Fruit;
+
+    /// Type of the `SegmentCollector` associated to this collector.
+    type Child: SegmentCollector<Fruit = Self::Fruit>;
+
     /// `set_segment` is called before beginning to enumerate
     /// on this segment.
-    fn set_segment(
-        &mut self,
+    fn for_segment(
+        &self,
         segment_local_id: SegmentLocalId,
         segment: &SegmentReader,
-    ) -> Result<()>;
-    /// The query pushes the scored document to the collector via this method.
-    fn collect(&mut self, doc: DocId, score: Score);
+    ) -> Result<Self::Child>;
 
     /// Returns true iff the collector requires to compute scores for documents.
     fn requires_scoring(&self) -> bool;
+
+    /// Combines the fruit associated to the collection of each segments
+    /// into one fruit.
+    fn merge_fruits(&self, segment_fruits: Vec<Self::Fruit>) -> Result<Self::Fruit>;
 }
 
-impl<'a, C: Collector> Collector for &'a mut C {
-    fn set_segment(
-        &mut self,
-        segment_local_id: SegmentLocalId,
-        segment: &SegmentReader,
-    ) -> Result<()> {
-        (*self).set_segment(segment_local_id, segment)
-    }
+/// The `SegmentCollector` is the trait in charge of defining the
+/// collect operation at the scale of the segment.
+///
+/// `.collect(doc, score)` will be called for every documents
+/// matching the query.
+pub trait SegmentCollector: 'static {
+    /// `Fruit` is the type for the result of our collection.
+    /// e.g. `usize` for the `Count` collector.
+    type Fruit: Fruit;
+
     /// The query pushes the scored document to the collector via this method.
-    fn collect(&mut self, doc: DocId, score: Score) {
-        C::collect(self, doc, score)
+    fn collect(&mut self, doc: DocId, score: Score);
+
+    /// Extract the fruit of the collection from the `SegmentCollector`.
+    fn harvest(self) -> Self::Fruit;
+}
+
+// -----------------------------------------------
+// Tuple implementations.
+
+impl<Left, Right> Collector for (Left, Right)
+where
+    Left: Collector,
+    Right: Collector,
+{
+    type Fruit = (Left::Fruit, Right::Fruit);
+    type Child = (Left::Child, Right::Child);
+
+    fn for_segment(&self, segment_local_id: u32, segment: &SegmentReader) -> Result<Self::Child> {
+        let left = self.0.for_segment(segment_local_id, segment)?;
+        let right = self.1.for_segment(segment_local_id, segment)?;
+        Ok((left, right))
     }
 
     fn requires_scoring(&self) -> bool {
-        C::requires_scoring(self)
+        self.0.requires_scoring() || self.1.requires_scoring()
     }
+
+    fn merge_fruits(
+        &self,
+        children: Vec<(Left::Fruit, Right::Fruit)>,
+    ) -> Result<(Left::Fruit, Right::Fruit)> {
+        let mut left_fruits = vec![];
+        let mut right_fruits = vec![];
+        for (left_fruit, right_fruit) in children {
+            left_fruits.push(left_fruit);
+            right_fruits.push(right_fruit);
+        }
+        Ok((
+            self.0.merge_fruits(left_fruits)?,
+            self.1.merge_fruits(right_fruits)?,
+        ))
+    }
+}
+
+impl<Left, Right> SegmentCollector for (Left, Right)
+where
+    Left: SegmentCollector,
+    Right: SegmentCollector,
+{
+    type Fruit = (Left::Fruit, Right::Fruit);
+
+    fn collect(&mut self, doc: DocId, score: Score) {
+        self.0.collect(doc, score);
+        self.1.collect(doc, score);
+    }
+
+    fn harvest(self) -> <Self as SegmentCollector>::Fruit {
+        (self.0.harvest(), self.1.harvest())
+    }
+}
+
+// 3-Tuple
+
+impl<One, Two, Three> Collector for (One, Two, Three)
+where
+    One: Collector,
+    Two: Collector,
+    Three: Collector,
+{
+    type Fruit = (One::Fruit, Two::Fruit, Three::Fruit);
+    type Child = (One::Child, Two::Child, Three::Child);
+
+    fn for_segment(&self, segment_local_id: u32, segment: &SegmentReader) -> Result<Self::Child> {
+        let one = self.0.for_segment(segment_local_id, segment)?;
+        let two = self.1.for_segment(segment_local_id, segment)?;
+        let three = self.2.for_segment(segment_local_id, segment)?;
+        Ok((one, two, three))
+    }
+
+    fn requires_scoring(&self) -> bool {
+        self.0.requires_scoring() || self.1.requires_scoring() || self.2.requires_scoring()
+    }
+
+    fn merge_fruits(&self, children: Vec<Self::Fruit>) -> Result<Self::Fruit> {
+        let mut one_fruits = vec![];
+        let mut two_fruits = vec![];
+        let mut three_fruits = vec![];
+        for (one_fruit, two_fruit, three_fruit) in children {
+            one_fruits.push(one_fruit);
+            two_fruits.push(two_fruit);
+            three_fruits.push(three_fruit);
+        }
+        Ok((
+            self.0.merge_fruits(one_fruits)?,
+            self.1.merge_fruits(two_fruits)?,
+            self.2.merge_fruits(three_fruits)?,
+        ))
+    }
+}
+
+impl<One, Two, Three> SegmentCollector for (One, Two, Three)
+where
+    One: SegmentCollector,
+    Two: SegmentCollector,
+    Three: SegmentCollector,
+{
+    type Fruit = (One::Fruit, Two::Fruit, Three::Fruit);
+
+    fn collect(&mut self, doc: DocId, score: Score) {
+        self.0.collect(doc, score);
+        self.1.collect(doc, score);
+        self.2.collect(doc, score);
+    }
+
+    fn harvest(self) -> <Self as SegmentCollector>::Fruit {
+        (self.0.harvest(), self.1.harvest(), self.2.harvest())
+    }
+}
+
+// 4-Tuple
+
+impl<One, Two, Three, Four> Collector for (One, Two, Three, Four)
+where
+    One: Collector,
+    Two: Collector,
+    Three: Collector,
+    Four: Collector,
+{
+    type Fruit = (One::Fruit, Two::Fruit, Three::Fruit, Four::Fruit);
+    type Child = (One::Child, Two::Child, Three::Child, Four::Child);
+
+    fn for_segment(&self, segment_local_id: u32, segment: &SegmentReader) -> Result<Self::Child> {
+        let one = self.0.for_segment(segment_local_id, segment)?;
+        let two = self.1.for_segment(segment_local_id, segment)?;
+        let three = self.2.for_segment(segment_local_id, segment)?;
+        let four = self.3.for_segment(segment_local_id, segment)?;
+        Ok((one, two, three, four))
+    }
+
+    fn requires_scoring(&self) -> bool {
+        self.0.requires_scoring()
+            || self.1.requires_scoring()
+            || self.2.requires_scoring()
+            || self.3.requires_scoring()
+    }
+
+    fn merge_fruits(&self, children: Vec<Self::Fruit>) -> Result<Self::Fruit> {
+        let mut one_fruits = vec![];
+        let mut two_fruits = vec![];
+        let mut three_fruits = vec![];
+        let mut four_fruits = vec![];
+        for (one_fruit, two_fruit, three_fruit, four_fruit) in children {
+            one_fruits.push(one_fruit);
+            two_fruits.push(two_fruit);
+            three_fruits.push(three_fruit);
+            four_fruits.push(four_fruit);
+        }
+        Ok((
+            self.0.merge_fruits(one_fruits)?,
+            self.1.merge_fruits(two_fruits)?,
+            self.2.merge_fruits(three_fruits)?,
+            self.3.merge_fruits(four_fruits)?,
+        ))
+    }
+}
+
+impl<One, Two, Three, Four> SegmentCollector for (One, Two, Three, Four)
+where
+    One: SegmentCollector,
+    Two: SegmentCollector,
+    Three: SegmentCollector,
+    Four: SegmentCollector,
+{
+    type Fruit = (One::Fruit, Two::Fruit, Three::Fruit, Four::Fruit);
+
+    fn collect(&mut self, doc: DocId, score: Score) {
+        self.0.collect(doc, score);
+        self.1.collect(doc, score);
+        self.2.collect(doc, score);
+        self.3.collect(doc, score);
+    }
+
+    fn harvest(self) -> <Self as SegmentCollector>::Fruit {
+        (
+            self.0.harvest(),
+            self.1.harvest(),
+            self.2.harvest(),
+            self.3.harvest(),
+        )
+    }
+}
+
+#[allow(missing_docs)]
+mod downcast_impl {
+    downcast!(super::Fruit);
 }
 
 #[cfg(test)]
-pub mod tests {
-
-    use super::*;
-    use core::SegmentReader;
-    use fastfield::BytesFastFieldReader;
-    use fastfield::FastFieldReader;
-    use schema::Field;
-    use DocId;
-    use Score;
-    use SegmentLocalId;
-
-    /// Stores all of the doc ids.
-    /// This collector is only used for tests.
-    /// It is unusable in practise, as it does not store
-    /// the segment ordinals
-    pub struct TestCollector {
-        offset: DocId,
-        segment_max_doc: DocId,
-        docs: Vec<DocId>,
-        scores: Vec<Score>,
-    }
-
-    impl TestCollector {
-        /// Return the exhalist of documents.
-        pub fn docs(self) -> Vec<DocId> {
-            self.docs
-        }
-
-        pub fn scores(self) -> Vec<Score> {
-            self.scores
-        }
-    }
-
-    impl Default for TestCollector {
-        fn default() -> TestCollector {
-            TestCollector {
-                offset: 0,
-                segment_max_doc: 0,
-                docs: Vec::new(),
-                scores: Vec::new(),
-            }
-        }
-    }
-
-    impl Collector for TestCollector {
-        fn set_segment(&mut self, _: SegmentLocalId, reader: &SegmentReader) -> Result<()> {
-            self.offset += self.segment_max_doc;
-            self.segment_max_doc = reader.max_doc();
-            Ok(())
-        }
-
-        fn collect(&mut self, doc: DocId, score: Score) {
-            self.docs.push(doc + self.offset);
-            self.scores.push(score);
-        }
-
-        fn requires_scoring(&self) -> bool {
-            true
-        }
-    }
-
-    /// Collects in order all of the fast fields for all of the
-    /// doc in the `DocSet`
-    ///
-    /// This collector is mainly useful for tests.
-    pub struct FastFieldTestCollector {
-        vals: Vec<u64>,
-        field: Field,
-        ff_reader: Option<FastFieldReader<u64>>,
-    }
-
-    impl FastFieldTestCollector {
-        pub fn for_field(field: Field) -> FastFieldTestCollector {
-            FastFieldTestCollector {
-                vals: Vec::new(),
-                field,
-                ff_reader: None,
-            }
-        }
-
-        pub fn vals(self) -> Vec<u64> {
-            self.vals
-        }
-    }
-
-    impl Collector for FastFieldTestCollector {
-        fn set_segment(&mut self, _: SegmentLocalId, reader: &SegmentReader) -> Result<()> {
-            self.ff_reader = Some(reader.fast_field_reader(self.field)?);
-            Ok(())
-        }
-
-        fn collect(&mut self, doc: DocId, _score: Score) {
-            let val = self.ff_reader.as_ref().unwrap().get(doc);
-            self.vals.push(val);
-        }
-        fn requires_scoring(&self) -> bool {
-            false
-        }
-    }
-
-    /// Collects in order all of the fast field bytes for all of the
-    /// docs in the `DocSet`
-    ///
-    /// This collector is mainly useful for tests.
-    pub struct BytesFastFieldTestCollector {
-        vals: Vec<u8>,
-        field: Field,
-        ff_reader: Option<BytesFastFieldReader>,
-    }
-
-    impl BytesFastFieldTestCollector {
-        pub fn for_field(field: Field) -> BytesFastFieldTestCollector {
-            BytesFastFieldTestCollector {
-                vals: Vec::new(),
-                field,
-                ff_reader: None,
-            }
-        }
-
-        pub fn vals(self) -> Vec<u8> {
-            self.vals
-        }
-    }
-
-    impl Collector for BytesFastFieldTestCollector {
-        fn set_segment(&mut self, _segment_local_id: u32, segment: &SegmentReader) -> Result<()> {
-            self.ff_reader = Some(segment.bytes_fast_field_reader(self.field)?);
-            Ok(())
-        }
-
-        fn collect(&mut self, doc: u32, _score: f32) {
-            let val = self.ff_reader.as_ref().unwrap().get_val(doc);
-            self.vals.extend(val);
-        }
-
-        fn requires_scoring(&self) -> bool {
-            false
-        }
-    }
-}
-
-#[cfg(all(test, feature = "unstable"))]
-mod bench {
-    use collector::{Collector, CountCollector};
-    use test::Bencher;
-
-    #[bench]
-    fn build_collector(b: &mut Bencher) {
-        b.iter(|| {
-            let mut count_collector = CountCollector::default();
-            let docs: Vec<u32> = (0..1_000_000).collect();
-            for doc in docs {
-                count_collector.collect(doc, 1f32);
-            }
-            count_collector.count()
-        });
-    }
-}
+pub mod tests;

src/collector/multi_collector.rs

@@ -1,9 +1,97 @@
 use super::Collector;
+use super::SegmentCollector;
+use collector::Fruit;
+use downcast::Downcast;
+use std::marker::PhantomData;
 use DocId;
 use Result;
 use Score;
 use SegmentLocalId;
 use SegmentReader;
+use TantivyError;
+
+pub struct MultiFruit {
+    sub_fruits: Vec<Option<Box<Fruit>>>,
+}
+
+pub struct CollectorWrapper<TCollector: Collector>(TCollector);
+
+impl<TCollector: Collector> Collector for CollectorWrapper<TCollector> {
+    type Fruit = Box<Fruit>;
+    type Child = Box<BoxableSegmentCollector>;
+
+    fn for_segment(
+        &self,
+        segment_local_id: u32,
+        reader: &SegmentReader,
+    ) -> Result<Box<BoxableSegmentCollector>> {
+        let child = self.0.for_segment(segment_local_id, reader)?;
+        Ok(Box::new(SegmentCollectorWrapper(child)))
+    }
+
+    fn requires_scoring(&self) -> bool {
+        self.0.requires_scoring()
+    }
+
+    fn merge_fruits(&self, children: Vec<<Self as Collector>::Fruit>) -> Result<Box<Fruit>> {
+        let typed_fruit: Vec<TCollector::Fruit> = children
+            .into_iter()
+            .map(|untyped_fruit| {
+                Downcast::<TCollector::Fruit>::downcast(untyped_fruit)
+                    .map(|boxed_but_typed| *boxed_but_typed)
+                    .map_err(|e| {
+                        let err_msg = format!("Failed to cast child collector fruit. {:?}", e);
+                        TantivyError::InvalidArgument(err_msg)
+                    })
+            })
+            .collect::<Result<_>>()?;
+        let merged_fruit = self.0.merge_fruits(typed_fruit)?;
+        Ok(Box::new(merged_fruit))
+    }
+}
+
+impl SegmentCollector for Box<BoxableSegmentCollector> {
+    type Fruit = Box<Fruit>;
+
+    fn collect(&mut self, doc: u32, score: f32) {
+        self.as_mut().collect(doc, score);
+    }
+
+    fn harvest(self) -> Box<Fruit> {
+        BoxableSegmentCollector::harvest_from_box(self)
+    }
+}
+
+pub trait BoxableSegmentCollector {
+    fn collect(&mut self, doc: u32, score: f32);
+    fn harvest_from_box(self: Box<Self>) -> Box<Fruit>;
+}
+
+pub struct SegmentCollectorWrapper<TSegmentCollector: SegmentCollector>(TSegmentCollector);
+
+impl<TSegmentCollector: SegmentCollector> BoxableSegmentCollector
+    for SegmentCollectorWrapper<TSegmentCollector>
+{
+    fn collect(&mut self, doc: u32, score: f32) {
+        self.0.collect(doc, score);
+    }
+
+    fn harvest_from_box(self: Box<Self>) -> Box<Fruit> {
+        Box::new(self.0.harvest())
+    }
+}
+
+pub struct FruitHandle<TFruit: Fruit> {
+    pos: usize,
+    _phantom: PhantomData<TFruit>,
+}
+
+impl<TFruit: Fruit> FruitHandle<TFruit> {
+    pub fn extract(self, fruits: &mut MultiFruit) -> TFruit {
+        let boxed_fruit = fruits.sub_fruits[self.pos].take().expect("");
+        *Downcast::<TFruit>::downcast(boxed_fruit).expect("Failed")
+    }
+}
 
 /// Multicollector makes it possible to collect on more than one collector.
 /// It should only be used for use cases where the Collector types is unknown
@@ -13,14 +101,14 @@ use SegmentReader;
 /// ```rust
 /// #[macro_use]
 /// extern crate tantivy;
-/// use tantivy::schema::{SchemaBuilder, TEXT};
+/// use tantivy::schema::{Schema, TEXT};
 /// use tantivy::{Index, Result};
-/// use tantivy::collector::{CountCollector, TopCollector, MultiCollector};
+/// use tantivy::collector::{Count, TopDocs, MultiCollector};
 /// use tantivy::query::QueryParser;
 ///
 /// # fn main() { example().unwrap(); }
 /// fn example() -> Result<()> {
-///     let mut schema_builder = SchemaBuilder::new();
+///     let mut schema_builder = Schema::builder();
 ///     let title = schema_builder.add_text_field("title", TEXT);
 ///     let schema = schema_builder.build();
 ///     let index = Index::create_in_ram(schema);
@@ -44,55 +132,115 @@ use SegmentReader;
 ///     index.load_searchers()?;
 ///     let searcher = index.searcher();
 ///
-///     {
-///         let mut top_collector = TopCollector::with_limit(2);
-///         let mut count_collector = CountCollector::default();
-///         {
-///             let mut collectors =
-///                 MultiCollector::from(vec![&mut top_collector, &mut count_collector]);
-///             let query_parser = QueryParser::for_index(&index, vec![title]);
-///             let query = query_parser.parse_query("diary")?;
-///             searcher.search(&*query, &mut collectors).unwrap();
-///         }
-///         assert_eq!(count_collector.count(), 2);
-///         assert!(top_collector.at_capacity());
-///     }
+///     let mut collectors = MultiCollector::new();
+///     let top_docs_handle = collectors.add_collector(TopDocs::with_limit(2));
+///     let count_handle = collectors.add_collector(Count);
+///     let query_parser = QueryParser::for_index(&index, vec![title]);
+///     let query = query_parser.parse_query("diary")?;
+///     let mut multi_fruit = searcher.search(&query, &collectors)?;
+///
+///     let count = count_handle.extract(&mut multi_fruit);
+///     let top_docs = top_docs_handle.extract(&mut multi_fruit);
+///
+///     # assert_eq!(count, 2);
+///     # assert_eq!(top_docs.len(), 2);
 ///
 ///     Ok(())
 /// }
 /// ```
+#[allow(clippy::type_complexity)]
+#[derive(Default)]
 pub struct MultiCollector<'a> {
-    collectors: Vec<&'a mut Collector>,
+    collector_wrappers:
+        Vec<Box<Collector<Child = Box<BoxableSegmentCollector>, Fruit = Box<Fruit>> + 'a>>,
 }
 
 impl<'a> MultiCollector<'a> {
-    /// Constructor
-    pub fn from(collectors: Vec<&'a mut Collector>) -> MultiCollector {
-        MultiCollector { collectors }
+    /// Create a new `MultiCollector`
+    pub fn new() -> Self {
+        Default::default()
+    }
+
+    /// Add a new collector to our `MultiCollector`.
+    pub fn add_collector<'b: 'a, TCollector: Collector + 'b>(
+        &mut self,
+        collector: TCollector,
+    ) -> FruitHandle<TCollector::Fruit> {
+        let pos = self.collector_wrappers.len();
+        self.collector_wrappers
+            .push(Box::new(CollectorWrapper(collector)));
+        FruitHandle {
+            pos,
+            _phantom: PhantomData,
+        }
     }
 }
 
 impl<'a> Collector for MultiCollector<'a> {
-    fn set_segment(
-        &mut self,
+    type Fruit = MultiFruit;
+    type Child = MultiCollectorChild;
+
+    fn for_segment(
+        &self,
         segment_local_id: SegmentLocalId,
         segment: &SegmentReader,
-    ) -> Result<()> {
-        for collector in &mut self.collectors {
-            collector.set_segment(segment_local_id, segment)?;
-        }
-        Ok(())
+    ) -> Result<MultiCollectorChild> {
+        let children = self
+            .collector_wrappers
+            .iter()
+            .map(|collector_wrapper| collector_wrapper.for_segment(segment_local_id, segment))
+            .collect::<Result<Vec<_>>>()?;
+        Ok(MultiCollectorChild { children })
     }
 
+    fn requires_scoring(&self) -> bool {
+        self.collector_wrappers.iter().any(|c| c.requires_scoring())
+    }
+
+    fn merge_fruits(&self, segments_multifruits: Vec<MultiFruit>) -> Result<MultiFruit> {
+        let mut segment_fruits_list: Vec<Vec<Box<Fruit>>> = (0..self.collector_wrappers.len())
+            .map(|_| Vec::with_capacity(segments_multifruits.len()))
+            .collect::<Vec<_>>();
+        for segment_multifruit in segments_multifruits {
+            for (idx, segment_fruit_opt) in segment_multifruit.sub_fruits.into_iter().enumerate() {
+                if let Some(segment_fruit) = segment_fruit_opt {
+                    segment_fruits_list[idx].push(segment_fruit);
+                }
+            }
+        }
+        let sub_fruits = self
+            .collector_wrappers
+            .iter()
+            .zip(segment_fruits_list)
+            .map(|(child_collector, segment_fruits)| {
+                Ok(Some(child_collector.merge_fruits(segment_fruits)?))
+            })
+            .collect::<Result<_>>()?;
+        Ok(MultiFruit { sub_fruits })
+    }
+}
+
+pub struct MultiCollectorChild {
+    children: Vec<Box<BoxableSegmentCollector>>,
+}
+
+impl SegmentCollector for MultiCollectorChild {
+    type Fruit = MultiFruit;
+
     fn collect(&mut self, doc: DocId, score: Score) {
-        for collector in &mut self.collectors {
-            collector.collect(doc, score);
+        for child in &mut self.children {
+            child.collect(doc, score);
         }
     }
-    fn requires_scoring(&self) -> bool {
-        self.collectors
-            .iter()
-            .any(|collector| collector.requires_scoring())
+
+    fn harvest(self) -> MultiFruit {
+        MultiFruit {
+            sub_fruits: self
+                .children
+                .into_iter()
+                .map(|child| Some(child.harvest()))
+                .collect(),
+        }
     }
 }
 
@@ -100,20 +248,42 @@ impl<'a> Collector for MultiCollector<'a> {
 mod tests {
 
     use super::*;
-    use collector::{Collector, CountCollector, TopScoreCollector};
+    use collector::{Count, TopDocs};
+    use query::TermQuery;
+    use schema::IndexRecordOption;
+    use schema::{Schema, TEXT};
+    use Index;
+    use Term;
 
     #[test]
     fn test_multi_collector() {
-        let mut top_collector = TopScoreCollector::with_limit(2);
-        let mut count_collector = CountCollector::default();
+        let mut schema_builder = Schema::builder();
+        let text = schema_builder.add_text_field("text", TEXT);
+        let schema = schema_builder.build();
+
+        let index = Index::create_in_ram(schema);
         {
-            let mut collectors =
-                MultiCollector::from(vec![&mut top_collector, &mut count_collector]);
-            collectors.collect(1, 0.2);
-            collectors.collect(2, 0.1);
-            collectors.collect(3, 0.5);
+            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            index_writer.add_document(doc!(text=>"abc"));
+            index_writer.add_document(doc!(text=>"abc abc abc"));
+            index_writer.add_document(doc!(text=>"abc abc"));
+            index_writer.commit().unwrap();
+            index_writer.add_document(doc!(text=>""));
+            index_writer.add_document(doc!(text=>"abc abc abc abc"));
+            index_writer.add_document(doc!(text=>"abc"));
+            index_writer.commit().unwrap();
         }
-        assert_eq!(count_collector.count(), 3);
-        assert!(top_collector.at_capacity());
+        index.load_searchers().unwrap();
+        let searcher = index.searcher();
+        let term = Term::from_field_text(text, "abc");
+        let query = TermQuery::new(term, IndexRecordOption::Basic);
+
+        let mut collectors = MultiCollector::new();
+        let topdocs_handler = collectors.add_collector(TopDocs::with_limit(2));
+        let count_handler = collectors.add_collector(Count);
+        let mut multifruits = searcher.search(&query, &mut collectors).unwrap();
+
+        assert_eq!(count_handler.extract(&mut multifruits), 5);
+        assert_eq!(topdocs_handler.extract(&mut multifruits).len(), 2);
     }
 }

src/collector/tests.rs

@@ -0,0 +1,201 @@
+use super::*;
+use core::SegmentReader;
+use fastfield::BytesFastFieldReader;
+use fastfield::FastFieldReader;
+use schema::Field;
+use DocAddress;
+use DocId;
+use Score;
+use SegmentLocalId;
+
+/// Stores all of the doc ids.
+/// This collector is only used for tests.
+/// It is unusable in pr
+///
+/// actise, as it does not store
+/// the segment ordinals
+pub struct TestCollector;
+
+pub struct TestSegmentCollector {
+    segment_id: SegmentLocalId,
+    fruit: TestFruit,
+}
+
+#[derive(Default)]
+pub struct TestFruit {
+    docs: Vec<DocAddress>,
+    scores: Vec<Score>,
+}
+
+impl TestFruit {
+    /// Return the list of matching documents exhaustively.
+    pub fn docs(&self) -> &[DocAddress] {
+        &self.docs[..]
+    }
+
+    pub fn scores(&self) -> &[Score] {
+        &self.scores[..]
+    }
+}
+
+impl Collector for TestCollector {
+    type Fruit = TestFruit;
+    type Child = TestSegmentCollector;
+
+    fn for_segment(
+        &self,
+        segment_id: SegmentLocalId,
+        _reader: &SegmentReader,
+    ) -> Result<TestSegmentCollector> {
+        Ok(TestSegmentCollector {
+            segment_id,
+            fruit: TestFruit::default(),
+        })
+    }
+
+    fn requires_scoring(&self) -> bool {
+        true
+    }
+
+    fn merge_fruits(&self, mut children: Vec<TestFruit>) -> Result<TestFruit> {
+        children.sort_by_key(|fruit| {
+            if fruit.docs().is_empty() {
+                0
+            } else {
+                fruit.docs()[0].segment_ord()
+            }
+        });
+        let mut docs = vec![];
+        let mut scores = vec![];
+        for child in children {
+            docs.extend(child.docs());
+            scores.extend(child.scores);
+        }
+        Ok(TestFruit { docs, scores })
+    }
+}
+
+impl SegmentCollector for TestSegmentCollector {
+    type Fruit = TestFruit;
+
+    fn collect(&mut self, doc: DocId, score: Score) {
+        self.fruit.docs.push(DocAddress(self.segment_id, doc));
+        self.fruit.scores.push(score);
+    }
+
+    fn harvest(self) -> <Self as SegmentCollector>::Fruit {
+        self.fruit
+    }
+}
+
+/// Collects in order all of the fast fields for all of the
+/// doc in the `DocSet`
+///
+/// This collector is mainly useful for tests.
+pub struct FastFieldTestCollector {
+    field: Field,
+}
+
+pub struct FastFieldSegmentCollector {
+    vals: Vec<u64>,
+    reader: FastFieldReader<u64>,
+}
+
+impl FastFieldTestCollector {
+    pub fn for_field(field: Field) -> FastFieldTestCollector {
+        FastFieldTestCollector { field }
+    }
+}
+
+impl Collector for FastFieldTestCollector {
+    type Fruit = Vec<u64>;
+    type Child = FastFieldSegmentCollector;
+
+    fn for_segment(
+        &self,
+        _: SegmentLocalId,
+        reader: &SegmentReader,
+    ) -> Result<FastFieldSegmentCollector> {
+        Ok(FastFieldSegmentCollector {
+            vals: Vec::new(),
+            reader: reader.fast_field_reader(self.field)?,
+        })
+    }
+
+    fn requires_scoring(&self) -> bool {
+        false
+    }
+
+    fn merge_fruits(&self, children: Vec<Vec<u64>>) -> Result<Vec<u64>> {
+        Ok(children.into_iter().flat_map(|v| v.into_iter()).collect())
+    }
+}
+
+impl SegmentCollector for FastFieldSegmentCollector {
+    type Fruit = Vec<u64>;
+
+    fn collect(&mut self, doc: DocId, _score: Score) {
+        let val = self.reader.get(doc);
+        self.vals.push(val);
+    }
+
+    fn harvest(self) -> Vec<u64> {
+        self.vals
+    }
+}
+
+/// Collects in order all of the fast field bytes for all of the
+/// docs in the `DocSet`
+///
+/// This collector is mainly useful for tests.
+pub struct BytesFastFieldTestCollector {
+    field: Field,
+}
+
+pub struct BytesFastFieldSegmentCollector {
+    vals: Vec<u8>,
+    reader: BytesFastFieldReader,
+}
+
+impl BytesFastFieldTestCollector {
+    pub fn for_field(field: Field) -> BytesFastFieldTestCollector {
+        BytesFastFieldTestCollector { field }
+    }
+}
+
+impl Collector for BytesFastFieldTestCollector {
+    type Fruit = Vec<u8>;
+    type Child = BytesFastFieldSegmentCollector;
+
+    fn for_segment(
+        &self,
+        _segment_local_id: u32,
+        segment: &SegmentReader,
+    ) -> Result<BytesFastFieldSegmentCollector> {
+        Ok(BytesFastFieldSegmentCollector {
+            vals: Vec::new(),
+            reader: segment.bytes_fast_field_reader(self.field)?,
+        })
+    }
+
+    fn requires_scoring(&self) -> bool {
+        false
+    }
+
+    fn merge_fruits(&self, children: Vec<Vec<u8>>) -> Result<Vec<u8>> {
+        Ok(children.into_iter().flat_map(|c| c.into_iter()).collect())
+    }
+}
+
+impl SegmentCollector for BytesFastFieldSegmentCollector {
+    type Fruit = Vec<u8>;
+
+    fn collect(&mut self, doc: u32, _score: f32) {
+        let data = self.reader.get_val(doc);
+        self.vals.extend(data);
+    }
+
+    fn harvest(self) -> <Self as SegmentCollector>::Fruit {
+        self.vals
+    }
+}

src/collector/top_collector.rs

@@ -1,56 +1,59 @@
+use serde::export::PhantomData;
 use std::cmp::Ordering;
 use std::collections::BinaryHeap;
 use DocAddress;
 use DocId;
+use Result;
 use SegmentLocalId;
+use SegmentReader;
 
 /// Contains a feature (field, score, etc.) of a document along with the document address.
 ///
 /// It has a custom implementation of `PartialOrd` that reverses the order. This is because the
 /// default Rust heap is a max heap, whereas a min heap is needed.
-#[derive(Clone, Copy)]
-pub struct ComparableDoc<T> {
+///
+/// WARNING: equality is not what you would expect here.
+/// Two elements are equal if their feature is equal, and regardless of whether `doc`
+/// is equal. This should be perfectly fine for this usage, but let's make sure this
+/// struct is never public.
+struct ComparableDoc<T, D> {
     feature: T,
-    doc_address: DocAddress,
+    doc: D,
 }
 
-impl<T: PartialOrd> PartialOrd for ComparableDoc<T> {
+impl<T: PartialOrd, D> PartialOrd for ComparableDoc<T, D> {
     fn partial_cmp(&self, other: &Self) -> Option<Ordering> {
         Some(self.cmp(other))
     }
 }
 
-impl<T: PartialOrd> Ord for ComparableDoc<T> {
+impl<T: PartialOrd, D> Ord for ComparableDoc<T, D> {
     #[inline]
     fn cmp(&self, other: &Self) -> Ordering {
         other
             .feature
             .partial_cmp(&self.feature)
-            .unwrap_or_else(|| other.doc_address.cmp(&self.doc_address))
+            .unwrap_or_else(|| Ordering::Equal)
     }
 }
 
-impl<T: PartialOrd> PartialEq for ComparableDoc<T> {
+impl<T: PartialOrd, D> PartialEq for ComparableDoc<T, D> {
     fn eq(&self, other: &Self) -> bool {
         self.cmp(other) == Ordering::Equal
     }
 }
 
-impl<T: PartialOrd> Eq for ComparableDoc<T> {}
+impl<T: PartialOrd, D> Eq for ComparableDoc<T, D> {}
 
-/// The Top Collector keeps track of the K documents
-/// sorted by type `T`.
-///
-/// The implementation is based on a `BinaryHeap`.
-/// The theorical complexity for collecting the top `K` out of `n` documents
-/// is `O(n log K)`.
-pub struct TopCollector<T> {
+pub(crate) struct TopCollector<T> {
     limit: usize,
-    heap: BinaryHeap<ComparableDoc<T>>,
-    segment_id: u32,
+    _marker: PhantomData<T>,
 }
 
-impl<T: PartialOrd + Clone> TopCollector<T> {
+impl<T> TopCollector<T>
+where
+    T: PartialOrd + Clone,
+{
     /// Creates a top collector, with a number of documents equal to "limit".
     ///
     /// # Panics
@@ -61,127 +64,156 @@ impl<T: PartialOrd + Clone> TopCollector<T> {
         }
         TopCollector {
             limit,
-            heap: BinaryHeap::with_capacity(limit),
-            segment_id: 0,
+            _marker: PhantomData,
         }
     }
 
-    /// Returns K best documents sorted in decreasing order.
-    ///
-    /// Calling this method triggers the sort.
-    /// The result of the sort is not cached.
-    pub fn docs(&self) -> Vec<DocAddress> {
-        self.top_docs()
-            .into_iter()
-            .map(|(_feature, doc)| doc)
-            .collect()
+    pub fn limit(&self) -> usize {
+        self.limit
     }
 
-    /// Returns K best FeatureDocuments sorted in decreasing order.
-    ///
-    /// Calling this method triggers the sort.
-    /// The result of the sort is not cached.
-    pub fn top_docs(&self) -> Vec<(T, DocAddress)> {
-        let mut feature_docs: Vec<ComparableDoc<T>> = self.heap.iter().cloned().collect();
-        feature_docs.sort();
-        feature_docs
+    pub fn merge_fruits(
+        &self,
+        children: Vec<Vec<(T, DocAddress)>>,
+    ) -> Result<Vec<(T, DocAddress)>> {
+        if self.limit == 0 {
+            return Ok(Vec::new());
+        }
+        let mut top_collector = BinaryHeap::new();
+        for child_fruit in children {
+            for (feature, doc) in child_fruit {
+                if top_collector.len() < self.limit {
+                    top_collector.push(ComparableDoc { feature, doc });
+                } else if let Some(mut head) = top_collector.peek_mut() {
+                    if head.feature < feature {
+                        *head = ComparableDoc { feature, doc };
+                    }
+                }
+            }
+        }
+        Ok(top_collector
+            .into_sorted_vec()
             .into_iter()
-            .map(
-                |ComparableDoc {
-                     feature,
-                     doc_address,
-                 }| (feature, doc_address),
-            ).collect()
+            .map(|cdoc| (cdoc.feature, cdoc.doc))
+            .collect())
+    }
+
+    pub(crate) fn for_segment(
+        &self,
+        segment_id: SegmentLocalId,
+        _: &SegmentReader,
+    ) -> Result<TopSegmentCollector<T>> {
+        Ok(TopSegmentCollector::new(segment_id, self.limit))
+    }
+}
+
+/// The Top Collector keeps track of the K documents
+/// sorted by type `T`.
+///
+/// The implementation is based on a `BinaryHeap`.
+/// The theorical complexity for collecting the top `K` out of `n` documents
+/// is `O(n log K)`.
+pub(crate) struct TopSegmentCollector<T> {
+    limit: usize,
+    heap: BinaryHeap<ComparableDoc<T, DocId>>,
+    segment_id: u32,
+}
+
+impl<T: PartialOrd> TopSegmentCollector<T> {
+    fn new(segment_id: SegmentLocalId, limit: usize) -> TopSegmentCollector<T> {
+        TopSegmentCollector {
+            limit,
+            heap: BinaryHeap::with_capacity(limit),
+            segment_id,
+        }
+    }
+}
+
+impl<T: PartialOrd + Clone> TopSegmentCollector<T> {
+    pub fn harvest(self) -> Vec<(T, DocAddress)> {
+        let segment_id = self.segment_id;
+        self.heap
+            .into_sorted_vec()
+            .into_iter()
+            .map(|comparable_doc| {
+                (
+                    comparable_doc.feature,
+                    DocAddress(segment_id, comparable_doc.doc),
+                )
+            })
+            .collect()
     }
 
     /// Return true iff at least K documents have gone through
     /// the collector.
-    #[inline]
-    pub fn at_capacity(&self) -> bool {
+    #[inline(always)]
+    pub(crate) fn at_capacity(&self) -> bool {
         self.heap.len() >= self.limit
     }
 
-    /// Sets the segment local ID for the collector
-    pub fn set_segment_id(&mut self, segment_id: SegmentLocalId) {
-        self.segment_id = segment_id;
-    }
-
     /// Collects a document scored by the given feature
     ///
     /// It collects documents until it has reached the max capacity. Once it reaches capacity, it
     /// will compare the lowest scoring item with the given one and keep whichever is greater.
+    #[inline(always)]
     pub fn collect(&mut self, doc: DocId, feature: T) {
         if self.at_capacity() {
             // It's ok to unwrap as long as a limit of 0 is forbidden.
-            let limit_doc: ComparableDoc<T> = self
-                .heap
-                .peek()
-                .expect("Top collector with size 0 is forbidden")
-                .clone();
-            if limit_doc.feature < feature {
-                let mut mut_head = self
-                    .heap
-                    .peek_mut()
-                    .expect("Top collector with size 0 is forbidden");
-                mut_head.feature = feature;
-                mut_head.doc_address = DocAddress(self.segment_id, doc);
+            if let Some(limit_feature) = self.heap.peek().map(|head| head.feature.clone()) {
+                if limit_feature < feature {
+                    if let Some(mut head) = self.heap.peek_mut() {
+                        head.feature = feature;
+                        head.doc = doc;
+                    }
+                }
             }
         } else {
-            let wrapped_doc = ComparableDoc {
-                feature,
-                doc_address: DocAddress(self.segment_id, doc),
-            };
-            self.heap.push(wrapped_doc);
+            // we have not reached capacity yet, so we can just push the
+            // element.
+            self.heap.push(ComparableDoc { feature, doc });
         }
     }
 }
 
 #[cfg(test)]
 mod tests {
-    use super::*;
-    use DocId;
+    use super::{TopCollector, TopSegmentCollector};
+    use DocAddress;
     use Score;
 
     #[test]
     fn test_top_collector_not_at_capacity() {
-        let mut top_collector = TopCollector::with_limit(4);
+        let mut top_collector = TopSegmentCollector::new(0, 4);
         top_collector.collect(1, 0.8);
         top_collector.collect(3, 0.2);
         top_collector.collect(5, 0.3);
-        assert!(!top_collector.at_capacity());
-        let score_docs: Vec<(Score, DocId)> = top_collector
-            .top_docs()
-            .into_iter()
-            .map(|(score, doc_address)| (score, doc_address.doc()))
-            .collect();
-        assert_eq!(score_docs, vec![(0.8, 1), (0.3, 5), (0.2, 3)]);
+        assert_eq!(
+            top_collector.harvest(),
+            vec![
+                (0.8, DocAddress(0, 1)),
+                (0.3, DocAddress(0, 5)),
+                (0.2, DocAddress(0, 3))
+            ]
+        );
     }
 
     #[test]
     fn test_top_collector_at_capacity() {
-        let mut top_collector = TopCollector::with_limit(4);
+        let mut top_collector = TopSegmentCollector::new(0, 4);
         top_collector.collect(1, 0.8);
         top_collector.collect(3, 0.2);
         top_collector.collect(5, 0.3);
         top_collector.collect(7, 0.9);
         top_collector.collect(9, -0.2);
-        assert!(top_collector.at_capacity());
-        {
-            let score_docs: Vec<(Score, DocId)> = top_collector
-                .top_docs()
-                .into_iter()
-                .map(|(score, doc_address)| (score, doc_address.doc()))
-                .collect();
-            assert_eq!(score_docs, vec![(0.9, 7), (0.8, 1), (0.3, 5), (0.2, 3)]);
-        }
-        {
-            let docs: Vec<DocId> = top_collector
-                .docs()
-                .into_iter()
-                .map(|doc_address| doc_address.doc())
-                .collect();
-            assert_eq!(docs, vec![7, 1, 5, 3]);
-        }
+        assert_eq!(
+            top_collector.harvest(),
+            vec![
+                (0.9, DocAddress(0, 7)),
+                (0.8, DocAddress(0, 1)),
+                (0.3, DocAddress(0, 5)),
+                (0.2, DocAddress(0, 3))
+            ]
+        );
     }
 
     #[test]
@@ -189,5 +221,4 @@ mod tests {
     fn test_top_0() {
         let _collector: TopCollector<Score> = TopCollector::with_limit(0);
     }
-
 }

src/collector/top_field_collector.rs

@@ -1,12 +1,13 @@
 use super::Collector;
 use collector::top_collector::TopCollector;
+use collector::top_collector::TopSegmentCollector;
+use collector::SegmentCollector;
 use fastfield::FastFieldReader;
 use fastfield::FastValue;
 use schema::Field;
 use DocAddress;
-use DocId;
 use Result;
-use Score;
+use SegmentLocalId;
 use SegmentReader;
 
 /// The Top Field Collector keeps track of the K documents
@@ -19,67 +20,57 @@ use SegmentReader;
 /// ```rust
 /// #[macro_use]
 /// extern crate tantivy;
-/// use tantivy::schema::{SchemaBuilder, TEXT, FAST};
-/// use tantivy::{Index, Result, DocId};
-/// use tantivy::collector::TopFieldCollector;
-/// use tantivy::query::QueryParser;
+/// # use tantivy::schema::{Schema, Field, FAST, TEXT};
+/// # use tantivy::{Index, Result, DocAddress};
+/// # use tantivy::query::{Query, QueryParser};
+/// use tantivy::collector::TopDocs;
 ///
-/// # fn main() { example().unwrap(); }
-/// fn example() -> Result<()> {
-///     let mut schema_builder = SchemaBuilder::new();
-///     let title = schema_builder.add_text_field("title", TEXT);
-///     let rating = schema_builder.add_u64_field("rating", FAST);
-///     let schema = schema_builder.build();
-///     let index = Index::create_in_ram(schema);
-///     {
-///         let mut index_writer = index.writer_with_num_threads(1, 3_000_000)?;
-///         index_writer.add_document(doc!(
-///             title => "The Name of the Wind",
-///             rating => 92u64,
-///         ));
-///         index_writer.add_document(doc!(
-///             title => "The Diary of Muadib",
-///             rating => 97u64,
-///         ));
-///         index_writer.add_document(doc!(
-///             title => "A Dairy Cow",
-///             rating => 63u64,
-///         ));
-///         index_writer.add_document(doc!(
-///             title => "The Diary of a Young Girl",
-///             rating => 80u64,
-///         ));
-///         index_writer.commit().unwrap();
-///     }
+/// # fn main() {
+/// #   let mut schema_builder = Schema::builder();
+/// #   let title = schema_builder.add_text_field("title", TEXT);
+/// #   let rating = schema_builder.add_u64_field("rating", FAST);
+/// #   let schema = schema_builder.build();
+/// #   let index = Index::create_in_ram(schema);
+/// #   let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+/// #   index_writer.add_document(doc!(
+/// #       title => "The Name of the Wind",
+/// #       rating => 92u64,
+/// #   ));
+/// #   index_writer.add_document(doc!(title => "The Diary of Muadib", rating => 97u64));
+/// #   index_writer.add_document(doc!(title => "A Dairy Cow", rating => 63u64));
+/// #   index_writer.add_document(doc!(title => "The Diary of a Young Girl", rating => 80u64));
+/// #   index_writer.commit().unwrap();
+/// #   index.load_searchers().unwrap();
+///	#   let query = QueryParser::for_index(&index, vec![title]).parse_query("diary").unwrap();
+/// #   let top_docs = docs_sorted_by_rating(&index, &query, rating).unwrap();
+/// #   assert_eq!(top_docs,
+/// #            vec![(97u64, DocAddress(0u32, 1)),
+/// #                 (80u64, DocAddress(0u32, 3))]);
+/// # }
+/// #
+/// /// Searches the document matching the given query, and
+/// /// collects the top 10 documents, order by the `field`
+/// /// given in argument.
+/// ///
+/// /// `field` is required to be a FAST field.
+/// fn docs_sorted_by_rating(index: &Index, query: &Query, sort_by_field: Field)
+///     -> Result<Vec<(u64, DocAddress)>> {
 ///
-///     index.load_searchers()?;
-///     let searcher = index.searcher();
+///     // This is where we build our collector!
+///     let top_docs_by_rating = TopDocs::with_limit(2).order_by_field(sort_by_field);
 ///
-///     {
-///	        let mut top_collector = TopFieldCollector::with_limit(rating, 2);
-///         let query_parser = QueryParser::for_index(&index, vec![title]);
-///         let query = query_parser.parse_query("diary")?;
-///         searcher.search(&*query, &mut top_collector).unwrap();
-///
-///         let score_docs: Vec<(u64, DocId)> = top_collector
-///           .top_docs()
-///           .into_iter()
-///           .map(|(field, doc_address)| (field, doc_address.doc()))
-///           .collect();
-///
-///         assert_eq!(score_docs, vec![(97u64, 1), (80, 3)]);
-///     }
-///
-///     Ok(())
+///     // ... and here is our documents. Not this is a simple vec.
+///     // The `u64` in the pair is the value of our fast field for each documents.
+///     index.searcher()
+///          .search(query, &top_docs_by_rating)
 /// }
 /// ```
-pub struct TopFieldCollector<T: FastValue> {
-    field: Field,
+pub struct TopDocsByField<T> {
     collector: TopCollector<T>,
-    fast_field: Option<FastFieldReader<T>>,
+    field: Field,
 }
 
-impl<T: FastValue + PartialOrd + Clone> TopFieldCollector<T> {
+impl<T: FastValue + PartialOrd + Clone> TopDocsByField<T> {
     /// Creates a top field collector, with a number of documents equal to "limit".
     ///
     /// The given field name must be a fast field, otherwise the collector have an error while
@@ -87,68 +78,72 @@ impl<T: FastValue + PartialOrd + Clone> TopFieldCollector<T> {
     ///
     /// # Panics
     /// The method panics if limit is 0
-    pub fn with_limit(field: Field, limit: usize) -> Self {
-        TopFieldCollector {
-            field,
+    pub(crate) fn new(field: Field, limit: usize) -> TopDocsByField<T> {
+        TopDocsByField {
             collector: TopCollector::with_limit(limit),
-            fast_field: None,
+            field,
         }
     }
-
-    /// Returns K best documents sorted the given field name in decreasing order.
-    ///
-    /// Calling this method triggers the sort.
-    /// The result of the sort is not cached.
-    pub fn docs(&self) -> Vec<DocAddress> {
-        self.collector.docs()
-    }
-
-    /// Returns K best FieldDocuments sorted in decreasing order.
-    ///
-    /// Calling this method triggers the sort.
-    /// The result of the sort is not cached.
-    pub fn top_docs(&self) -> Vec<(T, DocAddress)> {
-        self.collector.top_docs()
-    }
-
-    /// Return true iff at least K documents have gone through
-    /// the collector.
-    #[inline]
-    pub fn at_capacity(&self) -> bool {
-        self.collector.at_capacity()
-    }
 }
 
-impl<T: FastValue + PartialOrd + Clone> Collector for TopFieldCollector<T> {
-    fn set_segment(&mut self, segment_id: u32, segment: &SegmentReader) -> Result<()> {
-        self.collector.set_segment_id(segment_id);
-        self.fast_field = Some(segment.fast_field_reader(self.field)?);
-        Ok(())
-    }
+impl<T: FastValue + PartialOrd + Send + Sync + 'static> Collector for TopDocsByField<T> {
+    type Fruit = Vec<(T, DocAddress)>;
 
-    fn collect(&mut self, doc: DocId, _score: Score) {
-        let field_value = self
-            .fast_field
-            .as_ref()
-            .expect("collect() was called before set_segment. This should never happen.")
-            .get(doc);
-        self.collector.collect(doc, field_value);
+    type Child = TopFieldSegmentCollector<T>;
+
+    fn for_segment(
+        &self,
+        segment_local_id: SegmentLocalId,
+        reader: &SegmentReader,
+    ) -> Result<TopFieldSegmentCollector<T>> {
+        let collector = self.collector.for_segment(segment_local_id, reader)?;
+        let reader = reader.fast_field_reader(self.field)?;
+        Ok(TopFieldSegmentCollector { collector, reader })
     }
 
     fn requires_scoring(&self) -> bool {
         false
     }
+
+    fn merge_fruits(
+        &self,
+        segment_fruits: Vec<Vec<(T, DocAddress)>>,
+    ) -> Result<Vec<(T, DocAddress)>> {
+        self.collector.merge_fruits(segment_fruits)
+    }
+}
+
+pub struct TopFieldSegmentCollector<T: FastValue + PartialOrd> {
+    collector: TopSegmentCollector<T>,
+    reader: FastFieldReader<T>,
+}
+
+impl<T: FastValue + PartialOrd + Send + Sync + 'static> SegmentCollector
+    for TopFieldSegmentCollector<T>
+{
+    type Fruit = Vec<(T, DocAddress)>;
+
+    fn collect(&mut self, doc: u32, _score: f32) {
+        let field_value = self.reader.get(doc);
+        self.collector.collect(doc, field_value);
+    }
+
+    fn harvest(self) -> Vec<(T, DocAddress)> {
+        self.collector.harvest()
+    }
 }
 
 #[cfg(test)]
 mod tests {
-    use super::*;
+    use super::TopDocsByField;
+    use collector::Collector;
+    use collector::TopDocs;
     use query::Query;
     use query::QueryParser;
     use schema::Field;
     use schema::IntOptions;
-    use schema::Schema;
-    use schema::{SchemaBuilder, FAST, TEXT};
+    use schema::{Schema, FAST, TEXT};
+    use DocAddress;
     use Index;
     use IndexWriter;
     use TantivyError;
@@ -158,7 +153,7 @@ mod tests {
 
     #[test]
     fn test_top_collector_not_at_capacity() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let title = schema_builder.add_text_field(TITLE, TEXT);
         let size = schema_builder.add_u64_field(SIZE, FAST);
         let schema = schema_builder.build();
@@ -178,22 +173,22 @@ mod tests {
         });
         let searcher = index.searcher();
 
-        let mut top_collector = TopFieldCollector::with_limit(size, 4);
-        searcher.search(&*query, &mut top_collector).unwrap();
-        assert!(!top_collector.at_capacity());
-
-        let score_docs: Vec<(u64, DocId)> = top_collector
-            .top_docs()
-            .into_iter()
-            .map(|(field, doc_address)| (field, doc_address.doc()))
-            .collect();
-        assert_eq!(score_docs, vec![(64, 1), (16, 2), (12, 0)]);
+        let top_collector = TopDocs::with_limit(4).order_by_field(size);
+        let top_docs: Vec<(u64, DocAddress)> = searcher.search(&query, &top_collector).unwrap();
+        assert_eq!(
+            top_docs,
+            vec![
+                (64, DocAddress(0, 1)),
+                (16, DocAddress(0, 2)),
+                (12, DocAddress(0, 0))
+            ]
+        );
     }
 
     #[test]
     #[should_panic]
     fn test_field_does_not_exist() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let title = schema_builder.add_text_field(TITLE, TEXT);
         let size = schema_builder.add_u64_field(SIZE, FAST);
         let schema = schema_builder.build();
@@ -204,14 +199,16 @@ mod tests {
             ));
         });
         let searcher = index.searcher();
-        let segment = searcher.segment_reader(0);
-        let mut top_collector: TopFieldCollector<u64> = TopFieldCollector::with_limit(Field(2), 4);
-        let _ = top_collector.set_segment(0, segment);
+        let top_collector: TopDocsByField<u64> = TopDocs::with_limit(4).order_by_field(Field(2));
+        let segment_reader = searcher.segment_reader(0u32);
+        top_collector
+            .for_segment(0, segment_reader)
+            .expect("should panic");
     }
 
     #[test]
     fn test_field_not_fast_field() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let title = schema_builder.add_text_field(TITLE, TEXT);
         let size = schema_builder.add_u64_field(SIZE, IntOptions::default());
         let schema = schema_builder.build();
@@ -223,26 +220,16 @@ mod tests {
         });
         let searcher = index.searcher();
         let segment = searcher.segment_reader(0);
-        let mut top_collector: TopFieldCollector<u64> = TopFieldCollector::with_limit(size, 4);
+        let top_collector: TopDocsByField<u64> = TopDocs::with_limit(4).order_by_field(size);
         assert_matches!(
-            top_collector.set_segment(0, segment),
-            Err(TantivyError::FastFieldError(_))
+            top_collector
+                .for_segment(0, segment)
+                .map(|_| ())
+                .unwrap_err(),
+            TantivyError::FastFieldError(_)
         );
     }
 
-    #[test]
-    #[should_panic]
-    fn test_collect_before_set_segment() {
-        let mut top_collector: TopFieldCollector<u64> = TopFieldCollector::with_limit(Field(0), 4);
-        top_collector.collect(0, 0f32);
-    }
-
-    #[test]
-    #[should_panic]
-    fn test_top_0() {
-        let _: TopFieldCollector<u64> = TopFieldCollector::with_limit(Field(0), 0);
-    }
-
     fn index(
         query: &str,
         query_field: Field,

src/collector/top_score_collector.rs

@@ -1,5 +1,10 @@
 use super::Collector;
 use collector::top_collector::TopCollector;
+use collector::top_collector::TopSegmentCollector;
+use collector::SegmentCollector;
+use collector::TopDocsByField;
+use fastfield::FastValue;
+use schema::Field;
 use DocAddress;
 use DocId;
 use Result;
@@ -17,14 +22,15 @@ use SegmentReader;
 /// ```rust
 /// #[macro_use]
 /// extern crate tantivy;
-/// use tantivy::schema::{SchemaBuilder, TEXT};
-/// use tantivy::{Index, Result, DocId, Score};
-/// use tantivy::collector::TopScoreCollector;
+/// use tantivy::DocAddress;
+/// use tantivy::schema::{Schema, TEXT};
+/// use tantivy::{Index, Result};
+/// use tantivy::collector::TopDocs;
 /// use tantivy::query::QueryParser;
 ///
 /// # fn main() { example().unwrap(); }
 /// fn example() -> Result<()> {
-///     let mut schema_builder = SchemaBuilder::new();
+///     let mut schema_builder = Schema::builder();
 ///     let title = schema_builder.add_text_field("title", TEXT);
 ///     let schema = schema_builder.build();
 ///     let index = Index::create_in_ram(schema);
@@ -48,140 +54,147 @@ use SegmentReader;
 ///     index.load_searchers()?;
 ///     let searcher = index.searcher();
 ///
-///     {
-///	        let mut top_collector = TopScoreCollector::with_limit(2);
-///         let query_parser = QueryParser::for_index(&index, vec![title]);
-///         let query = query_parser.parse_query("diary")?;
-///         searcher.search(&*query, &mut top_collector).unwrap();
+///     let query_parser = QueryParser::for_index(&index, vec![title]);
+///     let query = query_parser.parse_query("diary")?;
+///     let top_docs = searcher.search(&query, &TopDocs::with_limit(2))?;
 ///
-///         let score_docs: Vec<(Score, DocId)> = top_collector
-///           .top_docs()
-///           .into_iter()
-///           .map(|(score, doc_address)| (score, doc_address.doc()))
-///           .collect();
-///
-///         assert_eq!(score_docs, vec![(0.7261542, 1), (0.6099695, 3)]);
-///     }
+///     assert_eq!(&top_docs[0], &(0.7261542, DocAddress(0, 1)));
+///     assert_eq!(&top_docs[1], &(0.6099695, DocAddress(0, 3)));
 ///
 ///     Ok(())
 /// }
 /// ```
-pub struct TopScoreCollector {
-    collector: TopCollector<Score>,
-}
+pub struct TopDocs(TopCollector<Score>);
 
-impl TopScoreCollector {
+impl TopDocs {
     /// Creates a top score collector, with a number of documents equal to "limit".
     ///
     /// # Panics
     /// The method panics if limit is 0
-    pub fn with_limit(limit: usize) -> TopScoreCollector {
-        TopScoreCollector {
-            collector: TopCollector::with_limit(limit),
-        }
+    pub fn with_limit(limit: usize) -> TopDocs {
+        TopDocs(TopCollector::with_limit(limit))
     }
 
-    /// Returns K best scored documents sorted in decreasing order.
+    /// Set top-K to rank documents by a given fast field.
     ///
-    /// Calling this method triggers the sort.
-    /// The result of the sort is not cached.
-    pub fn docs(&self) -> Vec<DocAddress> {
-        self.collector.docs()
-    }
-
-    /// Returns K best ScoredDocuments sorted in decreasing order.
-    ///
-    /// Calling this method triggers the sort.
-    /// The result of the sort is not cached.
-    pub fn top_docs(&self) -> Vec<(Score, DocAddress)> {
-        self.collector.top_docs()
-    }
-
-    /// Returns K best ScoredDocuments sorted in decreasing order.
-    ///
-    /// Calling this method triggers the sort.
-    /// The result of the sort is not cached.
-    #[deprecated]
-    pub fn score_docs(&self) -> Vec<(Score, DocAddress)> {
-        self.collector.top_docs()
-    }
-
-    /// Return true iff at least K documents have gone through
-    /// the collector.
-    #[inline]
-    pub fn at_capacity(&self) -> bool {
-        self.collector.at_capacity()
+    /// (By default, `TopDocs` collects the top-K documents sorted by
+    /// the similarity score.)
+    pub fn order_by_field<T: PartialOrd + FastValue + Clone>(
+        self,
+        field: Field,
+    ) -> TopDocsByField<T> {
+        TopDocsByField::new(field, self.0.limit())
     }
 }
 
-impl Collector for TopScoreCollector {
-    fn set_segment(&mut self, segment_id: SegmentLocalId, _: &SegmentReader) -> Result<()> {
-        self.collector.set_segment_id(segment_id);
-        Ok(())
-    }
+impl Collector for TopDocs {
+    type Fruit = Vec<(Score, DocAddress)>;
 
-    fn collect(&mut self, doc: DocId, score: Score) {
-        self.collector.collect(doc, score);
+    type Child = TopScoreSegmentCollector;
+
+    fn for_segment(
+        &self,
+        segment_local_id: SegmentLocalId,
+        reader: &SegmentReader,
+    ) -> Result<Self::Child> {
+        let collector = self.0.for_segment(segment_local_id, reader)?;
+        Ok(TopScoreSegmentCollector(collector))
     }
 
     fn requires_scoring(&self) -> bool {
         true
     }
+
+    fn merge_fruits(&self, child_fruits: Vec<Vec<(Score, DocAddress)>>) -> Result<Self::Fruit> {
+        self.0.merge_fruits(child_fruits)
+    }
+}
+
+/// Segment Collector associated to `TopDocs`.
+pub struct TopScoreSegmentCollector(TopSegmentCollector<Score>);
+
+impl SegmentCollector for TopScoreSegmentCollector {
+    type Fruit = Vec<(Score, DocAddress)>;
+
+    fn collect(&mut self, doc: DocId, score: Score) {
+        self.0.collect(doc, score)
+    }
+
+    fn harvest(self) -> Vec<(Score, DocAddress)> {
+        self.0.harvest()
+    }
 }
 
 #[cfg(test)]
 mod tests {
-    use super::*;
-    use collector::Collector;
-    use DocId;
+    use super::TopDocs;
+    use query::QueryParser;
+    use schema::Schema;
+    use schema::TEXT;
+    use DocAddress;
+    use Index;
     use Score;
 
+    fn make_index() -> Index {
+        let mut schema_builder = Schema::builder();
+        let text_field = schema_builder.add_text_field("text", TEXT);
+        let schema = schema_builder.build();
+        let index = Index::create_in_ram(schema);
+        {
+            // writing the segment
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
+            index_writer.add_document(doc!(text_field=>"Hello happy tax payer."));
+            index_writer.add_document(doc!(text_field=>"Droopy says hello happy tax payer"));
+            index_writer.add_document(doc!(text_field=>"I like Droopy"));
+            assert!(index_writer.commit().is_ok());
+        }
+        index.load_searchers().unwrap();
+        index
+    }
+
     #[test]
     fn test_top_collector_not_at_capacity() {
-        let mut top_collector = TopScoreCollector::with_limit(4);
-        top_collector.collect(1, 0.8);
-        top_collector.collect(3, 0.2);
-        top_collector.collect(5, 0.3);
-        assert!(!top_collector.at_capacity());
-        let score_docs: Vec<(Score, DocId)> = top_collector
-            .top_docs()
-            .into_iter()
-            .map(|(score, doc_address)| (score, doc_address.doc()))
-            .collect();
-        assert_eq!(score_docs, vec![(0.8, 1), (0.3, 5), (0.2, 3)]);
+        let index = make_index();
+        let field = index.schema().get_field("text").unwrap();
+        let query_parser = QueryParser::for_index(&index, vec![field]);
+        let text_query = query_parser.parse_query("droopy tax").unwrap();
+        let score_docs: Vec<(Score, DocAddress)> = index
+            .searcher()
+            .search(&text_query, &TopDocs::with_limit(4))
+            .unwrap();
+        assert_eq!(
+            score_docs,
+            vec![
+                (0.81221175, DocAddress(0u32, 1)),
+                (0.5376842, DocAddress(0u32, 2)),
+                (0.48527452, DocAddress(0, 0))
+            ]
+        );
     }
 
     #[test]
     fn test_top_collector_at_capacity() {
-        let mut top_collector = TopScoreCollector::with_limit(4);
-        top_collector.collect(1, 0.8);
-        top_collector.collect(3, 0.2);
-        top_collector.collect(5, 0.3);
-        top_collector.collect(7, 0.9);
-        top_collector.collect(9, -0.2);
-        assert!(top_collector.at_capacity());
-        {
-            let score_docs: Vec<(Score, DocId)> = top_collector
-                .top_docs()
-                .into_iter()
-                .map(|(score, doc_address)| (score, doc_address.doc()))
-                .collect();
-            assert_eq!(score_docs, vec![(0.9, 7), (0.8, 1), (0.3, 5), (0.2, 3)]);
-        }
-        {
-            let docs: Vec<DocId> = top_collector
-                .docs()
-                .into_iter()
-                .map(|doc_address| doc_address.doc())
-                .collect();
-            assert_eq!(docs, vec![7, 1, 5, 3]);
-        }
+        let index = make_index();
+        let field = index.schema().get_field("text").unwrap();
+        let query_parser = QueryParser::for_index(&index, vec![field]);
+        let text_query = query_parser.parse_query("droopy tax").unwrap();
+        let score_docs: Vec<(Score, DocAddress)> = index
+            .searcher()
+            .search(&text_query, &TopDocs::with_limit(2))
+            .unwrap();
+        assert_eq!(
+            score_docs,
+            vec![
+                (0.81221175, DocAddress(0u32, 1)),
+                (0.5376842, DocAddress(0u32, 2)),
+            ]
+        );
     }
 
     #[test]
     #[should_panic]
     fn test_top_0() {
-        TopScoreCollector::with_limit(0);
+        TopDocs::with_limit(0);
     }
 
 }

src/common/bitpacker.rs

@@ -1,9 +1,6 @@
-use common::serialize::BinarySerializable;
+use byteorder::{ByteOrder, LittleEndian, WriteBytesExt};
 use std::io;
-use std::io::Write;
-use std::mem;
 use std::ops::Deref;
-use std::ptr;
 
 pub(crate) struct BitPacker {
     mini_buffer: u64,
@@ -18,7 +15,7 @@ impl BitPacker {
         }
     }
 
-    pub fn write<TWrite: Write>(
+    pub fn write<TWrite: io::Write>(
         &mut self,
         val: u64,
         num_bits: u8,
@@ -28,14 +25,14 @@ impl BitPacker {
         let num_bits = num_bits as usize;
         if self.mini_buffer_written + num_bits > 64 {
             self.mini_buffer |= val_u64.wrapping_shl(self.mini_buffer_written as u32);
-            self.mini_buffer.serialize(output)?;
+            output.write_u64::<LittleEndian>(self.mini_buffer)?;
             self.mini_buffer = val_u64.wrapping_shr((64 - self.mini_buffer_written) as u32);
             self.mini_buffer_written = self.mini_buffer_written + num_bits - 64;
         } else {
             self.mini_buffer |= val_u64 << self.mini_buffer_written;
             self.mini_buffer_written += num_bits;
             if self.mini_buffer_written == 64 {
-                self.mini_buffer.serialize(output)?;
+                output.write_u64::<LittleEndian>(self.mini_buffer)?;
                 self.mini_buffer_written = 0;
                 self.mini_buffer = 0u64;
             }
@@ -43,17 +40,18 @@ impl BitPacker {
         Ok(())
     }
 
-    pub fn flush<TWrite: Write>(&mut self, output: &mut TWrite) -> io::Result<()> {
+    pub fn flush<TWrite: io::Write>(&mut self, output: &mut TWrite) -> io::Result<()> {
         if self.mini_buffer_written > 0 {
             let num_bytes = (self.mini_buffer_written + 7) / 8;
-            let arr: [u8; 8] = unsafe { mem::transmute::<u64, [u8; 8]>(self.mini_buffer.to_le()) };
+            let mut arr: [u8; 8] = [0u8; 8];
+            LittleEndian::write_u64(&mut arr, self.mini_buffer);
             output.write_all(&arr[..num_bytes])?;
             self.mini_buffer_written = 0;
         }
         Ok(())
     }
 
-    pub fn close<TWrite: Write>(&mut self, output: &mut TWrite) -> io::Result<()> {
+    pub fn close<TWrite: io::Write>(&mut self, output: &mut TWrite) -> io::Result<()> {
         self.flush(output)?;
         // Padding the write file to simplify reads.
         output.write_all(&[0u8; 7])?;
@@ -102,9 +100,7 @@ where
             addr + 8 <= data.len(),
             "The fast field field should have been padded with 7 bytes."
         );
-        #[cfg_attr(feature = "cargo-clippy", allow(clippy::cast_ptr_alignment))]
-        let val_unshifted_unmasked: u64 =
-            u64::from_le(unsafe { ptr::read_unaligned(data[addr..].as_ptr() as *const u64) });
+        let val_unshifted_unmasked: u64 = LittleEndian::read_u64(&data[addr..]);
         let val_shifted = (val_unshifted_unmasked >> bit_shift) as u64;
         val_shifted & mask
     }
@@ -126,9 +122,7 @@ where
             for output_val in output.iter_mut() {
                 let addr = addr_in_bits >> 3;
                 let bit_shift = addr_in_bits & 7;
-                #[cfg_attr(feature = "cargo-clippy", allow(clippy::cast_ptr_alignment))]
-                let val_unshifted_unmasked: u64 =
-                    unsafe { ptr::read_unaligned(data[addr..].as_ptr() as *const u64) };
+                let val_unshifted_unmasked: u64 = LittleEndian::read_u64(&data[addr..]);
                 let val_shifted = (val_unshifted_unmasked >> bit_shift) as u64;
                 *output_val = val_shifted & mask;
                 addr_in_bits += num_bits;

src/common/composite_file.rs

@@ -4,6 +4,8 @@ use common::VInt;
 use directory::ReadOnlySource;
 use directory::WritePtr;
 use schema::Field;
+use space_usage::FieldUsage;
+use space_usage::PerFieldSpaceUsage;
 use std::collections::HashMap;
 use std::io::Write;
 use std::io::{self, Read};
@@ -166,6 +168,17 @@ impl CompositeFile {
             .get(&FileAddr { field, idx })
             .map(|&(from, to)| self.data.slice(from, to))
     }
+
+    pub fn space_usage(&self) -> PerFieldSpaceUsage {
+        let mut fields = HashMap::new();
+        for (&field_addr, &(start, end)) in self.offsets_index.iter() {
+            fields
+                .entry(field_addr.field)
+                .or_insert_with(|| FieldUsage::empty(field_addr.field))
+                .add_field_idx(field_addr.idx, end - start);
+        }
+        PerFieldSpaceUsage::new(fields)
+    }
 }
 
 #[cfg(test)]

src/core/executor.rs

@@ -0,0 +1,137 @@
+use crossbeam::channel;
+use scoped_pool::{Pool, ThreadConfig};
+use Result;
+
+/// Search executor whether search request are single thread or multithread.
+///
+/// We don't expose Rayon thread pool directly here for several reasons.
+///
+/// First dependency hell. It is not a good idea to expose the
+/// API of a dependency, knowing it might conflict with a different version
+/// used by the client. Second, we may stop using rayon in the future.
+pub enum Executor {
+    SingleThread,
+    ThreadPool(Pool),
+}
+
+impl Executor {
+    /// Creates an Executor that performs all task in the caller thread.
+    pub fn single_thread() -> Executor {
+        Executor::SingleThread
+    }
+
+    // Creates an Executor that dispatches the tasks in a thread pool.
+    pub fn multi_thread(num_threads: usize, prefix: &'static str) -> Executor {
+        let thread_config = ThreadConfig::new().prefix(prefix);
+        let pool = Pool::with_thread_config(num_threads, thread_config);
+        Executor::ThreadPool(pool)
+    }
+
+    // Perform a map in the thread pool.
+    //
+    // Regardless of the executor (`SingleThread` or `ThreadPool`), panics in the task
+    // will propagate to the caller.
+    pub fn map<
+        A: Send,
+        R: Send,
+        AIterator: Iterator<Item = A>,
+        F: Sized + Sync + Fn(A) -> Result<R>,
+    >(
+        &self,
+        f: F,
+        args: AIterator,
+    ) -> Result<Vec<R>> {
+        match self {
+            Executor::SingleThread => args.map(f).collect::<Result<_>>(),
+            Executor::ThreadPool(pool) => {
+                let args_with_indices: Vec<(usize, A)> = args.enumerate().collect();
+                let num_fruits = args_with_indices.len();
+                let fruit_receiver = {
+                    let (fruit_sender, fruit_receiver) = channel::unbounded();
+                    pool.scoped(|scope| {
+                        for arg_with_idx in args_with_indices {
+                            scope.execute(|| {
+                                let (idx, arg) = arg_with_idx;
+                                let fruit = f(arg);
+                                if let Err(err) = fruit_sender.send((idx, fruit)) {
+                                    error!("Failed to send search task. It probably means all search threads have panicked. {:?}", err);
+                                }
+                            });
+                        }
+                    });
+                    fruit_receiver
+                    // This ends the scope of fruit_sender.
+                    // This is important as it makes it possible for the fruit_receiver iteration to
+                    // terminate.
+                };
+                // This is lame, but it does not use unsafe code.
+                let mut results_with_position = Vec::with_capacity(num_fruits);
+                for (pos, fruit_res) in fruit_receiver {
+                    let fruit = fruit_res?;
+                    results_with_position.push((pos, fruit));
+                }
+                results_with_position.sort_by_key(|(pos, _)| *pos);
+                assert_eq!(results_with_position.len(), num_fruits);
+                Ok(results_with_position
+                    .into_iter()
+                    .map(|(_, fruit)| fruit)
+                    .collect::<Vec<_>>())
+            }
+        }
+    }
+}
+
+#[cfg(test)]
+mod tests {
+
+    use super::Executor;
+
+    #[test]
+    #[should_panic(expected = "panic should propagate")]
+    fn test_panic_propagates_single_thread() {
+        let _result: Vec<usize> = Executor::single_thread()
+            .map(
+                |_| {
+                    panic!("panic should propagate");
+                },
+                vec![0].into_iter(),
+            )
+            .unwrap();
+    }
+
+    #[test]
+    #[should_panic] //< unfortunately the panic message is not propagated
+    fn test_panic_propagates_multi_thread() {
+        let _result: Vec<usize> = Executor::multi_thread(1, "search-test")
+            .map(
+                |_| {
+                    panic!("panic should propagate");
+                },
+                vec![0].into_iter(),
+            )
+            .unwrap();
+    }
+
+    #[test]
+    fn test_map_singlethread() {
+        let result: Vec<usize> = Executor::single_thread()
+            .map(|i| Ok(i * 2), 0..1_000)
+            .unwrap();
+        assert_eq!(result.len(), 1_000);
+        for i in 0..1_000 {
+            assert_eq!(result[i], i * 2);
+        }
+    }
+
+}
+
+#[test]
+fn test_map_multithread() {
+    let result: Vec<usize> = Executor::multi_thread(3, "search-test")
+        .map(|i| Ok(i * 2), 0..10)
+        .unwrap();
+    assert_eq!(result.len(), 10);
+    for i in 0..10 {
+        assert_eq!(result[i], i * 2);
+    }
+}

src/core/index.rs

@@ -3,6 +3,7 @@ use super::pool::Pool;
 use super::segment::create_segment;
 use super::segment::Segment;
 use core::searcher::Searcher;
+use core::Executor;
 use core::IndexMeta;
 use core::SegmentId;
 use core::SegmentMeta;
@@ -12,6 +13,7 @@ use directory::ManagedDirectory;
 #[cfg(feature = "mmap")]
 use directory::MmapDirectory;
 use directory::{Directory, RAMDirectory};
+use error::DataCorruption;
 use error::TantivyError;
 use indexer::index_writer::open_index_writer;
 use indexer::index_writer::HEAP_SIZE_MIN;
@@ -36,7 +38,13 @@ fn load_metas(directory: &Directory) -> Result<IndexMeta> {
     let meta_data = directory.atomic_read(&META_FILEPATH)?;
     let meta_string = String::from_utf8_lossy(&meta_data);
     serde_json::from_str(&meta_string)
-        .map_err(|_| TantivyError::CorruptedFile(META_FILEPATH.clone()))
+        .map_err(|e| {
+            DataCorruption::new(
+                META_FILEPATH.clone(),
+                format!("Meta file cannot be deserialized. {:?}.", e),
+            )
+        })
+        .map_err(From::from)
 }
 
 /// Search Index
@@ -45,10 +53,39 @@ pub struct Index {
     schema: Schema,
     num_searchers: Arc<AtomicUsize>,
     searcher_pool: Arc<Pool<Searcher>>,
+    executor: Arc<Executor>,
     tokenizers: TokenizerManager,
 }
 
 impl Index {
+    /// Examines the director to see if it contains an index
+    pub fn exists<Dir: Directory>(dir: &Dir) -> bool {
+        dir.exists(&META_FILEPATH)
+    }
+
+    /// Accessor to the search executor.
+    ///
+    /// This pool is used by default when calling `searcher.search(...)`
+    /// to perform search on the individual segments.
+    ///
+    /// By default the executor is single thread, and simply runs in the calling thread.
+    pub fn search_executor(&self) -> &Executor {
+        self.executor.as_ref()
+    }
+
+    /// Replace the default single thread search executor pool
+    /// by a thread pool with a given number of threads.
+    pub fn set_multithread_executor(&mut self, num_threads: usize) {
+        self.executor = Arc::new(Executor::multi_thread(num_threads, "thrd-tantivy-search-"));
+    }
+
+    /// Replace the default single thread search executor pool
+    /// by a thread pool with a given number of threads.
+    pub fn set_default_multithread_executor(&mut self) {
+        let default_num_threads = num_cpus::get();
+        self.set_multithread_executor(default_num_threads);
+    }
+
     /// Creates a new index using the `RAMDirectory`.
     ///
     /// The index will be allocated in anonymous memory.
@@ -65,9 +102,30 @@ impl Index {
     #[cfg(feature = "mmap")]
     pub fn create_in_dir<P: AsRef<Path>>(directory_path: P, schema: Schema) -> Result<Index> {
         let mmap_directory = MmapDirectory::open(directory_path)?;
+        if Index::exists(&mmap_directory) {
+            return Err(TantivyError::IndexAlreadyExists);
+        }
+
         Index::create(mmap_directory, schema)
     }
 
+    /// Opens or creates a new index in the provided directory
+    #[cfg(feature = "mmap")]
+    pub fn open_or_create<Dir: Directory>(dir: Dir, schema: Schema) -> Result<Index> {
+        if Index::exists(&dir) {
+            let index = Index::open(dir)?;
+            if index.schema() == schema {
+                Ok(index)
+            } else {
+                Err(TantivyError::SchemaError(
+                    "An index exists but the schema does not match.".to_string(),
+                ))
+            }
+        } else {
+            Index::create(dir, schema)
+        }
+    }
+
     /// Creates a new index in a temp directory.
     ///
     /// The index will use the `MMapDirectory` in a newly created directory.
@@ -84,11 +142,13 @@ impl Index {
 
     /// Creates a new index given an implementation of the trait `Directory`
     pub fn create<Dir: Directory>(dir: Dir, schema: Schema) -> Result<Index> {
-        let directory = ManagedDirectory::new(dir)?;
+        let directory = ManagedDirectory::wrap(dir)?;
         Index::from_directory(directory, schema)
     }
 
     /// Create a new index from a directory.
+    ///
+    /// This will overwrite existing meta.json
     fn from_directory(mut directory: ManagedDirectory, schema: Schema) -> Result<Index> {
         save_new_metas(schema.clone(), 0, directory.borrow_mut())?;
         let metas = IndexMeta::with_schema(schema);
@@ -105,6 +165,7 @@ impl Index {
             num_searchers: Arc::new(AtomicUsize::new(n_cpus)),
             searcher_pool: Arc::new(Pool::new()),
             tokenizers: TokenizerManager::default(),
+            executor: Arc::new(Executor::single_thread()),
         };
         index.load_searchers()?;
         Ok(index)
@@ -145,7 +206,7 @@ impl Index {
 
     /// Open the index using the provided directory
     pub fn open<D: Directory>(directory: D) -> Result<Index> {
-        let directory = ManagedDirectory::new(directory)?;
+        let directory = ManagedDirectory::wrap(directory)?;
         let metas = load_metas(&directory)?;
         Index::create_from_metas(directory, &metas)
     }
@@ -322,18 +383,20 @@ impl Clone for Index {
             num_searchers: Arc::clone(&self.num_searchers),
             searcher_pool: Arc::clone(&self.searcher_pool),
             tokenizers: self.tokenizers.clone(),
+            executor: self.executor.clone(),
         }
     }
 }
 
 #[cfg(test)]
 mod tests {
-    use schema::{SchemaBuilder, INT_INDEXED, TEXT};
+    use directory::RAMDirectory;
+    use schema::{Schema, INT_INDEXED, TEXT};
     use Index;
 
     #[test]
     fn test_indexer_for_field() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let num_likes_field = schema_builder.add_u64_field("num_likes", INT_INDEXED);
         let body_field = schema_builder.add_text_field("body", TEXT);
         let schema = schema_builder.build();
@@ -345,4 +408,54 @@ mod tests {
         );
     }
 
+    #[test]
+    fn test_index_exists() {
+        let directory = RAMDirectory::create();
+        assert!(!Index::exists(&directory));
+        assert!(Index::create(directory.clone(), throw_away_schema()).is_ok());
+        assert!(Index::exists(&directory));
+    }
+
+    #[test]
+    fn open_or_create_should_create() {
+        let directory = RAMDirectory::create();
+        assert!(!Index::exists(&directory));
+        assert!(Index::open_or_create(directory.clone(), throw_away_schema()).is_ok());
+        assert!(Index::exists(&directory));
+    }
+
+    #[test]
+    fn open_or_create_should_open() {
+        let directory = RAMDirectory::create();
+        assert!(Index::create(directory.clone(), throw_away_schema()).is_ok());
+        assert!(Index::exists(&directory));
+        assert!(Index::open_or_create(directory, throw_away_schema()).is_ok());
+    }
+
+    #[test]
+    fn create_should_wipeoff_existing() {
+        let directory = RAMDirectory::create();
+        assert!(Index::create(directory.clone(), throw_away_schema()).is_ok());
+        assert!(Index::exists(&directory));
+        assert!(Index::create(directory.clone(), Schema::builder().build()).is_ok());
+    }
+
+    #[test]
+    fn open_or_create_exists_but_schema_does_not_match() {
+        let directory = RAMDirectory::create();
+        assert!(Index::create(directory.clone(), throw_away_schema()).is_ok());
+        assert!(Index::exists(&directory));
+        assert!(Index::open_or_create(directory.clone(), throw_away_schema()).is_ok());
+        let err = Index::open_or_create(directory, Schema::builder().build());
+        assert_eq!(
+            format!("{:?}", err.unwrap_err()),
+            "SchemaError(\"An index exists but the schema does not match.\")"
+        );
+    }
+
+    fn throw_away_schema() -> Schema {
+        let mut schema_builder = Schema::builder();
+        let _ = schema_builder.add_u64_field("num_likes", INT_INDEXED);
+        schema_builder.build()
+    }
 }

src/core/index_meta.rs

@@ -46,13 +46,13 @@ impl fmt::Debug for IndexMeta {
 mod tests {
 
     use super::IndexMeta;
-    use schema::{SchemaBuilder, TEXT};
+    use schema::{Schema, TEXT};
     use serde_json;
 
     #[test]
     fn test_serialize_metas() {
         let schema = {
-            let mut schema_builder = SchemaBuilder::new();
+            let mut schema_builder = Schema::builder();
             schema_builder.add_text_field("text", TEXT);
             schema_builder.build()
         };

src/core/inverted_index_reader.rs

@@ -32,10 +32,7 @@ pub struct InvertedIndexReader {
 }
 
 impl InvertedIndexReader {
-    #[cfg_attr(
-        feature = "cargo-clippy",
-        allow(clippy::needless_pass_by_value)
-    )] // for symetry
+    #[cfg_attr(feature = "cargo-clippy", allow(clippy::needless_pass_by_value))] // for symetry
     pub(crate) fn new(
         termdict: TermDictionary,
         postings_source: ReadOnlySource,

src/core/mod.rs

@@ -1,3 +1,4 @@
+mod executor;
 pub mod index;
 mod index_meta;
 mod inverted_index_reader;
@@ -9,6 +10,7 @@ mod segment_id;
 mod segment_meta;
 mod segment_reader;
 
+pub use self::executor::Executor;
 pub use self::index::Index;
 pub use self::index_meta::IndexMeta;
 pub use self::inverted_index_reader::InvertedIndexReader;

src/core/searcher.rs

@@ -1,17 +1,43 @@
 use collector::Collector;
+use collector::SegmentCollector;
+use core::Executor;
 use core::InvertedIndexReader;
 use core::SegmentReader;
 use query::Query;
+use query::Scorer;
+use query::Weight;
 use schema::Document;
 use schema::Schema;
 use schema::{Field, Term};
+use space_usage::SearcherSpaceUsage;
 use std::fmt;
 use std::sync::Arc;
+use store::StoreReader;
 use termdict::TermMerger;
 use DocAddress;
 use Index;
 use Result;
 
+fn collect_segment<C: Collector>(
+    collector: &C,
+    weight: &Weight,
+    segment_ord: u32,
+    segment_reader: &SegmentReader,
+) -> Result<C::Fruit> {
+    let mut scorer = weight.scorer(segment_reader)?;
+    let mut segment_collector = collector.for_segment(segment_ord as u32, segment_reader)?;
+    if let Some(delete_bitset) = segment_reader.delete_bitset() {
+        scorer.for_each(&mut |doc, score| {
+            if !delete_bitset.is_deleted(doc) {
+                segment_collector.collect(doc, score);
+            }
+        });
+    } else {
+        scorer.for_each(&mut |doc, score| segment_collector.collect(doc, score));
+    }
+    Ok(segment_collector.harvest())
+}
+
 /// Holds a list of `SegmentReader`s ready for search.
 ///
 /// It guarantees that the `Segment` will not be removed before
@@ -21,6 +47,7 @@ pub struct Searcher {
     schema: Schema,
     index: Index,
     segment_readers: Vec<SegmentReader>,
+    store_readers: Vec<StoreReader>,
 }
 
 impl Searcher {
@@ -30,10 +57,15 @@ impl Searcher {
         index: Index,
         segment_readers: Vec<SegmentReader>,
     ) -> Searcher {
+        let store_readers = segment_readers
+            .iter()
+            .map(|segment_reader| segment_reader.get_store_reader())
+            .collect();
         Searcher {
             schema,
             index,
             segment_readers,
+            store_readers,
         }
     }
 
@@ -48,8 +80,8 @@ impl Searcher {
     /// the request to the right `Segment`.
     pub fn doc(&self, doc_address: DocAddress) -> Result<Document> {
         let DocAddress(segment_local_id, doc_id) = doc_address;
-        let segment_reader = &self.segment_readers[segment_local_id as usize];
-        segment_reader.doc(doc_id)
+        let store_reader = &self.store_readers[segment_local_id as usize];
+        store_reader.get(doc_id)
     }
 
     /// Access the schema associated to the index of this searcher.
@@ -72,7 +104,8 @@ impl Searcher {
             .iter()
             .map(|segment_reader| {
                 u64::from(segment_reader.inverted_index(term.field()).doc_freq(term))
-            }).sum::<u64>()
+            })
+            .sum::<u64>()
     }
 
     /// Return the list of segment readers
@@ -85,9 +118,58 @@ impl Searcher {
         &self.segment_readers[segment_ord as usize]
     }
 
-    /// Runs a query on the segment readers wrapped by the searcher
-    pub fn search<C: Collector>(&self, query: &Query, collector: &mut C) -> Result<()> {
-        query.search(self, collector)
+    /// Runs a query on the segment readers wrapped by the searcher.
+    ///
+    /// Search works as follows :
+    ///
+    ///  First the weight object associated to the query is created.
+    ///
+    ///  Then, the query loops over the segments and for each segment :
+    ///  - setup the collector and informs it that the segment being processed has changed.
+    ///  - creates a SegmentCollector for collecting documents associated to the segment
+    ///  - creates a `Scorer` object associated for this segment
+    ///  - iterate through the matched documents and push them to the segment collector.
+    ///
+    ///  Finally, the Collector merges each of the child collectors into itself for result usability
+    ///  by the caller.
+    pub fn search<C: Collector>(&self, query: &Query, collector: &C) -> Result<C::Fruit> {
+        let executor = self.index.search_executor();
+        self.search_with_executor(query, collector, executor)
+    }
+
+    /// Same as [`search(...)`](#method.search) but multithreaded.
+    ///
+    /// The current implementation is rather naive :
+    /// multithreading is by splitting search into as many task
+    /// as there are segments.
+    ///
+    /// It is powerless at making search faster if your index consists in
+    /// one large segment.
+    ///
+    /// Also, keep in my multithreading a single query on several
+    /// threads will not improve your throughput. It can actually
+    /// hurt it. It will however, decrease the average response time.
+    pub fn search_with_executor<C: Collector>(
+        &self,
+        query: &Query,
+        collector: &C,
+        executor: &Executor,
+    ) -> Result<C::Fruit> {
+        let scoring_enabled = collector.requires_scoring();
+        let weight = query.weight(self, scoring_enabled)?;
+        let segment_readers = self.segment_readers();
+        let fruits = executor.map(
+            |(segment_ord, segment_reader)| {
+                collect_segment(
+                    collector,
+                    weight.as_ref(),
+                    segment_ord as u32,
+                    segment_reader,
+                )
+            },
+            segment_readers.iter().enumerate(),
+        )?;
+        collector.merge_fruits(fruits)
     }
 
     /// Return the field searcher associated to a `Field`.
@@ -99,6 +181,15 @@ impl Searcher {
             .collect::<Vec<_>>();
         FieldSearcher::new(inv_index_readers)
     }
+
+    /// Summarize total space usage of this searcher.
+    pub fn space_usage(&self) -> SearcherSpaceUsage {
+        let mut space_usage = SearcherSpaceUsage::new();
+        for segment_reader in self.segment_readers.iter() {
+            space_usage.add_segment(segment_reader.space_usage());
+        }
+        space_usage
+    }
 }
 
 pub struct FieldSearcher {

src/core/segment_reader.rs

@@ -4,6 +4,7 @@ use core::InvertedIndexReader;
 use core::Segment;
 use core::SegmentComponent;
 use core::SegmentId;
+use directory::ReadOnlySource;
 use error::TantivyError;
 use fastfield::DeleteBitSet;
 use fastfield::FacetReader;
@@ -12,10 +13,10 @@ use fastfield::{self, FastFieldNotAvailableError};
 use fastfield::{BytesFastFieldReader, FastValue, MultiValueIntFastFieldReader};
 use fieldnorm::FieldNormReader;
 use schema::Cardinality;
-use schema::Document;
 use schema::Field;
 use schema::FieldType;
 use schema::Schema;
+use space_usage::SegmentSpaceUsage;
 use std::collections::HashMap;
 use std::fmt;
 use std::sync::Arc;
@@ -53,7 +54,7 @@ pub struct SegmentReader {
     fast_fields_composite: CompositeFile,
     fieldnorms_composite: CompositeFile,
 
-    store_reader: StoreReader,
+    store_source: ReadOnlySource,
     delete_bitset_opt: Option<DeleteBitSet>,
     schema: Schema,
 }
@@ -196,8 +197,7 @@ impl SegmentReader {
     /// Accessor to the segment's `Field norms`'s reader.
     ///
     /// Field norms are the length (in tokens) of the fields.
-    /// It is used in the computation of the [TfIdf]
-    /// (https://fulmicoton.gitbooks.io/tantivy-doc/content/tfidf.html).
+    /// It is used in the computation of the [TfIdf](https://fulmicoton.gitbooks.io/tantivy-doc/content/tfidf.html).
     ///
     /// They are simply stored as a fast field, serialized in
     /// the `.fieldnorm` file of the segment.
@@ -215,8 +215,8 @@ impl SegmentReader {
     }
 
     /// Accessor to the segment's `StoreReader`.
-    pub fn get_store_reader(&self) -> &StoreReader {
-        &self.store_reader
+    pub fn get_store_reader(&self) -> StoreReader {
+        StoreReader::from_source(self.store_source.clone())
     }
 
     /// Open a new segment for reading.
@@ -225,7 +225,6 @@ impl SegmentReader {
         let termdict_composite = CompositeFile::open(&termdict_source)?;
 
         let store_source = segment.open_read(SegmentComponent::STORE)?;
-        let store_reader = StoreReader::from_source(store_source);
 
         fail_point!("SegmentReader::open#middle");
 
@@ -271,7 +270,7 @@ impl SegmentReader {
             fast_fields_composite,
             fieldnorms_composite,
             segment_id: segment.id(),
-            store_reader,
+            store_source,
             delete_bitset_opt,
             positions_composite,
             positions_idx_composite,
@@ -350,14 +349,6 @@ impl SegmentReader {
         inv_idx_reader
     }
 
-    /// Returns the document (or to be accurate, its stored field)
-    /// bearing the given doc id.
-    /// This method is slow and should seldom be called from
-    /// within a collector.
-    pub fn doc(&self, doc_id: DocId) -> Result<Document> {
-        self.store_reader.get(doc_id)
-    }
-
     /// Returns the segment id
     pub fn segment_id(&self) -> SegmentId {
         self.segment_id
@@ -381,6 +372,24 @@ impl SegmentReader {
     pub fn doc_ids_alive(&self) -> SegmentReaderAliveDocsIterator {
         SegmentReaderAliveDocsIterator::new(&self)
     }
+
+    /// Summarize total space usage of this segment.
+    pub fn space_usage(&self) -> SegmentSpaceUsage {
+        SegmentSpaceUsage::new(
+            self.num_docs(),
+            self.termdict_composite.space_usage(),
+            self.postings_composite.space_usage(),
+            self.positions_composite.space_usage(),
+            self.positions_idx_composite.space_usage(),
+            self.fast_fields_composite.space_usage(),
+            self.fieldnorms_composite.space_usage(),
+            self.get_store_reader().space_usage(),
+            self.delete_bitset_opt
+                .as_ref()
+                .map(|x| x.space_usage())
+                .unwrap_or(0),
+        )
+    }
 }
 
 impl fmt::Debug for SegmentReader {
@@ -438,12 +447,12 @@ impl<'a> Iterator for SegmentReaderAliveDocsIterator<'a> {
 #[cfg(test)]
 mod test {
     use core::Index;
-    use schema::{SchemaBuilder, Term, STORED, TEXT};
+    use schema::{Schema, Term, STORED, TEXT};
     use DocId;
 
     #[test]
     fn test_alive_docs_iterator() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         schema_builder.add_text_field("name", TEXT | STORED);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema.clone());

src/directory/managed_directory.rs

@@ -1,7 +1,7 @@
 use core::MANAGED_FILEPATH;
 use directory::error::{DeleteError, IOError, OpenReadError, OpenWriteError};
 use directory::{ReadOnlySource, WritePtr};
-use error::TantivyError;
+use error::DataCorruption;
 use indexer::LockType;
 use serde_json;
 use std::collections::HashSet;
@@ -59,12 +59,17 @@ fn save_managed_paths(
 
 impl ManagedDirectory {
     /// Wraps a directory as managed directory.
-    pub fn new<Dir: Directory>(directory: Dir) -> Result<ManagedDirectory> {
+    pub fn wrap<Dir: Directory>(directory: Dir) -> Result<ManagedDirectory> {
         match directory.atomic_read(&MANAGED_FILEPATH) {
             Ok(data) => {
                 let managed_files_json = String::from_utf8_lossy(&data);
                 let managed_files: HashSet<PathBuf> = serde_json::from_str(&managed_files_json)
-                    .map_err(|_| TantivyError::CorruptedFile(MANAGED_FILEPATH.clone()))?;
+                    .map_err(|e| {
+                        DataCorruption::new(
+                            MANAGED_FILEPATH.clone(),
+                            format!("Managed file cannot be deserialized: {:?}. ", e),
+                        )
+                    })?;
                 Ok(ManagedDirectory {
                     directory: Box::new(directory),
                     meta_informations: Arc::new(RwLock::new(MetaInformation {
@@ -260,7 +265,7 @@ mod tests {
         let tempdir_path = PathBuf::from(tempdir.path());
         {
             let mmap_directory = MmapDirectory::open(&tempdir_path).unwrap();
-            let mut managed_directory = ManagedDirectory::new(mmap_directory).unwrap();
+            let mut managed_directory = ManagedDirectory::wrap(mmap_directory).unwrap();
             {
                 let mut write_file = managed_directory.open_write(*TEST_PATH1).unwrap();
                 write_file.flush().unwrap();
@@ -286,7 +291,7 @@ mod tests {
         }
         {
             let mmap_directory = MmapDirectory::open(&tempdir_path).unwrap();
-            let mut managed_directory = ManagedDirectory::new(mmap_directory).unwrap();
+            let mut managed_directory = ManagedDirectory::wrap(mmap_directory).unwrap();
             {
                 assert!(managed_directory.exists(*TEST_PATH1));
                 assert!(!managed_directory.exists(*TEST_PATH2));
@@ -310,7 +315,7 @@ mod tests {
         let living_files = HashSet::new();
 
         let mmap_directory = MmapDirectory::open(&tempdir_path).unwrap();
-        let mut managed_directory = ManagedDirectory::new(mmap_directory).unwrap();
+        let mut managed_directory = ManagedDirectory::wrap(mmap_directory).unwrap();
         managed_directory
             .atomic_write(*TEST_PATH1, &vec![0u8, 1u8])
             .unwrap();

src/directory/mmap_directory.rs

@@ -364,6 +364,11 @@ mod tests {
 
     use super::*;
 
+    #[test]
+    fn test_open_non_existant_path() {
+        assert!(MmapDirectory::open(PathBuf::from("./nowhere")).is_err());
+    }
+
     #[test]
     fn test_open_empty() {
         // empty file is actually an edge case because those

src/directory/mod.rs

@@ -12,7 +12,6 @@ mod managed_directory;
 mod ram_directory;
 mod read_only_source;
 mod shared_vec_slice;
-mod static_dictionnary;
 
 /// Errors specific to the directory module.
 pub mod error;
@@ -22,7 +21,6 @@ use std::io::{BufWriter, Seek, Write};
 pub use self::directory::{Directory, DirectoryClone};
 pub use self::ram_directory::RAMDirectory;
 pub use self::read_only_source::ReadOnlySource;
-pub use self::static_dictionnary::StaticDirectory;
 
 #[cfg(feature = "mmap")]
 pub use self::mmap_directory::MmapDirectory;

src/directory/ram_directory.rs

@@ -100,7 +100,8 @@ impl InnerDirectory {
                 );
                 let io_err = make_io_err(msg);
                 OpenReadError::IOError(IOError::with_path(path.to_owned(), io_err))
-            }).and_then(|readable_map| {
+            })
+            .and_then(|readable_map| {
                 readable_map
                     .get(path)
                     .ok_or_else(|| OpenReadError::FileDoesNotExist(PathBuf::from(path)))
@@ -120,7 +121,8 @@ impl InnerDirectory {
                 );
                 let io_err = make_io_err(msg);
                 DeleteError::IOError(IOError::with_path(path.to_owned(), io_err))
-            }).and_then(|mut writable_map| match writable_map.remove(path) {
+            })
+            .and_then(|mut writable_map| match writable_map.remove(path) {
                 Some(_) => Ok(()),
                 None => Err(DeleteError::FileDoesNotExist(PathBuf::from(path))),
             })

src/directory/read_only_source.rs

@@ -5,9 +5,6 @@ use fst::raw::MmapReadOnly;
 use stable_deref_trait::{CloneStableDeref, StableDeref};
 use std::ops::Deref;
 
-
-const EMPTY_SLICE: [u8; 0] = [];
-
 /// Read object that represents files in tantivy.
 ///
 /// These read objects are only in charge to deliver
@@ -20,8 +17,6 @@ pub enum ReadOnlySource {
     Mmap(MmapReadOnly),
     /// Wrapping a `Vec<u8>`
     Anonymous(SharedVecSlice),
-    /// Wrapping a static slice
-    Static(&'static [u8])
 }
 
 unsafe impl StableDeref for ReadOnlySource {}
@@ -38,7 +33,7 @@ impl Deref for ReadOnlySource {
 impl ReadOnlySource {
     /// Creates an empty ReadOnlySource
     pub fn empty() -> ReadOnlySource {
-        ReadOnlySource::Static(&EMPTY_SLICE)
+        ReadOnlySource::Anonymous(SharedVecSlice::empty())
     }
 
     /// Returns the data underlying the ReadOnlySource object.
@@ -47,7 +42,6 @@ impl ReadOnlySource {
             #[cfg(feature = "mmap")]
             ReadOnlySource::Mmap(ref mmap_read_only) => mmap_read_only.as_slice(),
             ReadOnlySource::Anonymous(ref shared_vec) => shared_vec.as_slice(),
-            ReadOnlySource::Static(data) => data,
         }
     }
 
@@ -85,9 +79,6 @@ impl ReadOnlySource {
             ReadOnlySource::Anonymous(ref shared_vec) => {
                 ReadOnlySource::Anonymous(shared_vec.slice(from_offset, to_offset))
             }
-            ReadOnlySource::Static(data) => {
-                ReadOnlySource::Static(&data[from_offset..to_offset])
-            }
         }
     }
 
@@ -127,9 +118,3 @@ impl From<Vec<u8>> for ReadOnlySource {
         ReadOnlySource::Anonymous(shared_data)
     }
 }
-
-impl From<&'static [u8]> for ReadOnlySource {
-    fn from(data: &'static [u8]) -> ReadOnlySource {
-        ReadOnlySource::Static(data)
-    }
-}

src/error.rs

@@ -8,18 +8,54 @@ use indexer::LockType;
 use query;
 use schema;
 use serde_json;
+use std::fmt;
 use std::path::PathBuf;
 use std::sync::PoisonError;
 
+pub struct DataCorruption {
+    filepath: Option<PathBuf>,
+    comment: String,
+}
+
+impl DataCorruption {
+    pub fn new(filepath: PathBuf, comment: String) -> DataCorruption {
+        DataCorruption {
+            filepath: Some(filepath),
+            comment,
+        }
+    }
+
+    pub fn comment_only(comment: String) -> DataCorruption {
+        DataCorruption {
+            filepath: None,
+            comment,
+        }
+    }
+}
+
+impl fmt::Debug for DataCorruption {
+    fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
+        write!(f, "Data corruption: ")?;
+        if let Some(ref filepath) = &self.filepath {
+            write!(f, "(in file `{:?}`)", filepath)?;
+        }
+        write!(f, ": {}.", self.comment)?;
+        Ok(())
+    }
+}
+
 /// The library's failure based error enum
 #[derive(Debug, Fail)]
 pub enum TantivyError {
     /// Path does not exist.
-    #[fail(display = "path does not exist: '{:?}'", _0)]
+    #[fail(display = "Path does not exist: '{:?}'", _0)]
     PathDoesNotExist(PathBuf),
     /// File already exists, this is a problem when we try to write into a new file.
-    #[fail(display = "file already exists: '{:?}'", _0)]
+    #[fail(display = "File already exists: '{:?}'", _0)]
     FileAlreadyExists(PathBuf),
+    /// Index already exists in this directory
+    #[fail(display = "Index already exists")]
+    IndexAlreadyExists,
     /// Failed to acquire file lock
     #[fail(
         display = "Failed to acquire Lockfile: {:?}. Possible causes: another IndexWriter instance or panic during previous lock drop.",
@@ -27,28 +63,35 @@ pub enum TantivyError {
     )]
     LockFailure(LockType),
     /// IO Error.
-    #[fail(display = "an IO error occurred: '{}'", _0)]
+    #[fail(display = "An IO error occurred: '{}'", _0)]
     IOError(#[cause] IOError),
-    /// The data within is corrupted.
-    ///
-    /// For instance, it contains invalid JSON.
-    #[fail(display = "file contains corrupted data: '{:?}'", _0)]
-    CorruptedFile(PathBuf),
+    /// Data corruption.
+    #[fail(display = "{:?}", _0)]
+    DataCorruption(DataCorruption),
     /// A thread holding the locked panicked and poisoned the lock.
-    #[fail(display = "a thread holding the locked panicked and poisoned the lock")]
+    #[fail(display = "A thread holding the locked panicked and poisoned the lock")]
     Poisoned,
     /// Invalid argument was passed by the user.
-    #[fail(display = "an invalid argument was passed: '{}'", _0)]
+    #[fail(display = "An invalid argument was passed: '{}'", _0)]
     InvalidArgument(String),
     /// An Error happened in one of the thread.
-    #[fail(display = "an error occurred in a thread: '{}'", _0)]
+    #[fail(display = "An error occurred in a thread: '{}'", _0)]
     ErrorInThread(String),
     /// An Error appeared related to the schema.
     #[fail(display = "Schema error: '{}'", _0)]
     SchemaError(String),
     /// Tried to access a fastfield reader for a field not configured accordingly.
-    #[fail(display = "fast field not available: '{:?}'", _0)]
+    #[fail(display = "Fast field not available: '{:?}'", _0)]
     FastFieldError(#[cause] FastFieldNotAvailableError),
+    /// System error. (e.g.: We failed spawning a new thread)
+    #[fail(display = "System error.'{}'", _0)]
+    SystemError(String),
+}
+
+impl From<DataCorruption> for TantivyError {
+    fn from(data_corruption: DataCorruption) -> TantivyError {
+        TantivyError::DataCorruption(data_corruption)
+    }
 }
 
 impl From<FastFieldNotAvailableError> for TantivyError {

src/fastfield/bytes/mod.rs

@@ -6,12 +6,12 @@ pub use self::writer::BytesFastFieldWriter;
 
 #[cfg(test)]
 mod tests {
-    use schema::SchemaBuilder;
+    use schema::Schema;
     use Index;
 
     #[test]
     fn test_bytes() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let field = schema_builder.add_bytes_field("bytesfield");
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);

src/fastfield/delete.rs

@@ -2,6 +2,7 @@ use bit_set::BitSet;
 use common::HasLen;
 use directory::ReadOnlySource;
 use directory::WritePtr;
+use space_usage::ByteCount;
 use std::io;
 use std::io::Write;
 use DocId;
@@ -63,6 +64,11 @@ impl DeleteBitSet {
             b & (1u8 << shift) != 0
         }
     }
+
+    /// Summarize total space usage of this bitset.
+    pub fn space_usage(&self) -> ByteCount {
+        self.data.len()
+    }
 }
 
 impl HasLen for DeleteBitSet {

src/fastfield/facet_reader.rs

@@ -1,5 +1,6 @@
 use super::MultiValueIntFastFieldReader;
 use schema::Facet;
+use std::str;
 use termdict::TermDictionary;
 use termdict::TermOrdinal;
 use DocId;
@@ -20,6 +21,7 @@ use DocId;
 pub struct FacetReader {
     term_ords: MultiValueIntFastFieldReader<u64>,
     term_dict: TermDictionary,
+    buffer: Vec<u8>,
 }
 
 impl FacetReader {
@@ -37,6 +39,7 @@ impl FacetReader {
         FacetReader {
             term_ords,
             term_dict,
+            buffer: vec![],
         }
     }
 
@@ -55,11 +58,18 @@ impl FacetReader {
     }
 
     /// Given a term ordinal returns the term associated to it.
-    pub fn facet_from_ord(&self, facet_ord: TermOrdinal, output: &mut Facet) {
+    pub fn facet_from_ord(
+        &mut self,
+        facet_ord: TermOrdinal,
+        output: &mut Facet,
+    ) -> Result<(), str::Utf8Error> {
         let found_term = self
             .term_dict
-            .ord_to_term(facet_ord as u64, output.inner_buffer_mut());
+            .ord_to_term(facet_ord as u64, &mut self.buffer);
         assert!(found_term, "Term ordinal {} no found.", facet_ord);
+        let facet_str = str::from_utf8(&self.buffer[..])?;
+        output.set_facet_str(facet_str);
+        Ok(())
     }
 
     /// Return the list of facet ordinals associated to a document.

src/fastfield/mod.rs

@@ -127,19 +127,19 @@ mod tests {
     use common::CompositeFile;
     use directory::{Directory, RAMDirectory, WritePtr};
     use fastfield::FastFieldReader;
-    use rand::Rng;
+    use rand::prelude::SliceRandom;
+    use rand::rngs::StdRng;
     use rand::SeedableRng;
-    use rand::XorShiftRng;
     use schema::Document;
     use schema::Field;
+    use schema::Schema;
     use schema::FAST;
-    use schema::{Schema, SchemaBuilder};
     use std::collections::HashMap;
     use std::path::Path;
 
     lazy_static! {
         pub static ref SCHEMA: Schema = {
-            let mut schema_builder = SchemaBuilder::default();
+            let mut schema_builder = Schema::builder();
             schema_builder.add_u64_field("field", FAST);
             schema_builder.build()
         };
@@ -298,7 +298,7 @@ mod tests {
     fn test_signed_intfastfield() {
         let path = Path::new("test");
         let mut directory: RAMDirectory = RAMDirectory::create();
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
 
         let i64_field = schema_builder.add_i64_field("field", FAST);
         let schema = schema_builder.build();
@@ -342,7 +342,7 @@ mod tests {
     fn test_signed_intfastfield_default_val() {
         let path = Path::new("test");
         let mut directory: RAMDirectory = RAMDirectory::create();
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let i64_field = schema_builder.add_i64_field("field", FAST);
         let schema = schema_builder.build();
 
@@ -367,11 +367,10 @@ mod tests {
         }
     }
 
+    // Warning: this generates the same permutation at each call
     pub fn generate_permutation() -> Vec<u64> {
-        let seed: [u8; 16] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16];
-        let mut rng = XorShiftRng::from_seed(seed);
         let mut permutation: Vec<u64> = (0u64..100_000u64).collect();
-        rng.shuffle(&mut permutation);
+        permutation.shuffle(&mut StdRng::from_seed([1u8; 32]));
         permutation
     }
 

src/fastfield/multivalued/mod.rs

@@ -9,12 +9,12 @@ mod tests {
 
     use schema::Cardinality;
     use schema::IntOptions;
-    use schema::SchemaBuilder;
+    use schema::Schema;
     use Index;
 
     #[test]
     fn test_multivalued_u64() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let field = schema_builder.add_u64_field(
             "multifield",
             IntOptions::default().set_fast(Cardinality::MultiValues),
@@ -49,7 +49,7 @@ mod tests {
 
     #[test]
     fn test_multivalued_i64() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let field = schema_builder.add_i64_field(
             "multifield",
             IntOptions::default().set_fast(Cardinality::MultiValues),

src/fastfield/multivalued/reader.rs

@@ -47,11 +47,11 @@ impl<Item: FastValue> MultiValueIntFastFieldReader<Item> {
 mod tests {
 
     use core::Index;
-    use schema::{Document, Facet, SchemaBuilder};
+    use schema::{Document, Facet, Schema};
 
     #[test]
     fn test_multifastfield_reader() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let facet_field = schema_builder.add_facet_field("facets");
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -82,20 +82,20 @@ mod tests {
 
         let mut facet = Facet::root();
         {
-            facet_reader.facet_from_ord(1, &mut facet);
+            facet_reader.facet_from_ord(1, &mut facet).unwrap();
             assert_eq!(facet, Facet::from("/category"));
         }
         {
-            facet_reader.facet_from_ord(2, &mut facet);
+            facet_reader.facet_from_ord(2, &mut facet).unwrap();
             assert_eq!(facet, Facet::from("/category/cat1"));
         }
         {
-            facet_reader.facet_from_ord(3, &mut facet);
+            facet_reader.facet_from_ord(3, &mut facet).unwrap();
             assert_eq!(format!("{}", facet), "/category/cat2");
             assert_eq!(facet, Facet::from("/category/cat2"));
         }
         {
-            facet_reader.facet_from_ord(4, &mut facet);
+            facet_reader.facet_from_ord(4, &mut facet).unwrap();
             assert_eq!(facet, Facet::from("/category/cat3"));
         }
 

src/fastfield/reader.rs

@@ -7,7 +7,7 @@ use directory::ReadOnlySource;
 use directory::{Directory, RAMDirectory, WritePtr};
 use fastfield::{FastFieldSerializer, FastFieldsWriter};
 use owning_ref::OwningRef;
-use schema::SchemaBuilder;
+use schema::Schema;
 use schema::FAST;
 use std::collections::HashMap;
 use std::marker::PhantomData;
@@ -108,7 +108,7 @@ impl<Item: FastValue> FastFieldReader<Item> {
 
 impl<Item: FastValue> From<Vec<Item>> for FastFieldReader<Item> {
     fn from(vals: Vec<Item>) -> FastFieldReader<Item> {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let field = schema_builder.add_u64_field("field", FAST);
         let schema = schema_builder.build();
         let path = Path::new("__dummy__");

src/fieldnorm/mod.rs

@@ -15,7 +15,7 @@
 //! precompute computationally expensive functions of the fieldnorm
 //! in a very short array.
 //!
-//! This trick is used by the [BM25 similarity]().
+//! This trick is used by the BM25 similarity.
 mod code;
 mod reader;
 mod serializer;

src/functional_test.rs

@@ -1,7 +1,6 @@
 use rand::thread_rng;
 use std::collections::HashSet;
 
-use rand::distributions::Range;
 use rand::Rng;
 use schema::*;
 use Index;
@@ -16,7 +15,7 @@ fn check_index_content(searcher: &Searcher, vals: &HashSet<u64>) {
 #[ignore]
 #[cfg(feature = "mmap")]
 fn test_indexing() {
-    let mut schema_builder = SchemaBuilder::default();
+    let mut schema_builder = Schema::builder();
 
     let id_field = schema_builder.add_u64_field("id", INT_INDEXED);
     let multiples_field = schema_builder.add_u64_field("multiples", INT_INDEXED);
@@ -24,7 +23,6 @@ fn test_indexing() {
 
     let index = Index::create_from_tempdir(schema).unwrap();
 
-    let universe = Range::new(0u64, 20u64);
     let mut rng = thread_rng();
 
     let mut index_writer = index.writer_with_num_threads(3, 120_000_000).unwrap();
@@ -33,7 +31,7 @@ fn test_indexing() {
     let mut uncommitted_docs: HashSet<u64> = HashSet::new();
 
     for _ in 0..200 {
-        let random_val = rng.sample(&universe);
+        let random_val = rng.gen_range(0, 20);
         if random_val == 0 {
             index_writer.commit().expect("Commit failed");
             committed_docs.extend(&uncommitted_docs);

src/indexer/delete_queue.rs

@@ -191,10 +191,7 @@ impl DeleteCursor {
         }
     }
 
-    #[cfg_attr(
-        feature = "cargo-clippy",
-        allow(clippy::wrong_self_convention)
-    )]
+    #[cfg_attr(feature = "cargo-clippy", allow(clippy::wrong_self_convention))]
     fn is_behind_opstamp(&mut self, target_opstamp: u64) -> bool {
         self.get()
             .map(|operation| operation.opstamp < target_opstamp)

src/indexer/index_writer.rs

@@ -8,7 +8,7 @@ use core::SegmentComponent;
 use core::SegmentId;
 use core::SegmentMeta;
 use core::SegmentReader;
-use crossbeam_channel as channel;
+use crossbeam::channel;
 use docset::DocSet;
 use error::TantivyError;
 use fastfield::write_delete_bitset;
@@ -61,7 +61,8 @@ fn initial_table_size(per_thread_memory_budget: usize) -> usize {
                 "Per thread memory is too small: {}",
                 per_thread_memory_budget
             )
-        }).min(19) // we cap it at 512K
+        })
+        .min(19) // we cap it at 512K
 }
 
 /// `IndexWriter` is the user entry-point to add document to an index.
@@ -139,7 +140,7 @@ pub fn open_index_writer(
     let stamper = Stamper::new(current_opstamp);
 
     let segment_updater =
-        SegmentUpdater::new(index.clone(), stamper.clone(), &delete_queue.cursor())?;
+        SegmentUpdater::create(index.clone(), stamper.clone(), &delete_queue.cursor())?;
 
     let mut index_writer = IndexWriter {
         _directory_lock: Some(directory_lock),
@@ -388,11 +389,13 @@ impl IndexWriter {
         let mem_budget = self.heap_size_in_bytes_per_thread;
         let join_handle: JoinHandle<Result<()>> = thread::Builder::new()
             .name(format!(
-                "indexing thread {} for gen {}",
+                "thrd-tantivy-index{}-gen{}",
                 self.worker_id, generation
-            )).spawn(move || {
+            ))
+            .spawn(move || {
                 loop {
-                    let mut document_iterator = document_receiver_clone.clone().peekable();
+                    let mut document_iterator =
+                        document_receiver_clone.clone().into_iter().peekable();
 
                     // the peeking here is to avoid
                     // creating a new segment's files
@@ -464,10 +467,8 @@ impl IndexWriter {
     ///
     /// Returns the former segment_ready channel.
     fn recreate_document_channel(&mut self) -> DocumentReceiver {
-        let (mut document_sender, mut document_receiver): (
-            DocumentSender,
-            DocumentReceiver,
-        ) = channel::bounded(PIPELINE_MAX_SIZE_IN_DOCS);
+        let (mut document_sender, mut document_receiver): (DocumentSender, DocumentReceiver) =
+            channel::bounded(PIPELINE_MAX_SIZE_IN_DOCS);
         swap(&mut self.document_sender, &mut document_sender);
         swap(&mut self.document_receiver, &mut document_receiver);
         document_receiver
@@ -640,7 +641,10 @@ impl IndexWriter {
     pub fn add_document(&mut self, document: Document) -> u64 {
         let opstamp = self.stamper.stamp();
         let add_operation = AddOperation { opstamp, document };
-        self.document_sender.send(add_operation);
+        let send_result = self.document_sender.send(add_operation);
+        if let Err(e) = send_result {
+            panic!("Failed to index document. Sending to indexing channel failed. This probably means all of the indexing threads have panicked. {:?}", e);
+        }
         opstamp
     }
 }
@@ -657,7 +661,7 @@ mod tests {
 
     #[test]
     fn test_lockfile_stops_duplicates() {
-        let schema_builder = schema::SchemaBuilder::default();
+        let schema_builder = schema::Schema::builder();
         let index = Index::create_in_ram(schema_builder.build());
         let _index_writer = index.writer(40_000_000).unwrap();
         match index.writer(40_000_000) {
@@ -668,7 +672,7 @@ mod tests {
 
     #[test]
     fn test_lockfile_already_exists_error_msg() {
-        let schema_builder = schema::SchemaBuilder::default();
+        let schema_builder = schema::Schema::builder();
         let index = Index::create_in_ram(schema_builder.build());
         let _index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
         match index.writer_with_num_threads(1, 3_000_000) {
@@ -683,7 +687,7 @@ mod tests {
 
     #[test]
     fn test_set_merge_policy() {
-        let schema_builder = schema::SchemaBuilder::default();
+        let schema_builder = schema::Schema::builder();
         let index = Index::create_in_ram(schema_builder.build());
         let index_writer = index.writer(40_000_000).unwrap();
         assert_eq!(
@@ -701,7 +705,7 @@ mod tests {
 
     #[test]
     fn test_lockfile_released_on_drop() {
-        let schema_builder = schema::SchemaBuilder::default();
+        let schema_builder = schema::Schema::builder();
         let index = Index::create_in_ram(schema_builder.build());
         {
             let _index_writer = index.writer(40_000_000).unwrap();
@@ -713,7 +717,7 @@ mod tests {
 
     #[test]
     fn test_commit_and_rollback() {
-        let mut schema_builder = schema::SchemaBuilder::default();
+        let mut schema_builder = schema::Schema::builder();
         let text_field = schema_builder.add_text_field("text", schema::TEXT);
         let index = Index::create_in_ram(schema_builder.build());
 
@@ -747,7 +751,7 @@ mod tests {
 
     #[test]
     fn test_with_merges() {
-        let mut schema_builder = schema::SchemaBuilder::default();
+        let mut schema_builder = schema::Schema::builder();
         let text_field = schema_builder.add_text_field("text", schema::TEXT);
         let index = Index::create_in_ram(schema_builder.build());
         let num_docs_containing = |s: &str| {
@@ -784,7 +788,7 @@ mod tests {
 
     #[test]
     fn test_prepare_with_commit_message() {
-        let mut schema_builder = schema::SchemaBuilder::default();
+        let mut schema_builder = schema::Schema::builder();
         let text_field = schema_builder.add_text_field("text", schema::TEXT);
         let index = Index::create_in_ram(schema_builder.build());
 
@@ -818,7 +822,7 @@ mod tests {
 
     #[test]
     fn test_prepare_but_rollback() {
-        let mut schema_builder = schema::SchemaBuilder::default();
+        let mut schema_builder = schema::Schema::builder();
         let text_field = schema_builder.add_text_field("text", schema::TEXT);
         let index = Index::create_in_ram(schema_builder.build());
 
@@ -866,7 +870,7 @@ mod tests {
     #[test]
     fn test_write_commit_fails() {
         use fail;
-        let mut schema_builder = schema::SchemaBuilder::default();
+        let mut schema_builder = schema::Schema::builder();
         let text_field = schema_builder.add_text_field("text", schema::TEXT);
         let index = Index::create_in_ram(schema_builder.build());
 

src/indexer/merger.rs

@@ -40,13 +40,15 @@ fn compute_total_num_tokens(readers: &[SegmentReader], field: Field) -> u64 {
             total_tokens += reader.inverted_index(field).total_num_tokens();
         }
     }
-    total_tokens + count
-        .iter()
-        .cloned()
-        .enumerate()
-        .map(|(fieldnorm_ord, count)| {
-            count as u64 * u64::from(FieldNormReader::id_to_fieldnorm(fieldnorm_ord as u8))
-        }).sum::<u64>()
+    total_tokens
+        + count
+            .iter()
+            .cloned()
+            .enumerate()
+            .map(|(fieldnorm_ord, count)| {
+                count as u64 * u64::from(FieldNormReader::id_to_fieldnorm(fieldnorm_ord as u8))
+            })
+            .sum::<u64>()
 }
 
 pub struct IndexMerger {
@@ -523,7 +525,8 @@ impl IndexMerger {
                         }
                     }
                     None
-                }).collect();
+                })
+                .collect();
 
             // At this point, `segment_postings` contains the posting list
             // of all of the segments containing the given term.
@@ -614,7 +617,7 @@ impl IndexMerger {
                     store_writer.store(&doc)?;
                 }
             } else {
-                store_writer.stack(store_reader)?;
+                store_writer.stack(&store_reader)?;
             }
         }
         Ok(())
@@ -635,10 +638,9 @@ impl SerializableSegment for IndexMerger {
 #[cfg(test)]
 mod tests {
     use byteorder::{BigEndian, ReadBytesExt, WriteBytesExt};
-    use collector::chain;
     use collector::tests::TestCollector;
     use collector::tests::{BytesFastFieldTestCollector, FastFieldTestCollector};
-    use collector::FacetCollector;
+    use collector::{Count, FacetCollector};
     use core::Index;
     use futures::Future;
     use query::AllQuery;
@@ -647,6 +649,7 @@ mod tests {
     use schema;
     use schema::Cardinality;
     use schema::Document;
+    use schema::Facet;
     use schema::IndexRecordOption;
     use schema::IntOptions;
     use schema::Term;
@@ -658,13 +661,14 @@ mod tests {
 
     #[test]
     fn test_index_merger_no_deletes() {
-        let mut schema_builder = schema::SchemaBuilder::default();
+        let mut schema_builder = schema::Schema::builder();
         let text_fieldtype = schema::TextOptions::default()
             .set_indexing_options(
                 TextFieldIndexing::default()
                     .set_tokenizer("default")
                     .set_index_option(IndexRecordOption::WithFreqs),
-            ).set_stored();
+            )
+            .set_stored();
         let text_field = schema_builder.add_text_field("text", text_fieldtype);
         let score_fieldtype = schema::IntOptions::default().set_fast(Cardinality::SingleValue);
         let score_field = schema_builder.add_u64_field("score", score_fieldtype);
@@ -742,27 +746,32 @@ mod tests {
             index.load_searchers().unwrap();
             let searcher = index.searcher();
             let get_doc_ids = |terms: Vec<Term>| {
-                let mut collector = TestCollector::default();
                 let query = BooleanQuery::new_multiterms_query(terms);
-                assert!(searcher.search(&query, &mut collector).is_ok());
-                collector.docs()
+                let top_docs = searcher.search(&query, &TestCollector).unwrap();
+                top_docs.docs().to_vec()
             };
             {
                 assert_eq!(
                     get_doc_ids(vec![Term::from_field_text(text_field, "a")]),
-                    vec![1, 2, 4]
+                    vec![DocAddress(0, 1), DocAddress(0, 2), DocAddress(0, 4)]
                 );
                 assert_eq!(
                     get_doc_ids(vec![Term::from_field_text(text_field, "af")]),
-                    vec![0, 3]
+                    vec![DocAddress(0, 0), DocAddress(0, 3)]
                 );
                 assert_eq!(
                     get_doc_ids(vec![Term::from_field_text(text_field, "g")]),
-                    vec![4]
+                    vec![DocAddress(0, 4)]
                 );
                 assert_eq!(
                     get_doc_ids(vec![Term::from_field_text(text_field, "b")]),
-                    vec![0, 1, 2, 3, 4]
+                    vec![
+                        DocAddress(0, 0),
+                        DocAddress(0, 1),
+                        DocAddress(0, 2),
+                        DocAddress(0, 3),
+                        DocAddress(0, 4)
+                    ]
                 );
             }
             {
@@ -788,17 +797,18 @@ mod tests {
             {
                 let get_fast_vals = |terms: Vec<Term>| {
                     let query = BooleanQuery::new_multiterms_query(terms);
-                    let mut collector = FastFieldTestCollector::for_field(score_field);
-                    assert!(searcher.search(&query, &mut collector).is_ok());
-                    collector.vals()
+                    searcher
+                        .search(&query, &FastFieldTestCollector::for_field(score_field))
+                        .unwrap()
                 };
                 let get_fast_vals_bytes = |terms: Vec<Term>| {
                     let query = BooleanQuery::new_multiterms_query(terms);
-                    let mut collector = BytesFastFieldTestCollector::for_field(bytes_score_field);
                     searcher
-                        .search(&query, &mut collector)
-                        .expect("failed to search");
-                    collector.vals()
+                        .search(
+                            &query,
+                            &BytesFastFieldTestCollector::for_field(bytes_score_field),
+                        )
+                        .expect("failed to search")
                 };
                 assert_eq!(
                     get_fast_vals(vec![Term::from_field_text(text_field, "a")]),
@@ -814,11 +824,12 @@ mod tests {
 
     #[test]
     fn test_index_merger_with_deletes() {
-        let mut schema_builder = schema::SchemaBuilder::default();
+        let mut schema_builder = schema::Schema::builder();
         let text_fieldtype = schema::TextOptions::default()
             .set_indexing_options(
                 TextFieldIndexing::default().set_index_option(IndexRecordOption::WithFreqs),
-            ).set_stored();
+            )
+            .set_stored();
         let text_field = schema_builder.add_text_field("text", text_fieldtype);
         let score_fieldtype = schema::IntOptions::default().set_fast(Cardinality::SingleValue);
         let score_field = schema_builder.add_u64_field("score", score_fieldtype);
@@ -827,21 +838,13 @@ mod tests {
         let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
 
         let search_term = |searcher: &Searcher, term: Term| {
-            let mut collector = FastFieldTestCollector::for_field(score_field);
-            let mut bytes_collector = BytesFastFieldTestCollector::for_field(bytes_score_field);
+            let collector = FastFieldTestCollector::for_field(score_field);
+            let bytes_collector = BytesFastFieldTestCollector::for_field(bytes_score_field);
             let term_query = TermQuery::new(term, IndexRecordOption::Basic);
-
-            {
-                let mut combined_collector =
-                    chain().push(&mut collector).push(&mut bytes_collector);
-                searcher
-                    .search(&term_query, &mut combined_collector)
-                    .unwrap();
-            }
-
-            let scores = collector.vals();
-
-            let mut score_bytes = Cursor::new(bytes_collector.vals());
+            let (scores, bytes) = searcher
+                .search(&term_query, &(collector, bytes_collector))
+                .unwrap();
+            let mut score_bytes = Cursor::new(bytes);
             for &score in &scores {
                 assert_eq!(score as u32, score_bytes.read_u32::<BigEndian>().unwrap());
             }
@@ -854,21 +857,21 @@ mod tests {
         {
             // a first commit
             index_writer.add_document(doc!(
-                    text_field => "a b d",
-                    score_field => 1u64,
-                    bytes_score_field => vec![0u8, 0, 0, 1],
-                ));
+                text_field => "a b d",
+                score_field => 1u64,
+                bytes_score_field => vec![0u8, 0, 0, 1],
+            ));
             index_writer.add_document(doc!(
-                    text_field => "b c",
-                    score_field => 2u64,
-                    bytes_score_field => vec![0u8, 0, 0, 2],
-                ));
+                text_field => "b c",
+                score_field => 2u64,
+                bytes_score_field => vec![0u8, 0, 0, 2],
+            ));
             index_writer.delete_term(Term::from_field_text(text_field, "c"));
             index_writer.add_document(doc!(
-                    text_field => "c d",
-                    score_field => 3u64,
-                    bytes_score_field => vec![0u8, 0, 0, 3],
-                ));
+                text_field => "c d",
+                score_field => 3u64,
+                bytes_score_field => vec![0u8, 0, 0, 3],
+            ));
             index_writer.commit().expect("committed");
             index.load_searchers().unwrap();
             let ref searcher = *index.searcher();
@@ -895,37 +898,37 @@ mod tests {
         {
             // a second commit
             index_writer.add_document(doc!(
-                    text_field => "a d e",
-                    score_field => 4_000u64,
-                    bytes_score_field => vec![0u8, 0, 0, 4],
-                ));
+                text_field => "a d e",
+                score_field => 4_000u64,
+                bytes_score_field => vec![0u8, 0, 0, 4],
+            ));
             index_writer.add_document(doc!(
-                    text_field => "e f",
-                    score_field => 5_000u64,
-                    bytes_score_field => vec![0u8, 0, 0, 5],
-                ));
+                text_field => "e f",
+                score_field => 5_000u64,
+                bytes_score_field => vec![0u8, 0, 0, 5],
+            ));
             index_writer.delete_term(Term::from_field_text(text_field, "a"));
             index_writer.delete_term(Term::from_field_text(text_field, "f"));
             index_writer.add_document(doc!(
-                    text_field => "f g",
-                    score_field => 6_000u64,
-                    bytes_score_field => vec![0u8, 0, 23, 112],
-                ));
+                text_field => "f g",
+                score_field => 6_000u64,
+                bytes_score_field => vec![0u8, 0, 23, 112],
+            ));
             index_writer.add_document(doc!(
-                    text_field => "g h",
-                    score_field => 7_000u64,
-                    bytes_score_field => vec![0u8, 0, 27, 88],
-                ));
+                text_field => "g h",
+                score_field => 7_000u64,
+                bytes_score_field => vec![0u8, 0, 27, 88],
+            ));
             index_writer.commit().expect("committed");
             index.load_searchers().unwrap();
             let searcher = index.searcher();
 
             assert_eq!(searcher.segment_readers().len(), 2);
             assert_eq!(searcher.num_docs(), 3);
-            assert_eq!(searcher.segment_readers()[0].num_docs(), 1);
-            assert_eq!(searcher.segment_readers()[0].max_doc(), 3);
-            assert_eq!(searcher.segment_readers()[1].num_docs(), 2);
-            assert_eq!(searcher.segment_readers()[1].max_doc(), 4);
+            assert_eq!(searcher.segment_readers()[0].num_docs(), 2);
+            assert_eq!(searcher.segment_readers()[0].max_doc(), 4);
+            assert_eq!(searcher.segment_readers()[1].num_docs(), 1);
+            assert_eq!(searcher.segment_readers()[1].max_doc(), 3);
             assert_eq!(
                 search_term(&searcher, Term::from_field_text(text_field, "a")),
                 empty_vec
@@ -959,15 +962,15 @@ mod tests {
                 .segment_reader(0)
                 .fast_field_reader::<u64>(score_field)
                 .unwrap();
-            assert_eq!(score_field_reader.min_value(), 1);
-            assert_eq!(score_field_reader.max_value(), 3);
+            assert_eq!(score_field_reader.min_value(), 4000);
+            assert_eq!(score_field_reader.max_value(), 7000);
 
             let score_field_reader = searcher
                 .segment_reader(1)
                 .fast_field_reader::<u64>(score_field)
                 .unwrap();
-            assert_eq!(score_field_reader.min_value(), 4000);
-            assert_eq!(score_field_reader.max_value(), 7000);
+            assert_eq!(score_field_reader.min_value(), 1);
+            assert_eq!(score_field_reader.max_value(), 3);
         }
         {
             // merging the segments
@@ -1140,10 +1143,9 @@ mod tests {
 
     #[test]
     fn test_merge_facets() {
-        let mut schema_builder = schema::SchemaBuilder::default();
+        let mut schema_builder = schema::Schema::builder();
         let facet_field = schema_builder.add_facet_field("facet");
         let index = Index::create_in_ram(schema_builder.build());
-        use schema::Facet;
         {
             let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             let index_doc = |index_writer: &mut IndexWriter, doc_facets: &[&str]| {
@@ -1172,20 +1174,16 @@ mod tests {
             index_doc(&mut index_writer, &["/top/e", "/top/f"]);
             index_writer.commit().expect("committed");
         }
+
         index.load_searchers().unwrap();
         let test_searcher = |expected_num_docs: usize, expected: &[(&str, u64)]| {
             let searcher = index.searcher();
             let mut facet_collector = FacetCollector::for_field(facet_field);
             facet_collector.add_facet(Facet::from("/top"));
-            use collector::{CountCollector, MultiCollector};
-            let mut count_collector = CountCollector::default();
-            {
-                let mut multi_collectors =
-                    MultiCollector::from(vec![&mut count_collector, &mut facet_collector]);
-                searcher.search(&AllQuery, &mut multi_collectors).unwrap();
-            }
-            assert_eq!(count_collector.count(), expected_num_docs);
-            let facet_counts = facet_collector.harvest();
+            let (count, facet_counts) = searcher
+                .search(&AllQuery, &(Count, facet_collector))
+                .unwrap();
+            assert_eq!(count, expected_num_docs);
             let facets: Vec<(String, u64)> = facet_counts
                 .get("/top")
                 .map(|(facet, count)| (facet.to_string(), count))
@@ -1209,7 +1207,6 @@ mod tests {
                 ("/top/f", 1),
             ],
         );
-
         // Merging the segments
         {
             let segment_ids = index
@@ -1222,7 +1219,6 @@ mod tests {
                 .wait()
                 .expect("Merging failed");
             index_writer.wait_merging_threads().unwrap();
-
             index.load_searchers().unwrap();
             test_searcher(
                 11,
@@ -1261,7 +1257,7 @@ mod tests {
 
     #[test]
     fn test_merge_multivalued_int_fields_all_deleted() {
-        let mut schema_builder = schema::SchemaBuilder::default();
+        let mut schema_builder = schema::Schema::builder();
         let int_options = IntOptions::default()
             .set_fast(Cardinality::MultiValues)
             .set_indexed();
@@ -1302,7 +1298,7 @@ mod tests {
 
     #[test]
     fn test_merge_multivalued_int_fields() {
-        let mut schema_builder = schema::SchemaBuilder::default();
+        let mut schema_builder = schema::Schema::builder();
         let int_options = IntOptions::default()
             .set_fast(Cardinality::MultiValues)
             .set_indexed();
@@ -1368,15 +1364,17 @@ mod tests {
             assert_eq!(&vals, &[17]);
         }
 
-        {
-            let segment = searcher.segment_reader(1u32);
-            let ff_reader = segment.multi_fast_field_reader(int_field).unwrap();
-            ff_reader.get_vals(0, &mut vals);
-            assert_eq!(&vals, &[20]);
-        }
+        println!(
+            "{:?}",
+            searcher
+                .segment_readers()
+                .iter()
+                .map(|reader| reader.max_doc())
+                .collect::<Vec<_>>()
+        );
 
         {
-            let segment = searcher.segment_reader(2u32);
+            let segment = searcher.segment_reader(1u32);
             let ff_reader = segment.multi_fast_field_reader(int_field).unwrap();
             ff_reader.get_vals(0, &mut vals);
             assert_eq!(&vals, &[28, 27]);
@@ -1385,6 +1383,13 @@ mod tests {
             assert_eq!(&vals, &[1_000]);
         }
 
+        {
+            let segment = searcher.segment_reader(2u32);
+            let ff_reader = segment.multi_fast_field_reader(int_field).unwrap();
+            ff_reader.get_vals(0, &mut vals);
+            assert_eq!(&vals, &[20]);
+        }
+
         // Merging the segments
         {
             let segment_ids = index
@@ -1403,6 +1408,14 @@ mod tests {
 
         {
             let searcher = index.searcher();
+            println!(
+                "{:?}",
+                searcher
+                    .segment_readers()
+                    .iter()
+                    .map(|reader| reader.max_doc())
+                    .collect::<Vec<_>>()
+            );
             let segment = searcher.segment_reader(0u32);
             let ff_reader = segment.multi_fast_field_reader(int_field).unwrap();
 
@@ -1428,13 +1441,13 @@ mod tests {
             assert_eq!(&vals, &[17]);
 
             ff_reader.get_vals(7, &mut vals);
-            assert_eq!(&vals, &[20]);
-
-            ff_reader.get_vals(8, &mut vals);
             assert_eq!(&vals, &[28, 27]);
 
-            ff_reader.get_vals(9, &mut vals);
+            ff_reader.get_vals(8, &mut vals);
             assert_eq!(&vals, &[1_000]);
+
+            ff_reader.get_vals(9, &mut vals);
+            assert_eq!(&vals, &[20]);
         }
     }
 }

src/indexer/segment_updater.rs

@@ -138,7 +138,7 @@ struct InnerSegmentUpdater {
 }
 
 impl SegmentUpdater {
-    pub fn new(
+    pub fn create(
         index: Index,
         stamper: Stamper,
         delete_cursor: &DeleteCursor,
@@ -195,7 +195,8 @@ impl SegmentUpdater {
                 segment_updater.0.segment_manager.add_segment(segment_entry);
                 segment_updater.consider_merge_options();
                 true
-            }).forget();
+            })
+            .forget();
             true
         } else {
             false
@@ -227,20 +228,38 @@ impl SegmentUpdater {
         if self.is_alive() {
             let index = &self.0.index;
             let directory = index.directory();
+            let mut commited_segment_metas = self.0.segment_manager.committed_segment_metas();
+
+            // We sort segment_readers by number of documents.
+            // This is an heuristic to make multithreading more efficient.
+            //
+            // This is not done at the searcher level because I had a strange
+            // use case in which I was dealing with a large static index,
+            // dispatched over 5 SSD drives.
+            //
+            // A `UnionDirectory` makes it possible to read from these
+            // 5 different drives and creates a meta.json on the fly.
+            // In order to optimize the throughput, it creates a lasagna of segments
+            // from the different drives.
+            //
+            // Segment 1 from disk 1, Segment 1 from disk 2, etc.
+            commited_segment_metas.sort_by_key(|segment_meta| -(segment_meta.max_doc() as i32));
             save_metas(
-                self.0.segment_manager.committed_segment_metas(),
+                commited_segment_metas,
                 index.schema(),
                 opstamp,
                 commit_message,
                 directory.box_clone().borrow_mut(),
-            ).expect("Could not save metas.");
+            )
+            .expect("Could not save metas.");
         }
     }
 
     pub fn garbage_collect_files(&self) -> Result<()> {
         self.run_async(move |segment_updater| {
             segment_updater.garbage_collect_files_exec();
-        }).wait()
+        })
+        .wait()
     }
 
     fn garbage_collect_files_exec(&self) {
@@ -262,7 +281,8 @@ impl SegmentUpdater {
                 segment_updater.garbage_collect_files_exec();
                 segment_updater.consider_merge_options();
             }
-        }).wait()
+        })
+        .wait()
     }
 
     pub fn start_merge(&self, segment_ids: &[SegmentId]) -> Result<Receiver<SegmentMeta>> {
@@ -270,7 +290,8 @@ impl SegmentUpdater {
         let segment_ids_vec = segment_ids.to_vec();
         self.run_async(move |segment_updater| {
             segment_updater.start_merge_impl(&segment_ids_vec[..])
-        }).wait()?
+        })
+        .wait()?
     }
 
     // `segment_ids` is required to be non-empty.
@@ -336,7 +357,8 @@ impl SegmentUpdater {
                     .unwrap()
                     .remove(&merging_thread_id);
                 Ok(())
-            }).expect("Failed to spawn a thread.");
+            })
+            .expect("Failed to spawn a thread.");
         self.0
             .merging_threads
             .write()
@@ -427,7 +449,8 @@ impl SegmentUpdater {
             let previous_metas = segment_updater.0.index.load_metas().unwrap();
             segment_updater.save_metas(previous_metas.opstamp, previous_metas.payload);
             segment_updater.garbage_collect_files_exec();
-        }).wait()
+        })
+        .wait()
     }
 
     /// Wait for current merging threads.
@@ -484,7 +507,7 @@ mod tests {
 
     #[test]
     fn test_delete_during_merge() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
 

src/indexer/segment_writer.rs

@@ -62,7 +62,8 @@ impl SegmentWriter {
                             segment.index().tokenizers().get(tokenizer_name)
                         }),
                     _ => None,
-                }).collect();
+                })
+                .collect();
         Ok(SegmentWriter {
             max_doc: 0,
             multifield_postings,
@@ -110,18 +111,18 @@ impl SegmentWriter {
             }
             match *field_options.field_type() {
                 FieldType::HierarchicalFacet => {
-                    let facets: Vec<&[u8]> = field_values
+                    let facets: Vec<&str> = field_values
                         .iter()
                         .flat_map(|field_value| match *field_value.value() {
-                            Value::Facet(ref facet) => Some(facet.encoded_bytes()),
+                            Value::Facet(ref facet) => Some(facet.encoded_str()),
                             _ => {
                                 panic!("Expected hierarchical facet");
                             }
-                        }).collect();
+                        })
+                        .collect();
                     let mut term = Term::for_field(field); // we set the Term
-                    for facet_bytes in facets {
+                    for fake_str in facets {
                         let mut unordered_term_id_opt = None;
-                        let fake_str = unsafe { str::from_utf8_unchecked(facet_bytes) };
                         FacetTokenizer.token_stream(fake_str).process(&mut |token| {
                             term.set_text(&token.text);
                             let unordered_term_id =
@@ -145,7 +146,8 @@ impl SegmentWriter {
                             .flat_map(|field_value| match *field_value.value() {
                                 Value::Str(ref text) => Some(text.as_str()),
                                 _ => None,
-                            }).collect();
+                            })
+                            .collect();
                         if texts.is_empty() {
                             0
                         } else {

src/lib.rs

@@ -1,6 +1,5 @@
 #![doc(html_logo_url = "http://fulmicoton.com/tantivy-logo/tantivy-logo.png")]
 #![cfg_attr(all(feature = "unstable", test), feature(test))]
-#![cfg_attr(feature = "cargo-clippy", feature(tool_lints))]
 #![cfg_attr(feature = "cargo-clippy", allow(clippy::module_inception))]
 #![doc(test(attr(allow(unused_variables), deny(warnings))))]
 #![warn(missing_docs)]
@@ -24,7 +23,8 @@
 //! # use tempdir::TempDir;
 //! # use tantivy::Index;
 //! # use tantivy::schema::*;
-//! # use tantivy::collector::TopCollector;
+//! # use tantivy::{Score, DocAddress};
+//! # use tantivy::collector::TopDocs;
 //! # use tantivy::query::QueryParser;
 //! #
 //! # fn main() {
@@ -46,7 +46,7 @@
 //! // in a compressed, row-oriented key-value store.
 //! // This store is useful to reconstruct the
 //! // documents that were selected during the search phase.
-//! let mut schema_builder = SchemaBuilder::default();
+//! let mut schema_builder = Schema::builder();
 //! let title = schema_builder.add_text_field("title", TEXT | STORED);
 //! let body = schema_builder.add_text_field("body", TEXT);
 //! let schema = schema_builder.build();
@@ -86,13 +86,13 @@
 //! // A ticket has been opened regarding this problem.
 //! let query = query_parser.parse_query("sea whale")?;
 //!
-//! let mut top_collector = TopCollector::with_limit(10);
-//! searcher.search(&*query, &mut top_collector)?;
+//! // Perform search.
+//! // `topdocs` contains the 10 most relevant doc ids, sorted by decreasing scores...
+//! let top_docs: Vec<(Score, DocAddress)> =
+//!     searcher.search(&query, &TopDocs::with_limit(10))?;
 //!
-//! // Our top collector now contains the 10
-//! // most relevant doc ids...
-//! let doc_addresses = top_collector.docs();
-//! for doc_address in doc_addresses {
+//! for (_score, doc_address) in top_docs {
+//!     // Retrieve the actual content of documents given its `doc_address`.
 //!     let retrieved_doc = searcher.doc(doc_address)?;
 //!     println!("{}", schema.to_json(&retrieved_doc));
 //! }
@@ -129,14 +129,14 @@ extern crate base64;
 extern crate bit_set;
 extern crate bitpacking;
 extern crate byteorder;
+extern crate scoped_pool;
 
 extern crate combine;
 
 extern crate crossbeam;
-extern crate crossbeam_channel;
 extern crate fnv;
 extern crate fst;
-
+extern crate fst_regex;
 extern crate futures;
 extern crate futures_cpupool;
 extern crate htmlescape;
@@ -152,8 +152,6 @@ extern crate tempdir;
 extern crate tempfile;
 extern crate uuid;
 
-
-
 #[cfg(test)]
 #[macro_use]
 extern crate matches;
@@ -185,10 +183,7 @@ mod macros;
 
 pub use error::TantivyError;
 
-#[deprecated(
-    since = "0.7.0",
-    note = "please use `tantivy::TantivyError` instead"
-)]
+#[deprecated(since = "0.7.0", note = "please use `tantivy::TantivyError` instead")]
 pub use error::TantivyError as Error;
 
 extern crate census;
@@ -213,11 +208,12 @@ pub(crate) mod positions;
 pub mod postings;
 pub mod query;
 pub mod schema;
+pub mod space_usage;
 pub mod store;
 pub mod termdict;
 
 mod snippet;
-pub use self::snippet::SnippetGenerator;
+pub use self::snippet::{Snippet, SnippetGenerator};
 
 mod docset;
 pub use self::docset::{DocSet, SkipResult};
@@ -300,9 +296,11 @@ mod tests {
     use docset::DocSet;
     use query::BooleanQuery;
     use rand::distributions::Bernoulli;
-    use rand::distributions::Range;
-    use rand::{Rng, SeedableRng, XorShiftRng};
+    use rand::distributions::Uniform;
+    use rand::rngs::StdRng;
+    use rand::{Rng, SeedableRng};
     use schema::*;
+    use DocAddress;
     use Index;
     use IndexWriter;
     use Postings;
@@ -321,16 +319,15 @@ mod tests {
     }
 
     pub fn generate_nonunique_unsorted(max_value: u32, n_elems: usize) -> Vec<u32> {
-        let seed: [u8; 16] = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15];
-        XorShiftRng::from_seed(seed)
-            .sample_iter(&Range::new(0u32, max_value))
+        let seed: [u8; 32] = [1; 32];
+        StdRng::from_seed(seed)
+            .sample_iter(&Uniform::new(0u32, max_value))
             .take(n_elems)
             .collect::<Vec<u32>>()
     }
 
     pub fn sample_with_seed(n: u32, ratio: f64, seed_val: u8) -> Vec<u32> {
-        let seed: [u8; 16] = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, seed_val];
-        XorShiftRng::from_seed(seed)
+        StdRng::from_seed([seed_val; 32])
             .sample_iter(&Bernoulli::new(ratio))
             .take(n as usize)
             .enumerate()
@@ -345,7 +342,7 @@ mod tests {
     #[test]
     #[cfg(feature = "mmap")]
     fn test_indexing() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_from_tempdir(schema).unwrap();
@@ -370,7 +367,7 @@ mod tests {
 
     #[test]
     fn test_docfreq1() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let index = Index::create_in_ram(schema_builder.build());
         let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
@@ -410,7 +407,7 @@ mod tests {
 
     #[test]
     fn test_fieldnorm_no_docs_with_field() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let title_field = schema_builder.add_text_field("title", TEXT);
         let text_field = schema_builder.add_text_field("text", TEXT);
         let index = Index::create_in_ram(schema_builder.build());
@@ -439,7 +436,7 @@ mod tests {
 
     #[test]
     fn test_fieldnorm() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let index = Index::create_in_ram(schema_builder.build());
         {
@@ -480,7 +477,7 @@ mod tests {
 
     #[test]
     fn test_delete_postings1() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let term_abcd = Term::from_field_text(text_field, "abcd");
         let term_a = Term::from_field_text(text_field, "a");
@@ -491,42 +488,21 @@ mod tests {
         {
             // writing the segment
             let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
-            {
-                // 0
-                let doc = doc!(text_field=>"a b");
-                index_writer.add_document(doc);
-            }
-            {
-                // 1
-                let doc = doc!(text_field=>" a c");
-                index_writer.add_document(doc);
-            }
-            {
-                // 2
-                let doc = doc!(text_field=>" b c");
-                index_writer.add_document(doc);
-            }
-            {
-                // 3
-                let doc = doc!(text_field=>" b d");
-                index_writer.add_document(doc);
-            }
-            {
-                index_writer.delete_term(Term::from_field_text(text_field, "c"));
-            }
-            {
-                index_writer.delete_term(Term::from_field_text(text_field, "a"));
-            }
-            {
-                // 4
-                let doc = doc!(text_field=>" b c");
-                index_writer.add_document(doc);
-            }
-            {
-                // 5
-                let doc = doc!(text_field=>" a");
-                index_writer.add_document(doc);
-            }
+            // 0
+            index_writer.add_document(doc!(text_field=>"a b"));
+            // 1
+            index_writer.add_document(doc!(text_field=>" a c"));
+            // 2
+            index_writer.add_document(doc!(text_field=>" b c"));
+            // 3
+            index_writer.add_document(doc!(text_field=>" b d"));
+
+            index_writer.delete_term(Term::from_field_text(text_field, "c"));
+            index_writer.delete_term(Term::from_field_text(text_field, "a"));
+            // 4
+            index_writer.add_document(doc!(text_field=>" b c"));
+            // 5
+            index_writer.add_document(doc!(text_field=>" a"));
             index_writer.commit().unwrap();
         }
         {
@@ -534,11 +510,9 @@ mod tests {
             let searcher = index.searcher();
             let reader = searcher.segment_reader(0);
             let inverted_index = reader.inverted_index(text_field);
-            assert!(
-                inverted_index
-                    .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
-                    .is_none()
-            );
+            assert!(inverted_index
+                .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
+                .is_none());
             {
                 let mut postings = inverted_index
                     .read_postings(&term_a, IndexRecordOption::WithFreqsAndPositions)
@@ -561,15 +535,10 @@ mod tests {
         {
             // writing the segment
             let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
-            {
-                // 0
-                let doc = doc!(text_field=>"a b");
-                index_writer.add_document(doc);
-            }
-            {
-                // 1
-                index_writer.delete_term(Term::from_field_text(text_field, "c"));
-            }
+            // 0
+            index_writer.add_document(doc!(text_field=>"a b"));
+            // 1
+            index_writer.delete_term(Term::from_field_text(text_field, "c"));
             index_writer.rollback().unwrap();
         }
         {
@@ -578,11 +547,9 @@ mod tests {
             let reader = searcher.segment_reader(0);
             let inverted_index = reader.inverted_index(term_abcd.field());
 
-            assert!(
-                inverted_index
-                    .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
-                    .is_none()
-            );
+            assert!(inverted_index
+                .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
+                .is_none());
             {
                 let mut postings = inverted_index
                     .read_postings(&term_a, IndexRecordOption::WithFreqsAndPositions)
@@ -605,13 +572,8 @@ mod tests {
         {
             // writing the segment
             let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
-            {
-                let doc = doc!(text_field=>"a b");
-                index_writer.add_document(doc);
-            }
-            {
-                index_writer.delete_term(Term::from_field_text(text_field, "c"));
-            }
+            index_writer.add_document(doc!(text_field=>"a b"));
+            index_writer.delete_term(Term::from_field_text(text_field, "c"));
             index_writer.rollback().unwrap();
             index_writer.delete_term(Term::from_field_text(text_field, "a"));
             index_writer.commit().unwrap();
@@ -621,11 +583,9 @@ mod tests {
             let searcher = index.searcher();
             let reader = searcher.segment_reader(0);
             let inverted_index = reader.inverted_index(term_abcd.field());
-            assert!(
-                inverted_index
-                    .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
-                    .is_none()
-            );
+            assert!(inverted_index
+                .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
+                .is_none());
             {
                 let mut postings = inverted_index
                     .read_postings(&term_a, IndexRecordOption::WithFreqsAndPositions)
@@ -655,7 +615,7 @@ mod tests {
 
     #[test]
     fn test_indexed_u64() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let field = schema_builder.add_u64_field("value", INT_INDEXED);
         let schema = schema_builder.build();
 
@@ -678,7 +638,7 @@ mod tests {
 
     #[test]
     fn test_indexed_i64() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let value_field = schema_builder.add_i64_field("value", INT_INDEXED);
         let schema = schema_builder.build();
 
@@ -702,7 +662,7 @@ mod tests {
 
     #[test]
     fn test_indexedfield_not_in_documents() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let absent_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
@@ -718,7 +678,7 @@ mod tests {
 
     #[test]
     fn test_delete_postings2() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -754,7 +714,7 @@ mod tests {
 
     #[test]
     fn test_termfreq() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -773,11 +733,9 @@ mod tests {
             let reader = searcher.segment_reader(0);
             let inverted_index = reader.inverted_index(text_field);
             let term_abcd = Term::from_field_text(text_field, "abcd");
-            assert!(
-                inverted_index
-                    .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
-                    .is_none()
-            );
+            assert!(inverted_index
+                .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
+                .is_none());
             let term_af = Term::from_field_text(text_field, "af");
             let mut postings = inverted_index
                 .read_postings(&term_af, IndexRecordOption::WithFreqsAndPositions)
@@ -791,7 +749,7 @@ mod tests {
 
     #[test]
     fn test_searcher_1() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -799,18 +757,9 @@ mod tests {
         {
             // writing the segment
             let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
-            {
-                let doc = doc!(text_field=>"af af af b");
-                index_writer.add_document(doc);
-            }
-            {
-                let doc = doc!(text_field=>"a b c");
-                index_writer.add_document(doc);
-            }
-            {
-                let doc = doc!(text_field=>"a b c d");
-                index_writer.add_document(doc);
-            }
+            index_writer.add_document(doc!(text_field=>"af af af b"));
+            index_writer.add_document(doc!(text_field=>"a b c"));
+            index_writer.add_document(doc!(text_field=>"a b c d"));
             index_writer.commit().unwrap();
         }
         {
@@ -818,55 +767,42 @@ mod tests {
             let searcher = index.searcher();
             let get_doc_ids = |terms: Vec<Term>| {
                 let query = BooleanQuery::new_multiterms_query(terms);
-                let mut collector = TestCollector::default();
-                assert!(searcher.search(&query, &mut collector).is_ok());
-                collector.docs()
+                let topdocs = searcher.search(&query, &TestCollector).unwrap();
+                topdocs.docs().to_vec()
             };
-            {
-                assert_eq!(
-                    get_doc_ids(vec![Term::from_field_text(text_field, "a")]),
-                    vec![1, 2]
-                );
-            }
-            {
-                assert_eq!(
-                    get_doc_ids(vec![Term::from_field_text(text_field, "af")]),
-                    vec![0]
-                );
-            }
-            {
-                assert_eq!(
-                    get_doc_ids(vec![Term::from_field_text(text_field, "b")]),
-                    vec![0, 1, 2]
-                );
-            }
-            {
-                assert_eq!(
-                    get_doc_ids(vec![Term::from_field_text(text_field, "c")]),
-                    vec![1, 2]
-                );
-            }
-            {
-                assert_eq!(
-                    get_doc_ids(vec![Term::from_field_text(text_field, "d")]),
-                    vec![2]
-                );
-            }
-            {
-                assert_eq!(
-                    get_doc_ids(vec![
-                        Term::from_field_text(text_field, "b"),
-                        Term::from_field_text(text_field, "a"),
-                    ]),
-                    vec![0, 1, 2]
-                );
-            }
+            assert_eq!(
+                get_doc_ids(vec![Term::from_field_text(text_field, "a")]),
+                vec![DocAddress(0, 1), DocAddress(0, 2)]
+            );
+            assert_eq!(
+                get_doc_ids(vec![Term::from_field_text(text_field, "af")]),
+                vec![DocAddress(0, 0)]
+            );
+            assert_eq!(
+                get_doc_ids(vec![Term::from_field_text(text_field, "b")]),
+                vec![DocAddress(0, 0), DocAddress(0, 1), DocAddress(0, 2)]
+            );
+            assert_eq!(
+                get_doc_ids(vec![Term::from_field_text(text_field, "c")]),
+                vec![DocAddress(0, 1), DocAddress(0, 2)]
+            );
+            assert_eq!(
+                get_doc_ids(vec![Term::from_field_text(text_field, "d")]),
+                vec![DocAddress(0, 2)]
+            );
+            assert_eq!(
+                get_doc_ids(vec![
+                    Term::from_field_text(text_field, "b"),
+                    Term::from_field_text(text_field, "a"),
+                ]),
+                vec![DocAddress(0, 0), DocAddress(0, 1), DocAddress(0, 2)]
+            );
         }
     }
 
     #[test]
     fn test_searcher_2() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -893,7 +829,7 @@ mod tests {
 
     #[test]
     fn test_doc_macro() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let other_text_field = schema_builder.add_text_field("text2", TEXT);
         let document = doc!(text_field => "tantivy",
@@ -911,7 +847,7 @@ mod tests {
 
     #[test]
     fn test_wrong_fast_field_type() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let fast_field_unsigned = schema_builder.add_u64_field("unsigned", FAST);
         let fast_field_signed = schema_builder.add_i64_field("signed", FAST);
         let text_field = schema_builder.add_text_field("text", TEXT);

src/macros.rs

@@ -26,12 +26,12 @@
 /// #[macro_use]
 /// extern crate tantivy;
 ///
-/// use tantivy::schema::{SchemaBuilder, TEXT, FAST};
+/// use tantivy::schema::{Schema, TEXT, FAST};
 ///
 /// //...
 ///
 /// # fn main() {
-/// let mut schema_builder = SchemaBuilder::new();
+/// let mut schema_builder = Schema::builder();
 /// let title = schema_builder.add_text_field("title", TEXT);
 /// let author = schema_builder.add_text_field("text", TEXT);
 /// let likes = schema_builder.add_u64_field("num_u64", FAST);
@@ -67,33 +67,33 @@ macro_rules! doc(
 
 #[cfg(test)]
 mod test {
-    use schema::{SchemaBuilder, FAST, TEXT};
+    use schema::{Schema, FAST, TEXT};
 
     #[test]
     fn test_doc_basic() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let title = schema_builder.add_text_field("title", TEXT);
         let author = schema_builder.add_text_field("text", TEXT);
         let likes = schema_builder.add_u64_field("num_u64", FAST);
         let _schema = schema_builder.build();
         let _doc = doc!(
-            title => "Life Aquatic",
-            author => "Wes Anderson",
-            likes => 4u64
-            );
+        title => "Life Aquatic",
+        author => "Wes Anderson",
+        likes => 4u64
+        );
     }
 
     #[test]
     fn test_doc_trailing_comma() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let title = schema_builder.add_text_field("title", TEXT);
         let author = schema_builder.add_text_field("text", TEXT);
         let likes = schema_builder.add_u64_field("num_u64", FAST);
         let _schema = schema_builder.build();
         let _doc = doc!(
-            title => "Life Aquatic",
-            author => "Wes Anderson",
-            likes => 4u64,
-            );
+        title => "Life Aquatic",
+        author => "Wes Anderson",
+        likes => 4u64,
+        );
     }
 }

src/postings/compression/mod.rs

@@ -266,21 +266,17 @@ pub mod tests {
 mod bench {
 
     use super::*;
-    use rand::Rng;
     use rand::SeedableRng;
-    use rand::XorShiftRng;
+    use rand::{Rng, XorShiftRng};
     use test::Bencher;
 
-    fn generate_array_with_seed(n: usize, ratio: f32, seed_val: u32) -> Vec<u32> {
-        let seed: &[u32; 4] = &[1, 2, 3, seed_val];
+    fn generate_array_with_seed(n: usize, ratio: f64, seed_val: u8) -> Vec<u32> {
+        let seed: &[u8; 16] = &[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, seed_val];
         let mut rng: XorShiftRng = XorShiftRng::from_seed(*seed);
-        (0..u32::max_value())
-            .filter(|_| rng.next_f32() < ratio)
-            .take(n)
-            .collect()
+        (0u32..).filter(|_| rng.gen_bool(ratio)).take(n).collect()
     }
 
-    pub fn generate_array(n: usize, ratio: f32) -> Vec<u32> {
+    pub fn generate_array(n: usize, ratio: f64) -> Vec<u32> {
         generate_array_with_seed(n, ratio, 4)
     }
 
@@ -297,24 +293,23 @@ mod bench {
     fn bench_uncompress(b: &mut Bencher) {
         let mut encoder = BlockEncoder::new();
         let data = generate_array(COMPRESSION_BLOCK_SIZE, 0.1);
-        let (_, compressed) = encoder.compress_block_sorted(&data, 0u32);
+        let (num_bits, compressed) = encoder.compress_block_sorted(&data, 0u32);
         let mut decoder = BlockDecoder::new();
         b.iter(|| {
-            decoder.uncompress_block_sorted(compressed, 0u32);
+            decoder.uncompress_block_sorted(compressed, 0u32, num_bits);
         });
     }
 
     #[test]
     fn test_all_docs_compression_numbits() {
-        for num_bits in 0..33 {
+        for expected_num_bits in 0u8.. {
             let mut data = [0u32; 128];
-            if num_bits > 0 {
-                data[0] = 1 << (num_bits - 1);
+            if expected_num_bits > 0 {
+                data[0] = (1u64 << (expected_num_bits as usize) - 1) as u32;
             }
             let mut encoder = BlockEncoder::new();
-            let compressed = encoder.compress_block_unsorted(&data);
-            assert_eq!(compressed[0] as usize, num_bits);
-            assert_eq!(compressed.len(), compressed_block_size(compressed[0]));
+            let (num_bits, compressed) = encoder.compress_block_unsorted(&data);
+            assert_eq!(compressed.len(), compressed_block_size(num_bits));
         }
     }
 

src/postings/mod.rs

@@ -54,17 +54,18 @@ pub mod tests {
     use indexer::operation::AddOperation;
     use indexer::SegmentWriter;
     use query::Scorer;
-    use rand::{Rng, SeedableRng, XorShiftRng};
+    use rand::rngs::StdRng;
+    use rand::{Rng, SeedableRng};
     use schema::Field;
     use schema::IndexRecordOption;
-    use schema::{Document, SchemaBuilder, Term, INT_INDEXED, STRING, TEXT};
+    use schema::{Document, Schema, Term, INT_INDEXED, STRING, TEXT};
     use std::iter;
     use DocId;
     use Score;
 
     #[test]
     pub fn test_position_write() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -88,7 +89,7 @@ pub mod tests {
 
     #[test]
     pub fn test_skip_positions() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let title = schema_builder.add_text_field("title", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -163,7 +164,7 @@ pub mod tests {
     #[test]
     pub fn test_position_and_fieldnorm1() {
         let mut positions = Vec::new();
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema.clone());
@@ -220,12 +221,10 @@ pub mod tests {
             }
             {
                 let term_a = Term::from_field_text(text_field, "abcdef");
-                assert!(
-                    segment_reader
-                        .inverted_index(term_a.field())
-                        .read_postings(&term_a, IndexRecordOption::WithFreqsAndPositions)
-                        .is_none()
-                );
+                assert!(segment_reader
+                    .inverted_index(term_a.field())
+                    .read_postings(&term_a, IndexRecordOption::WithFreqsAndPositions)
+                    .is_none());
             }
             {
                 let term_a = Term::from_field_text(text_field, "a");
@@ -276,7 +275,7 @@ pub mod tests {
     #[test]
     pub fn test_position_and_fieldnorm2() {
         let mut positions: Vec<u32> = Vec::new();
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -317,7 +316,7 @@ pub mod tests {
         let num_docs = 300u32;
 
         let index = {
-            let mut schema_builder = SchemaBuilder::default();
+            let mut schema_builder = Schema::builder();
             let value_field = schema_builder.add_u64_field("value", INT_INDEXED);
             let schema = schema_builder.build();
 
@@ -498,12 +497,11 @@ pub mod tests {
             Term::from_field_text(field, "d")
         };
         pub static ref INDEX: Index = {
-            let mut schema_builder = SchemaBuilder::default();
+            let mut schema_builder = Schema::builder();
             let text_field = schema_builder.add_text_field("text", STRING);
             let schema = schema_builder.build();
 
-            let seed: [u8; 16] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16];
-            let mut rng: XorShiftRng = XorShiftRng::from_seed(seed);
+            let mut rng: StdRng = StdRng::from_seed([1u8; 32]);
 
             let index = Index::create_in_ram(schema);
             let posting_list_size = 1_000_000;
@@ -654,7 +652,7 @@ mod bench {
         });
     }
 
-    fn bench_skip_next(p: f32, b: &mut Bencher) {
+    fn bench_skip_next(p: f64, b: &mut Bencher) {
         let searcher = INDEX.searcher();
         let segment_reader = searcher.segment_reader(0);
         let docs = tests::sample(segment_reader.num_docs(), p);

src/postings/postings_writer.rs

@@ -29,7 +29,8 @@ fn posting_from_field_entry(field_entry: &FieldEntry) -> Box<PostingsWriter> {
                 IndexRecordOption::WithFreqsAndPositions => {
                     SpecializedPostingsWriter::<TFAndPositionRecorder>::new_boxed()
                 }
-            }).unwrap_or_else(|| SpecializedPostingsWriter::<NothingRecorder>::new_boxed()),
+            })
+            .unwrap_or_else(|| SpecializedPostingsWriter::<NothingRecorder>::new_boxed()),
         FieldType::U64(_) | FieldType::I64(_) | FieldType::HierarchicalFacet => {
             SpecializedPostingsWriter::<NothingRecorder>::new_boxed()
         }
@@ -107,10 +108,8 @@ impl MultiFieldPostingsWriter {
             .map(|(key, _, _)| Term::wrap(key).field())
             .enumerate();
 
-        let mut unordered_term_mappings: HashMap<
-            Field,
-            HashMap<UnorderedTermId, TermOrdinal>,
-        > = HashMap::new();
+        let mut unordered_term_mappings: HashMap<Field, HashMap<UnorderedTermId, TermOrdinal>> =
+            HashMap::new();
 
         let mut prev_field = Field(u32::max_value());
         for (offset, field) in term_offsets_it {
@@ -138,7 +137,8 @@ impl MultiFieldPostingsWriter {
                         .enumerate()
                         .map(|(term_ord, unord_term_id)| {
                             (unord_term_id as UnorderedTermId, term_ord as TermOrdinal)
-                        }).collect();
+                        })
+                        .collect();
                     unordered_term_mappings.insert(field, mapping);
                 }
                 FieldType::U64(_) | FieldType::I64(_) => {}

src/postings/segment_postings.rs

@@ -126,7 +126,6 @@ impl SegmentPostings {
 fn exponential_search(target: u32, arr: &[u32]) -> (usize, usize) {
     let mut start = 0;
     let end = arr.len();
-    debug_assert!(target >= arr[start]);
     debug_assert!(target <= arr[end - 1]);
     let mut jump = 1;
     loop {
@@ -216,11 +215,10 @@ impl DocSet for SegmentPostings {
 
         // we're in the right block now, start with an exponential search
         let block_docs = self.block_cursor.docs();
-
-        debug_assert!(target >= self.doc());
         let new_cur = self
             .cur
             .wrapping_add(search_within_block(&block_docs[self.cur..], target));
+
         if need_positions {
             sum_freqs_skipped += self.block_cursor.freqs()[self.cur..new_cur]
                 .iter()
@@ -533,7 +531,8 @@ impl BlockSegmentPostings {
                     } else {
                         BlockSegmentPostingsSkipResult::Terminated
                     }
-                }).unwrap_or(BlockSegmentPostingsSkipResult::Terminated);
+                })
+                .unwrap_or(BlockSegmentPostingsSkipResult::Terminated);
         }
         BlockSegmentPostingsSkipResult::Terminated
     }
@@ -621,6 +620,7 @@ impl<'b> Streamer<'b> for BlockSegmentPostings {
 #[cfg(test)]
 mod tests {
 
+    use super::exponential_search;
     use super::search_within_block;
     use super::BlockSegmentPostings;
     use super::BlockSegmentPostingsSkipResult;
@@ -630,10 +630,11 @@ mod tests {
     use docset::DocSet;
     use fst::Streamer;
     use schema::IndexRecordOption;
-    use schema::SchemaBuilder;
+    use schema::Schema;
     use schema::Term;
     use schema::INT_INDEXED;
     use DocId;
+    use SkipResult;
 
     #[test]
     fn test_empty_segment_postings() {
@@ -661,6 +662,16 @@ mod tests {
             .0
     }
 
+    #[test]
+    fn test_exponentiel_search() {
+        assert_eq!(exponential_search(0, &[1, 2]), (0, 1));
+        assert_eq!(exponential_search(1, &[1, 2]), (0, 1));
+        assert_eq!(
+            exponential_search(7, &[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]),
+            (3, 7)
+        );
+    }
+
     fn util_test_search_within_block(block: &[u32], target: u32) {
         assert_eq!(
             search_within_block(block, target),
@@ -692,7 +703,7 @@ mod tests {
 
     #[test]
     fn test_block_segment_postings() {
-        let mut block_segments = build_block_postings((0..100_000).collect::<Vec<u32>>());
+        let mut block_segments = build_block_postings(&(0..100_000).collect::<Vec<u32>>());
         let mut offset: u32 = 0u32;
         // checking that the block before calling advance is empty
         assert!(block_segments.docs().is_empty());
@@ -706,14 +717,44 @@ mod tests {
         }
     }
 
-    fn build_block_postings(docs: Vec<DocId>) -> BlockSegmentPostings {
-        let mut schema_builder = SchemaBuilder::default();
+    #[test]
+    fn test_skip_right_at_new_block() {
+        let mut doc_ids = (0..128).collect::<Vec<u32>>();
+        doc_ids.push(129);
+        doc_ids.push(130);
+        {
+            let block_segments = build_block_postings(&doc_ids);
+            let mut docset = SegmentPostings::from_block_postings(block_segments, None);
+            assert_eq!(docset.skip_next(128), SkipResult::OverStep);
+            assert_eq!(docset.doc(), 129);
+            assert!(docset.advance());
+            assert_eq!(docset.doc(), 130);
+            assert!(!docset.advance());
+        }
+        {
+            let block_segments = build_block_postings(&doc_ids);
+            let mut docset = SegmentPostings::from_block_postings(block_segments, None);
+            assert_eq!(docset.skip_next(129), SkipResult::Reached);
+            assert_eq!(docset.doc(), 129);
+            assert!(docset.advance());
+            assert_eq!(docset.doc(), 130);
+            assert!(!docset.advance());
+        }
+        {
+            let block_segments = build_block_postings(&doc_ids);
+            let mut docset = SegmentPostings::from_block_postings(block_segments, None);
+            assert_eq!(docset.skip_next(131), SkipResult::End);
+        }
+    }
+
+    fn build_block_postings(docs: &[DocId]) -> BlockSegmentPostings {
+        let mut schema_builder = Schema::builder();
         let int_field = schema_builder.add_u64_field("id", INT_INDEXED);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
         let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
         let mut last_doc = 0u32;
-        for doc in docs {
+        for &doc in docs {
             for _ in last_doc..doc {
                 index_writer.add_document(doc!(int_field=>1u64));
             }
@@ -733,7 +774,7 @@ mod tests {
     #[test]
     fn test_block_segment_postings_skip() {
         for i in 0..4 {
-            let mut block_postings = build_block_postings(vec![3]);
+            let mut block_postings = build_block_postings(&[3]);
             assert_eq!(
                 block_postings.skip_to(i),
                 BlockSegmentPostingsSkipResult::Success(0u32)
@@ -743,7 +784,7 @@ mod tests {
                 BlockSegmentPostingsSkipResult::Terminated
             );
         }
-        let mut block_postings = build_block_postings(vec![3]);
+        let mut block_postings = build_block_postings(&[3]);
         assert_eq!(
             block_postings.skip_to(4u32),
             BlockSegmentPostingsSkipResult::Terminated
@@ -756,7 +797,7 @@ mod tests {
         for i in 0..1300 {
             docs.push((i * i / 100) + i);
         }
-        let mut block_postings = build_block_postings(docs.clone());
+        let mut block_postings = build_block_postings(&docs[..]);
         for i in vec![0, 424, 10000] {
             assert_eq!(
                 block_postings.skip_to(i),
@@ -778,7 +819,7 @@ mod tests {
 
     #[test]
     fn test_reset_block_segment_postings() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let int_field = schema_builder.add_u64_field("id", INT_INDEXED);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);

src/postings/serializer.rs

@@ -55,7 +55,7 @@ pub struct InvertedIndexSerializer {
 
 impl InvertedIndexSerializer {
     /// Open a new `PostingsSerializer` for the given segment
-    fn new(
+    fn create(
         terms_write: CompositeWrite<WritePtr>,
         postings_write: CompositeWrite<WritePtr>,
         positions_write: CompositeWrite<WritePtr>,
@@ -74,7 +74,7 @@ impl InvertedIndexSerializer {
     /// Open a new `PostingsSerializer` for the given segment
     pub fn open(segment: &mut Segment) -> Result<InvertedIndexSerializer> {
         use SegmentComponent::{POSITIONS, POSITIONSSKIP, POSTINGS, TERMS};
-        InvertedIndexSerializer::new(
+        InvertedIndexSerializer::create(
             CompositeWrite::wrap(segment.open_write(TERMS)?),
             CompositeWrite::wrap(segment.open_write(POSTINGS)?),
             CompositeWrite::wrap(segment.open_write(POSITIONS)?),
@@ -99,7 +99,7 @@ impl InvertedIndexSerializer {
         let positions_write = self.positions_write.for_field(field);
         let positionsidx_write = self.positionsidx_write.for_field(field);
         let field_type: FieldType = (*field_entry.field_type()).clone();
-        FieldSerializer::new(
+        FieldSerializer::create(
             &field_type,
             term_dictionary_write,
             postings_write,
@@ -130,7 +130,7 @@ pub struct FieldSerializer<'a> {
 }
 
 impl<'a> FieldSerializer<'a> {
-    fn new(
+    fn create(
         field_type: &FieldType,
         term_dictionary_write: &'a mut CountingWriter<WritePtr>,
         postings_write: &'a mut CountingWriter<WritePtr>,
@@ -152,7 +152,7 @@ impl<'a> FieldSerializer<'a> {
             _ => (false, false),
         };
         let term_dictionary_builder =
-            TermDictionaryBuilder::new(term_dictionary_write, &field_type)?;
+            TermDictionaryBuilder::create(term_dictionary_write, &field_type)?;
         let postings_serializer =
             PostingsSerializer::new(postings_write, term_freq_enabled, position_enabled);
         let positions_serializer_opt = if position_enabled {

src/postings/stacker/expull.rs

@@ -174,8 +174,8 @@ mod tests {
 
 #[cfg(all(test, feature = "unstable"))]
 mod bench {
+    use super::super::MemoryArena;
     use super::ExpUnrolledLinkedList;
-    use tantivy_memory_arena::MemoryArena;
     use test::Bencher;
 
     const NUM_STACK: usize = 10_000;
@@ -199,20 +199,19 @@ mod bench {
 
     #[bench]
     fn bench_push_stack(bench: &mut Bencher) {
-        let heap = MemoryArena::new();
         bench.iter(|| {
+            let mut heap = MemoryArena::new();
             let mut stacks = Vec::with_capacity(100);
             for _ in 0..NUM_STACK {
-                let (_, stack) = heap.allocate_object::<ExpUnrolledLinkedList>();
+                let mut stack = ExpUnrolledLinkedList::new(&mut heap);
                 stacks.push(stack);
             }
             for s in 0..NUM_STACK {
                 for i in 0u32..STACK_SIZE {
                     let t = s * 392017 % NUM_STACK;
-                    stacks[t].push(i, &heap);
+                    stacks[t].push(i, &mut heap);
                 }
             }
-            heap.clear();
         });
     }
 }

src/postings/stacker/mod.rs

@@ -1,9 +1,7 @@
 mod expull;
 mod memory_arena;
-mod murmurhash2;
 mod term_hashmap;
 
 pub use self::expull::ExpUnrolledLinkedList;
 pub use self::memory_arena::{Addr, ArenaStorable, MemoryArena};
-use self::murmurhash2::murmurhash2;
 pub use self::term_hashmap::{compute_table_size, TermHashMap};

src/postings/stacker/murmurhash2.rs

@@ -1,87 +0,0 @@
-use std::ptr;
-const SEED: u32 = 3_242_157_231u32;
-const M: u32 = 0x5bd1_e995;
-
-#[inline(always)]
-pub fn murmurhash2(key: &[u8]) -> u32 {
-    #[cfg_attr(feature = "cargo-clippy", allow(clippy::cast_ptr_alignment))]
-    let mut key_ptr: *const u32 = key.as_ptr() as *const u32;
-    let len = key.len() as u32;
-    let mut h: u32 = SEED ^ len;
-
-    let num_blocks = len >> 2;
-    for _ in 0..num_blocks {
-        let mut k: u32 = unsafe { ptr::read_unaligned(key_ptr) }; // ok because of num_blocks definition
-        k = k.wrapping_mul(M);
-        k ^= k >> 24;
-        k = k.wrapping_mul(M);
-        h = h.wrapping_mul(M);
-        h ^= k;
-        key_ptr = key_ptr.wrapping_offset(1);
-    }
-
-    // Handle the last few bytes of the input array
-    let remaining: &[u8] = &key[key.len() & !3..];
-    match remaining.len() {
-        3 => {
-            h ^= u32::from(remaining[2]) << 16;
-            h ^= u32::from(remaining[1]) << 8;
-            h ^= u32::from(remaining[0]);
-            h = h.wrapping_mul(M);
-        }
-        2 => {
-            h ^= u32::from(remaining[1]) << 8;
-            h ^= u32::from(remaining[0]);
-            h = h.wrapping_mul(M);
-        }
-        1 => {
-            h ^= u32::from(remaining[0]);
-            h = h.wrapping_mul(M);
-        }
-        _ => {}
-    }
-    h ^= h >> 13;
-    h = h.wrapping_mul(M);
-    h ^ (h >> 15)
-}
-
-#[cfg(test)]
-mod test {
-
-    use super::murmurhash2;
-    use std::collections::HashSet;
-
-    #[test]
-    fn test_murmur() {
-        let s1 = "abcdef";
-        let s2 = "abcdeg";
-        for i in 0..5 {
-            assert_eq!(
-                murmurhash2(&s1[i..5].as_bytes()),
-                murmurhash2(&s2[i..5].as_bytes())
-            );
-        }
-    }
-
-    #[test]
-    fn test_murmur_against_reference_impl() {
-        assert_eq!(murmurhash2("".as_bytes()), 3632506080);
-        assert_eq!(murmurhash2("a".as_bytes()), 455683869);
-        assert_eq!(murmurhash2("ab".as_bytes()), 2448092234);
-        assert_eq!(murmurhash2("abc".as_bytes()), 2066295634);
-        assert_eq!(murmurhash2("abcd".as_bytes()), 2588571162);
-        assert_eq!(murmurhash2("abcde".as_bytes()), 2988696942);
-        assert_eq!(murmurhash2("abcdefghijklmnop".as_bytes()), 2350868870);
-    }
-
-    #[test]
-    fn test_murmur_collisions() {
-        let mut set: HashSet<u32> = HashSet::default();
-        for i in 0..10_000 {
-            let s = format!("hash{}", i);
-            let hash = murmurhash2(s.as_bytes());
-            set.insert(hash);
-        }
-        assert_eq!(set.len(), 10_000);
-    }
-}

src/postings/stacker/term_hashmap.rs

@@ -1,4 +1,7 @@
-use super::murmurhash2;
+extern crate murmurhash32;
+
+use self::murmurhash32::murmurhash2;
+
 use super::{Addr, ArenaStorable, MemoryArena};
 use std::iter;
 use std::mem;
@@ -206,7 +209,7 @@ impl TermHashMap {
             self.resize();
         }
         let key_bytes: &[u8] = key.as_ref();
-        let hash = murmurhash2::murmurhash2(key.as_ref());
+        let hash = murmurhash2(key.as_ref());
         let mut probe = self.probe(hash);
         loop {
             let bucket = probe.next_probe();

src/query/all_query.rs

@@ -86,12 +86,12 @@ mod tests {
 
     use super::AllQuery;
     use query::Query;
-    use schema::{SchemaBuilder, TEXT};
+    use schema::{Schema, TEXT};
     use Index;
 
     #[test]
     fn test_all_query() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);

src/query/automaton_weight.rs

@@ -11,7 +11,7 @@ use Result;
 /// A weight struct for Fuzzy Term and Regex Queries
 pub struct AutomatonWeight<A>
 where
-    A: Automaton,
+    A: Automaton + Send + Sync + 'static,
 {
     field: Field,
     automaton: A,
@@ -19,7 +19,7 @@ where
 
 impl<A> AutomatonWeight<A>
 where
-    A: Automaton,
+    A: Automaton + Send + Sync + 'static,
 {
     /// Create a new AutomationWeight
     pub fn new(field: Field, automaton: A) -> AutomatonWeight<A> {
@@ -34,7 +34,7 @@ where
 
 impl<A> Weight for AutomatonWeight<A>
 where
-    A: Automaton,
+    A: Automaton + Send + Sync + 'static,
 {
     fn scorer(&self, reader: &SegmentReader) -> Result<Box<Scorer>> {
         let max_doc = reader.max_doc();

src/query/bm25.rs

@@ -63,7 +63,8 @@ impl BM25Weight {
             .map(|term| {
                 let term_doc_freq = searcher.doc_freq(term);
                 idf(term_doc_freq, total_num_docs)
-            }).sum::<f32>();
+            })
+            .sum::<f32>();
         BM25Weight::new(idf, average_fieldnorm)
     }
 

src/query/boolean_query/boolean_query.rs

@@ -47,7 +47,8 @@ impl Query for BooleanQuery {
             .iter()
             .map(|&(ref occur, ref subquery)| {
                 Ok((*occur, subquery.weight(searcher, scoring_enabled)?))
-            }).collect::<Result<_>>()?;
+            })
+            .collect::<Result<_>>()?;
         Ok(Box::new(BooleanWeight::new(sub_weights, scoring_enabled)))
     }
 
@@ -68,7 +69,8 @@ impl BooleanQuery {
                 let term_query: Box<Query> =
                     Box::new(TermQuery::new(term, IndexRecordOption::WithFreqs));
                 (Occur::Should, term_query)
-            }).collect();
+            })
+            .collect();
         BooleanQuery::from(occur_term_queries)
     }
 

src/query/boolean_query/mod.rs

@@ -19,10 +19,11 @@ mod tests {
     use query::Scorer;
     use query::TermQuery;
     use schema::*;
+    use DocId;
     use Index;
 
     fn aux_test_helper() -> (Index, Field) {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -130,9 +131,13 @@ mod tests {
 
         let matching_docs = |boolean_query: &Query| {
             let searcher = index.searcher();
-            let mut test_collector = TestCollector::default();
-            searcher.search(boolean_query, &mut test_collector).unwrap();
-            test_collector.docs()
+            let test_docs = searcher.search(boolean_query, &TestCollector).unwrap();
+            test_docs
+                .docs()
+                .iter()
+                .cloned()
+                .map(|doc| doc.1)
+                .collect::<Vec<DocId>>()
         };
 
         {
@@ -186,9 +191,8 @@ mod tests {
 
         let score_docs = |boolean_query: &Query| {
             let searcher = index.searcher();
-            let mut test_collector = TestCollector::default();
-            searcher.search(boolean_query, &mut test_collector).unwrap();
-            test_collector.scores()
+            let fruit = searcher.search(boolean_query, &TestCollector).unwrap();
+            fruit.scores().to_vec()
         };
 
         {

src/query/fuzzy_query.rs

@@ -25,14 +25,14 @@ lazy_static! {
 /// ```rust
 /// #[macro_use]
 /// extern crate tantivy;
-/// use tantivy::schema::{SchemaBuilder, TEXT};
+/// use tantivy::schema::{Schema, TEXT};
 /// use tantivy::{Index, Result, Term};
-/// use tantivy::collector::{CountCollector, TopCollector, chain};
+/// use tantivy::collector::{Count, TopDocs};
 /// use tantivy::query::FuzzyTermQuery;
 ///
 /// # fn main() { example().unwrap(); }
 /// fn example() -> Result<()> {
-///     let mut schema_builder = SchemaBuilder::new();
+///     let mut schema_builder = Schema::builder();
 ///     let title = schema_builder.add_text_field("title", TEXT);
 ///     let schema = schema_builder.build();
 ///     let index = Index::create_in_ram(schema);
@@ -57,16 +57,12 @@ lazy_static! {
 ///     let searcher = index.searcher();
 ///
 ///     {
-///         let mut top_collector = TopCollector::with_limit(2);
-///         let mut count_collector = CountCollector::default();
-///         {
-///             let mut collectors = chain().push(&mut top_collector).push(&mut count_collector);
-///             let term = Term::from_field_text(title, "Diary");
-///             let query = FuzzyTermQuery::new(term, 1, true);
-///             searcher.search(&query, &mut collectors).unwrap();
-///         }
-///         assert_eq!(count_collector.count(), 2);
-///         assert!(top_collector.at_capacity());
+///
+///         let term = Term::from_field_text(title, "Diary");
+///         let query = FuzzyTermQuery::new(term, 1, true);
+///         let (top_docs, count) = searcher.search(&query, &(TopDocs::with_limit(2), Count)).unwrap();
+///         assert_eq!(count, 2);
+///         assert_eq!(top_docs.len(), 2);
 ///     }
 ///
 ///     Ok(())
@@ -122,8 +118,8 @@ impl Query for FuzzyTermQuery {
 #[cfg(test)]
 mod test {
     use super::FuzzyTermQuery;
-    use collector::TopCollector;
-    use schema::SchemaBuilder;
+    use collector::TopDocs;
+    use schema::Schema;
     use schema::TEXT;
     use tests::assert_nearly_equals;
     use Index;
@@ -131,7 +127,7 @@ mod test {
 
     #[test]
     pub fn test_fuzzy_term() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let country_field = schema_builder.add_text_field("country", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -148,14 +144,14 @@ mod test {
         index.load_searchers().unwrap();
         let searcher = index.searcher();
         {
-            let mut collector = TopCollector::with_limit(2);
             let term = Term::from_field_text(country_field, "japon");
 
             let fuzzy_query = FuzzyTermQuery::new(term, 1, true);
-            searcher.search(&fuzzy_query, &mut collector).unwrap();
-            let scored_docs = collector.top_docs();
-            assert_eq!(scored_docs.len(), 1, "Expected only 1 document");
-            let (score, _) = scored_docs[0];
+            let top_docs = searcher
+                .search(&fuzzy_query, &TopDocs::with_limit(2))
+                .unwrap();
+            assert_eq!(top_docs.len(), 1, "Expected only 1 document");
+            let (score, _) = top_docs[0];
             assert_nearly_equals(1f32, score);
         }
     }

src/query/mod.rs

@@ -16,10 +16,7 @@ mod phrase_query;
 mod query;
 mod query_parser;
 mod range_query;
-
-#[cfg(feature="regex_query")]
 mod regex_query;
-
 mod reqopt_scorer;
 mod scorer;
 mod term_query;
@@ -50,10 +47,7 @@ pub use self::query::Query;
 pub use self::query_parser::QueryParser;
 pub use self::query_parser::QueryParserError;
 pub use self::range_query::RangeQuery;
-
-#[cfg(feature="regex_query")]
 pub use self::regex_query::RegexQuery;
-
 pub use self::reqopt_scorer::RequiredOptionalScorer;
 pub use self::scorer::ConstScorer;
 pub use self::scorer::Scorer;
@@ -62,15 +56,15 @@ pub use self::weight::Weight;
 
 #[cfg(test)]
 mod tests {
-    use Index;
-    use schema::{SchemaBuilder, TEXT};
     use query::QueryParser;
-    use Term;
+    use schema::{Schema, TEXT};
     use std::collections::BTreeSet;
+    use Index;
+    use Term;
 
     #[test]
     fn test_query_terms() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -79,33 +73,48 @@ mod tests {
         let term_b = Term::from_field_text(text_field, "b");
         {
             let mut terms_set: BTreeSet<Term> = BTreeSet::new();
-            query_parser.parse_query("a").unwrap().query_terms(&mut terms_set);
+            query_parser
+                .parse_query("a")
+                .unwrap()
+                .query_terms(&mut terms_set);
             let terms: Vec<&Term> = terms_set.iter().collect();
             assert_eq!(vec![&term_a], terms);
         }
         {
             let mut terms_set: BTreeSet<Term> = BTreeSet::new();
-            query_parser.parse_query("a b").unwrap().query_terms(&mut terms_set);
+            query_parser
+                .parse_query("a b")
+                .unwrap()
+                .query_terms(&mut terms_set);
             let terms: Vec<&Term> = terms_set.iter().collect();
             assert_eq!(vec![&term_a, &term_b], terms);
         }
         {
             let mut terms_set: BTreeSet<Term> = BTreeSet::new();
-            query_parser.parse_query("\"a b\"").unwrap().query_terms(&mut terms_set);
+            query_parser
+                .parse_query("\"a b\"")
+                .unwrap()
+                .query_terms(&mut terms_set);
             let terms: Vec<&Term> = terms_set.iter().collect();
             assert_eq!(vec![&term_a, &term_b], terms);
         }
         {
             let mut terms_set: BTreeSet<Term> = BTreeSet::new();
-            query_parser.parse_query("a a a a a").unwrap().query_terms(&mut terms_set);
+            query_parser
+                .parse_query("a a a a a")
+                .unwrap()
+                .query_terms(&mut terms_set);
             let terms: Vec<&Term> = terms_set.iter().collect();
             assert_eq!(vec![&term_a], terms);
         }
         {
             let mut terms_set: BTreeSet<Term> = BTreeSet::new();
-            query_parser.parse_query("a -b").unwrap().query_terms(&mut terms_set);
+            query_parser
+                .parse_query("a -b")
+                .unwrap()
+                .query_terms(&mut terms_set);
             let terms: Vec<&Term> = terms_set.iter().collect();
             assert_eq!(vec![&term_a, &term_b], terms);
         }
     }
-}
+}

src/query/phrase_query/mod.rs

@@ -13,11 +13,13 @@ mod tests {
     use collector::tests::TestCollector;
     use core::Index;
     use error::TantivyError;
-    use schema::{SchemaBuilder, Term, TEXT};
+    use schema::{Schema, Term, TEXT};
     use tests::assert_nearly_equals;
+    use DocAddress;
+    use DocId;
 
     fn create_index(texts: &[&'static str]) -> Index {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -47,16 +49,19 @@ mod tests {
         index.load_searchers().unwrap();
         let searcher = index.searcher();
         let test_query = |texts: Vec<&str>| {
-            let mut test_collector = TestCollector::default();
             let terms: Vec<Term> = texts
                 .iter()
                 .map(|text| Term::from_field_text(text_field, text))
                 .collect();
             let phrase_query = PhraseQuery::new(terms);
-            searcher
-                .search(&phrase_query, &mut test_collector)
+            let test_fruits = searcher
+                .search(&phrase_query, &TestCollector)
                 .expect("search should succeed");
-            test_collector.docs()
+            test_fruits
+                .docs()
+                .iter()
+                .map(|docaddr| docaddr.1)
+                .collect::<Vec<_>>()
         };
         assert_eq!(test_query(vec!["a", "b", "c"]), vec![2, 4]);
         assert_eq!(test_query(vec!["a", "b"]), vec![1, 2, 3, 4]);
@@ -67,7 +72,7 @@ mod tests {
 
     #[test]
     pub fn test_phrase_query_no_positions() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         use schema::IndexRecordOption;
         use schema::TextFieldIndexing;
         use schema::TextOptions;
@@ -91,9 +96,9 @@ mod tests {
             Term::from_field_text(text_field, "a"),
             Term::from_field_text(text_field, "b"),
         ]);
-        let mut test_collector = TestCollector::default();
         if let TantivyError::SchemaError(ref msg) = searcher
-            .search(&phrase_query, &mut test_collector)
+            .search(&phrase_query, &TestCollector)
+            .map(|_| ())
             .unwrap_err()
         {
             assert_eq!(
@@ -113,16 +118,16 @@ mod tests {
         index.load_searchers().unwrap();
         let searcher = index.searcher();
         let test_query = |texts: Vec<&str>| {
-            let mut test_collector = TestCollector::default();
             let terms: Vec<Term> = texts
                 .iter()
                 .map(|text| Term::from_field_text(text_field, text))
                 .collect();
             let phrase_query = PhraseQuery::new(terms);
             searcher
-                .search(&phrase_query, &mut test_collector)
-                .expect("search should succeed");
-            test_collector.scores()
+                .search(&phrase_query, &TestCollector)
+                .expect("search should succeed")
+                .scores()
+                .to_vec()
         };
         let scores = test_query(vec!["a", "b"]);
         assert_nearly_equals(scores[0], 0.40618482);
@@ -131,51 +136,39 @@ mod tests {
 
     #[test] // motivated by #234
     pub fn test_phrase_query_docfreq_order() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
         {
             let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
-            {
-                // 0
-                let doc = doc!(text_field=>"b");
-                index_writer.add_document(doc);
-            }
-            {
-                // 1
-                let doc = doc!(text_field=>"a b");
-                index_writer.add_document(doc);
-            }
-            {
-                // 2
-                let doc = doc!(text_field=>"b a");
-                index_writer.add_document(doc);
-            }
+            index_writer.add_document(doc!(text_field=>"b"));
+            index_writer.add_document(doc!(text_field=>"a b"));
+            index_writer.add_document(doc!(text_field=>"b a"));
             assert!(index_writer.commit().is_ok());
         }
 
         index.load_searchers().unwrap();
         let searcher = index.searcher();
         let test_query = |texts: Vec<&str>| {
-            let mut test_collector = TestCollector::default();
             let terms: Vec<Term> = texts
                 .iter()
                 .map(|text| Term::from_field_text(text_field, text))
                 .collect();
             let phrase_query = PhraseQuery::new(terms);
             searcher
-                .search(&phrase_query, &mut test_collector)
-                .expect("search should succeed");
-            test_collector.docs()
+                .search(&phrase_query, &TestCollector)
+                .expect("search should succeed")
+                .docs()
+                .to_vec()
         };
-        assert_eq!(test_query(vec!["a", "b"]), vec![1]);
-        assert_eq!(test_query(vec!["b", "a"]), vec![2]);
+        assert_eq!(test_query(vec!["a", "b"]), vec![DocAddress(0, 1)]);
+        assert_eq!(test_query(vec!["b", "a"]), vec![DocAddress(0, 2)]);
     }
 
     #[test] // motivated by #234
     pub fn test_phrase_query_non_trivial_offsets() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -187,16 +180,18 @@ mod tests {
         index.load_searchers().unwrap();
         let searcher = index.searcher();
         let test_query = |texts: Vec<(usize, &str)>| {
-            let mut test_collector = TestCollector::default();
             let terms: Vec<(usize, Term)> = texts
                 .iter()
                 .map(|(offset, text)| (*offset, Term::from_field_text(text_field, text)))
                 .collect();
             let phrase_query = PhraseQuery::new_with_offset(terms);
             searcher
-                .search(&phrase_query, &mut test_collector)
-                .expect("search should succeed");
-            test_collector.docs()
+                .search(&phrase_query, &TestCollector)
+                .expect("search should succeed")
+                .docs()
+                .iter()
+                .map(|doc_address| doc_address.1)
+                .collect::<Vec<DocId>>()
         };
         assert_eq!(test_query(vec![(0, "a"), (1, "b")]), vec![0]);
         assert_eq!(test_query(vec![(1, "b"), (0, "a")]), vec![0]);

src/query/phrase_query/phrase_scorer.rs

@@ -134,7 +134,8 @@ impl<TPostings: Postings> PhraseScorer<TPostings> {
             .into_iter()
             .map(|(offset, postings)| {
                 PostingsWithOffset::new(postings, (max_offset - offset) as u32)
-            }).collect::<Vec<_>>();
+            })
+            .collect::<Vec<_>>();
         PhraseScorer {
             intersection_docset: Intersection::new(postings_with_offsets),
             num_docsets,

src/query/query.rs

@@ -1,11 +1,9 @@
 use super::Weight;
-use collector::Collector;
 use core::searcher::Searcher;
 use downcast;
 use std::collections::BTreeSet;
 use std::fmt;
 use Result;
-use SegmentLocalId;
 use Term;
 
 /// The `Query` trait defines a set of documents and a scoring method
@@ -63,26 +61,6 @@ pub trait Query: QueryClone + downcast::Any + fmt::Debug {
     /// Extract all of the terms associated to the query and insert them in the
     /// term set given in arguments.
     fn query_terms(&self, _term_set: &mut BTreeSet<Term>) {}
-
-    /// Search works as follows :
-    ///
-    /// First the weight object associated to the query is created.
-    ///
-    /// Then, the query loops over the segments and for each segment :
-    /// - setup the collector and informs it that the segment being processed has changed.
-    /// - creates a `Scorer` object associated for this segment
-    /// - iterate throw the matched documents and push them to the collector.
-    ///
-    fn search(&self, searcher: &Searcher, collector: &mut Collector) -> Result<()> {
-        let scoring_enabled = collector.requires_scoring();
-        let weight = self.weight(searcher, scoring_enabled)?;
-        for (segment_ord, segment_reader) in searcher.segment_readers().iter().enumerate() {
-            collector.set_segment(segment_ord as SegmentLocalId, segment_reader)?;
-            let mut scorer = weight.scorer(segment_reader)?;
-            scorer.collect(collector, segment_reader.delete_bitset());
-        }
-        Ok(())
-    }
 }
 
 pub trait QueryClone {
@@ -98,6 +76,26 @@ where
     }
 }
 
+impl Query for Box<Query> {
+    fn weight(&self, searcher: &Searcher, scoring_enabled: bool) -> Result<Box<Weight>> {
+        self.as_ref().weight(searcher, scoring_enabled)
+    }
+
+    fn count(&self, searcher: &Searcher) -> Result<usize> {
+        self.as_ref().count(searcher)
+    }
+
+    fn query_terms(&self, term_set: &mut BTreeSet<Term<Vec<u8>>>) {
+        self.as_ref().query_terms(term_set);
+    }
+}
+
+impl QueryClone for Box<Query> {
+    fn box_clone(&self) -> Box<Query> {
+        self.as_ref().box_clone()
+    }
+}
+
 #[allow(missing_docs)]
 mod downcast_impl {
     downcast!(super::Query);

src/query/query_parser/query_parser.rs

@@ -68,7 +68,8 @@ fn trim_ast(logical_ast: LogicalAST) -> Option<LogicalAST> {
                 .into_iter()
                 .flat_map(|(occur, child)| {
                     trim_ast(child).map(|trimmed_child| (occur, trimmed_child))
-                }).collect::<Vec<_>>();
+                })
+                .collect::<Vec<_>>();
             if trimmed_children.is_empty() {
                 None
             } else {
@@ -128,6 +129,7 @@ fn trim_ast(logical_ast: LogicalAST) -> Option<LogicalAST> {
 ///
 /// *  all docs query: A plain `*` will match all documents in the index.
 ///
+#[derive(Clone)]
 pub struct QueryParser {
     schema: Schema,
     default_fields: Vec<Field>,
@@ -421,7 +423,8 @@ impl QueryParser {
                             lower: self.resolve_bound(field, &lower)?,
                             upper: self.resolve_bound(field, &upper)?,
                         })))
-                    }).collect::<Result<Vec<_>, QueryParserError>>()?;
+                    })
+                    .collect::<Result<Vec<_>, QueryParserError>>()?;
                 let result_ast = if clauses.len() == 1 {
                     clauses.pop().unwrap()
                 } else {
@@ -484,12 +487,12 @@ mod test {
     use query::Query;
     use schema::Field;
     use schema::{IndexRecordOption, TextFieldIndexing, TextOptions};
-    use schema::{SchemaBuilder, Term, INT_INDEXED, STORED, STRING, TEXT};
+    use schema::{Schema, Term, INT_INDEXED, STORED, STRING, TEXT};
     use tokenizer::{LowerCaser, SimpleTokenizer, StopWordFilter, Tokenizer, TokenizerManager};
     use Index;
 
     fn make_query_parser() -> QueryParser {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field_indexing = TextFieldIndexing::default()
             .set_tokenizer("en_with_stop_words")
             .set_index_option(IndexRecordOption::WithFreqsAndPositions);
@@ -597,25 +600,19 @@ mod test {
         assert!(query_parser.parse_query("signed:2324").is_ok());
         assert!(query_parser.parse_query("signed:\"22\"").is_ok());
         assert!(query_parser.parse_query("signed:\"-2234\"").is_ok());
-        assert!(
-            query_parser
-                .parse_query("signed:\"-9999999999999\"")
-                .is_ok()
-        );
+        assert!(query_parser
+            .parse_query("signed:\"-9999999999999\"")
+            .is_ok());
         assert!(query_parser.parse_query("signed:\"a\"").is_err());
         assert!(query_parser.parse_query("signed:\"2a\"").is_err());
-        assert!(
-            query_parser
-                .parse_query("signed:\"18446744073709551615\"")
-                .is_err()
-        );
+        assert!(query_parser
+            .parse_query("signed:\"18446744073709551615\"")
+            .is_err());
         assert!(query_parser.parse_query("unsigned:\"2\"").is_ok());
         assert!(query_parser.parse_query("unsigned:\"-2\"").is_err());
-        assert!(
-            query_parser
-                .parse_query("unsigned:\"18446744073709551615\"")
-                .is_ok()
-        );
+        assert!(query_parser
+            .parse_query("unsigned:\"18446744073709551615\"")
+            .is_ok());
         test_parse_query_to_logical_ast_helper(
             "unsigned:2324",
             "Term([0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 9, 20])",
@@ -720,7 +717,7 @@ mod test {
 
     #[test]
     pub fn test_unknown_tokenizer() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field_indexing = TextFieldIndexing::default()
             .set_tokenizer("nonexistingtokenizer")
             .set_index_option(IndexRecordOption::Basic);
@@ -738,7 +735,7 @@ mod test {
 
     #[test]
     pub fn test_query_parser_no_positions() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field_indexing = TextFieldIndexing::default()
             .set_tokenizer("customtokenizer")
             .set_index_option(IndexRecordOption::Basic);

src/query/range_query.rs

@@ -40,14 +40,13 @@ fn map_bound<TFrom, TTo, Transform: Fn(&TFrom) -> TTo>(
 /// # #[macro_use]
 /// # extern crate tantivy;
 /// # use tantivy::Index;
-/// # use tantivy::schema::{SchemaBuilder, INT_INDEXED};
-/// # use tantivy::collector::CountCollector;
-/// # use tantivy::query::Query;
+/// # use tantivy::schema::{Schema, INT_INDEXED};
+/// # use tantivy::collector::Count;
 /// # use tantivy::Result;
 /// # use tantivy::query::RangeQuery;
 /// #
 /// # fn run() -> Result<()> {
-/// #     let mut schema_builder = SchemaBuilder::new();
+/// #     let mut schema_builder = Schema::builder();
 /// #     let year_field = schema_builder.add_u64_field("year", INT_INDEXED);
 /// #     let schema = schema_builder.build();
 /// #
@@ -67,10 +66,7 @@ fn map_bound<TFrom, TTo, Transform: Fn(&TFrom) -> TTo>(
 ///
 /// let docs_in_the_sixties = RangeQuery::new_u64(year_field, 1960..1970);
 ///
-/// let mut count_collector = CountCollector::default();
-/// docs_in_the_sixties.search(&searcher, &mut count_collector)?;
-///
-/// let num_60s_books = count_collector.count();
+/// let num_60s_books = searcher.search(&docs_in_the_sixties, &Count)?;
 ///
 /// #     assert_eq!(num_60s_books, 2285);
 /// #     Ok(())
@@ -296,9 +292,8 @@ impl Weight for RangeWeight {
 mod tests {
 
     use super::RangeQuery;
-    use collector::CountCollector;
-    use query::Query;
-    use schema::{Document, Field, SchemaBuilder, INT_INDEXED};
+    use collector::Count;
+    use schema::{Document, Field, Schema, INT_INDEXED};
     use std::collections::Bound;
     use Index;
     use Result;
@@ -306,7 +301,7 @@ mod tests {
     #[test]
     fn test_range_query_simple() {
         fn run() -> Result<()> {
-            let mut schema_builder = SchemaBuilder::new();
+            let mut schema_builder = Schema::builder();
             let year_field = schema_builder.add_u64_field("year", INT_INDEXED);
             let schema = schema_builder.build();
 
@@ -327,9 +322,8 @@ mod tests {
             let docs_in_the_sixties = RangeQuery::new_u64(year_field, 1960u64..1970u64);
 
             // ... or `1960..=1969` if inclusive range is enabled.
-            let mut count_collector = CountCollector::default();
-            docs_in_the_sixties.search(&searcher, &mut count_collector)?;
-            assert_eq!(count_collector.count(), 2285);
+            let count = searcher.search(&docs_in_the_sixties, &Count)?;
+            assert_eq!(count, 2285);
             Ok(())
         }
 
@@ -340,7 +334,7 @@ mod tests {
     fn test_range_query() {
         let int_field: Field;
         let schema = {
-            let mut schema_builder = SchemaBuilder::new();
+            let mut schema_builder = Schema::builder();
             int_field = schema_builder.add_i64_field("intfield", INT_INDEXED);
             schema_builder.build()
         };
@@ -363,11 +357,8 @@ mod tests {
         }
         index.load_searchers().unwrap();
         let searcher = index.searcher();
-        let count_multiples = |range_query: RangeQuery| {
-            let mut count_collector = CountCollector::default();
-            range_query.search(&searcher, &mut count_collector).unwrap();
-            count_collector.count()
-        };
+        let count_multiples =
+            |range_query: RangeQuery| searcher.search(&range_query, &Count).unwrap();
 
         assert_eq!(count_multiples(RangeQuery::new_i64(int_field, 10..11)), 9);
         assert_eq!(

src/query/regex_query.rs

@@ -1,7 +1,5 @@
-extern crate fst_regex;
-
 use error::TantivyError;
-use self::fst_regex::Regex;
+use fst_regex::Regex;
 use query::{AutomatonWeight, Query, Weight};
 use schema::Field;
 use std::clone::Clone;
@@ -18,14 +16,14 @@ use Searcher;
 /// ```rust
 /// #[macro_use]
 /// extern crate tantivy;
-/// use tantivy::schema::{SchemaBuilder, TEXT};
+/// use tantivy::schema::{Schema, TEXT};
 /// use tantivy::{Index, Result, Term};
-/// use tantivy::collector::{CountCollector, TopCollector, chain};
+/// use tantivy::collector::Count;
 /// use tantivy::query::RegexQuery;
 ///
 /// # fn main() { example().unwrap(); }
 /// fn example() -> Result<()> {
-///     let mut schema_builder = SchemaBuilder::new();
+///     let mut schema_builder = Schema::builder();
 ///     let title = schema_builder.add_text_field("title", TEXT);
 ///     let schema = schema_builder.build();
 ///     let index = Index::create_in_ram(schema);
@@ -49,19 +47,10 @@ use Searcher;
 ///     index.load_searchers()?;
 ///     let searcher = index.searcher();
 ///
-///     {
-///         let mut top_collector = TopCollector::with_limit(2);
-///         let mut count_collector = CountCollector::default();
-///         {
-///             let mut collectors = chain().push(&mut top_collector).push(&mut count_collector);
-///             let term = Term::from_field_text(title, "Diary");
-///             let query = RegexQuery::new("d[ai]{2}ry".to_string(), title);
-///             searcher.search(&query, &mut collectors).unwrap();
-///         }
-///         assert_eq!(count_collector.count(), 3);
-///         assert!(top_collector.at_capacity());
-///     }
-///
+///     let term = Term::from_field_text(title, "Diary");
+///     let query = RegexQuery::new("d[ai]{2}ry".to_string(), title);
+///     let count = searcher.search(&query, &Count)?;
+///     assert_eq!(count, 3);
 ///     Ok(())
 /// }
 /// ```
@@ -97,15 +86,15 @@ impl Query for RegexQuery {
 #[cfg(test)]
 mod test {
     use super::RegexQuery;
-    use collector::TopCollector;
-    use schema::SchemaBuilder;
+    use collector::TopDocs;
+    use schema::Schema;
     use schema::TEXT;
     use tests::assert_nearly_equals;
     use Index;
 
     #[test]
     pub fn test_regex_query() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let country_field = schema_builder.add_text_field("country", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -122,20 +111,18 @@ mod test {
         index.load_searchers().unwrap();
         let searcher = index.searcher();
         {
-            let mut collector = TopCollector::with_limit(2);
             let regex_query = RegexQuery::new("jap[ao]n".to_string(), country_field);
-            searcher.search(&regex_query, &mut collector).unwrap();
-            let scored_docs = collector.top_docs();
+            let scored_docs = searcher
+                .search(&regex_query, &TopDocs::with_limit(2))
+                .unwrap();
             assert_eq!(scored_docs.len(), 1, "Expected only 1 document");
             let (score, _) = scored_docs[0];
             assert_nearly_equals(1f32, score);
         }
-        {
-            let mut collector = TopCollector::with_limit(2);
-            let regex_query = RegexQuery::new("jap[A-Z]n".to_string(), country_field);
-            searcher.search(&regex_query, &mut collector).unwrap();
-            let scored_docs = collector.top_docs();
-            assert_eq!(scored_docs.len(), 0, "Expected ZERO document");
-        }
+        let regex_query = RegexQuery::new("jap[A-Z]n".to_string(), country_field);
+        let top_docs = searcher
+            .search(&regex_query, &TopDocs::with_limit(2))
+            .unwrap();
+        assert!(top_docs.is_empty(), "Expected ZERO document");
     }
 }

src/query/scorer.rs

@@ -1,8 +1,6 @@
-use collector::Collector;
 use common::BitSet;
 use docset::{DocSet, SkipResult};
 use downcast;
-use fastfield::DeleteBitSet;
 use std::ops::DerefMut;
 use DocId;
 use Score;
@@ -16,20 +14,11 @@ pub trait Scorer: downcast::Any + DocSet + 'static {
     /// This method will perform a bit of computation and is not cached.
     fn score(&mut self) -> Score;
 
-    /// Consumes the complete `DocSet` and
-    /// push the scored documents to the collector.
-    fn collect(&mut self, collector: &mut Collector, delete_bitset_opt: Option<&DeleteBitSet>) {
-        if let Some(delete_bitset) = delete_bitset_opt {
-            while self.advance() {
-                let doc = self.doc();
-                if !delete_bitset.is_deleted(doc) {
-                    collector.collect(doc, self.score());
-                }
-            }
-        } else {
-            while self.advance() {
-                collector.collect(self.doc(), self.score());
-            }
+    /// Iterates through all of the document matched by the DocSet
+    /// `DocSet` and push the scored documents to the collector.
+    fn for_each(&mut self, callback: &mut FnMut(DocId, Score)) {
+        while self.advance() {
+            callback(self.doc(), self.score());
         }
     }
 }
@@ -44,9 +33,9 @@ impl Scorer for Box<Scorer> {
         self.deref_mut().score()
     }
 
-    fn collect(&mut self, collector: &mut Collector, delete_bitset: Option<&DeleteBitSet>) {
+    fn for_each(&mut self, callback: &mut FnMut(DocId, Score)) {
         let scorer = self.deref_mut();
-        scorer.collect(collector, delete_bitset);
+        scorer.for_each(callback);
     }
 }
 

src/query/term_query/mod.rs

@@ -9,17 +9,17 @@ pub use self::term_weight::TermWeight;
 #[cfg(test)]
 mod tests {
 
-    use collector::TopCollector;
+    use collector::TopDocs;
     use docset::DocSet;
     use query::{Query, QueryParser, Scorer, TermQuery};
-    use schema::{IndexRecordOption, SchemaBuilder, STRING, TEXT};
+    use schema::{IndexRecordOption, Schema, STRING, TEXT};
     use tests::assert_nearly_equals;
     use Index;
     use Term;
 
     #[test]
     pub fn test_term_query_no_freq() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", STRING);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -49,7 +49,7 @@ mod tests {
 
     #[test]
     pub fn test_term_weight() {
-        let mut schema_builder = SchemaBuilder::new();
+        let mut schema_builder = Schema::builder();
         let left_field = schema_builder.add_text_field("left", TEXT);
         let right_field = schema_builder.add_text_field("right", TEXT);
         let large_field = schema_builder.add_text_field("large", TEXT);
@@ -68,37 +68,35 @@ mod tests {
         index.load_searchers().unwrap();
         let searcher = index.searcher();
         {
-            let mut collector = TopCollector::with_limit(2);
             let term = Term::from_field_text(left_field, "left2");
             let term_query = TermQuery::new(term, IndexRecordOption::WithFreqs);
-            searcher.search(&term_query, &mut collector).unwrap();
-            let scored_docs = collector.top_docs();
-            assert_eq!(scored_docs.len(), 1);
-            let (score, _) = scored_docs[0];
+            let topdocs = searcher
+                .search(&term_query, &TopDocs::with_limit(2))
+                .unwrap();
+            assert_eq!(topdocs.len(), 1);
+            let (score, _) = topdocs[0];
             assert_nearly_equals(0.77802235, score);
         }
         {
-            let mut collector = TopCollector::with_limit(2);
             let term = Term::from_field_text(left_field, "left1");
             let term_query = TermQuery::new(term, IndexRecordOption::WithFreqs);
-            searcher.search(&term_query, &mut collector).unwrap();
-            let scored_docs = collector.top_docs();
-            assert_eq!(scored_docs.len(), 2);
-            let (score1, _) = scored_docs[0];
+            let top_docs = searcher
+                .search(&term_query, &TopDocs::with_limit(2))
+                .unwrap();
+            assert_eq!(top_docs.len(), 2);
+            let (score1, _) = top_docs[0];
             assert_nearly_equals(0.27101856, score1);
-            let (score2, _) = scored_docs[1];
+            let (score2, _) = top_docs[1];
             assert_nearly_equals(0.13736556, score2);
         }
         {
             let query_parser = QueryParser::for_index(&index, vec![]);
             let query = query_parser.parse_query("left:left2 left:left1").unwrap();
-            let mut collector = TopCollector::with_limit(2);
-            searcher.search(&*query, &mut collector).unwrap();
-            let scored_docs = collector.top_docs();
-            assert_eq!(scored_docs.len(), 2);
-            let (score1, _) = scored_docs[0];
+            let top_docs = searcher.search(&query, &TopDocs::with_limit(2)).unwrap();
+            assert_eq!(top_docs.len(), 2);
+            let (score1, _) = top_docs[0];
             assert_nearly_equals(0.9153879, score1);
-            let (score2, _) = scored_docs[1];
+            let (score2, _) = top_docs[1];
             assert_nearly_equals(0.27101856, score2);
         }
     }

src/query/term_query/term_query.rs

@@ -21,14 +21,14 @@ use Term;
 /// ```rust
 /// #[macro_use]
 /// extern crate tantivy;
-/// use tantivy::schema::{SchemaBuilder, TEXT, IndexRecordOption};
+/// use tantivy::schema::{Schema, TEXT, IndexRecordOption};
 /// use tantivy::{Index, Result, Term};
-/// use tantivy::collector::{CountCollector, TopCollector, chain};
+/// use tantivy::collector::{Count, TopDocs};
 /// use tantivy::query::TermQuery;
 ///
 /// # fn main() { example().unwrap(); }
 /// fn example() -> Result<()> {
-///     let mut schema_builder = SchemaBuilder::new();
+///     let mut schema_builder = Schema::builder();
 ///     let title = schema_builder.add_text_field("title", TEXT);
 ///     let schema = schema_builder.build();
 ///     let index = Index::create_in_ram(schema);
@@ -52,20 +52,12 @@ use Term;
 ///     index.load_searchers()?;
 ///     let searcher = index.searcher();
 ///
-///     {
-///         let mut top_collector = TopCollector::with_limit(2);
-///         let mut count_collector = CountCollector::default();
-///         {
-///             let mut collectors = chain().push(&mut top_collector).push(&mut count_collector);
-///             let query = TermQuery::new(
-///                 Term::from_field_text(title, "diary"),
-///                 IndexRecordOption::Basic,
-///             );
-///             searcher.search(&query, &mut collectors).unwrap();
-///         }
-///         assert_eq!(count_collector.count(), 2);
-///         assert!(top_collector.at_capacity());
-///     }
+///     let query = TermQuery::new(
+///         Term::from_field_text(title, "diary"),
+///         IndexRecordOption::Basic,
+///     );
+///     let (top_docs, count) = searcher.search(&query, &(TopDocs::with_limit(2), Count)).unwrap();
+///     assert_eq!(count, 2);
 ///
 ///     Ok(())
 /// }

src/query/union.rs

@@ -55,7 +55,8 @@ where
                         None
                     }
                 },
-            ).collect();
+            )
+            .collect();
         Union {
             docsets: non_empty_docsets,
             bitsets: Box::new([TinySet::empty(); HORIZON_NUM_TINYBITSETS]),
@@ -214,10 +215,7 @@ where
 
             // The target is outside of the buffered horizon.
             // advance all docsets to a doc >= to the target.
-            #[cfg_attr(
-                feature = "cargo-clippy",
-                allow(clippy::clippy::collapsible_if)
-            )]
+            #[cfg_attr(feature = "cargo-clippy", allow(clippy::clippy::collapsible_if))]
             unordered_drain_filter(&mut self.docsets, |docset| {
                 if docset.doc() < target {
                     if docset.skip_next(target) == SkipResult::End {

src/query/weight.rs

@@ -6,7 +6,7 @@ use Result;
 /// for a given set of segments.
 ///
 /// See [`Query`](./trait.Query.html).
-pub trait Weight {
+pub trait Weight: Send + Sync + 'static {
     /// Returns the scorer for the given segment.
     /// See [`Query`](./trait.Query.html).
     fn scorer(&self, reader: &SegmentReader) -> Result<Box<Scorer>>;

src/schema/document.rs

@@ -161,7 +161,7 @@ mod tests {
 
     #[test]
     fn test_doc() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("title", TEXT);
         let mut doc = Document::default();
         doc.add_text(text_field, "My title");

src/schema/facet.rs

@@ -6,6 +6,7 @@ use std::borrow::Cow;
 use std::fmt::{self, Debug, Display, Formatter};
 use std::io::{self, Read, Write};
 use std::str;
+use std::string::FromUtf8Error;
 
 const SLASH_BYTE: u8 = b'/';
 const ESCAPE_BYTE: u8 = b'\\';
@@ -14,6 +15,10 @@ const ESCAPE_BYTE: u8 = b'\\';
 /// representation of facets.
 pub const FACET_SEP_BYTE: u8 = 0u8;
 
+/// `char` used as a level separation in the binary
+/// representation of facets. (It is the null codepoint.)
+pub const FACET_SEP_CHAR: char = '\u{0}';
+
 /// A Facet represent a point in a given hierarchy.
 ///
 /// They are typically represented similarly to a filepath.
@@ -26,18 +31,18 @@ pub const FACET_SEP_BYTE: u8 = 0u8;
 /// its facet. In the example above, `/electronics/tv_and_video/`
 /// and `/electronics`.
 #[derive(Clone, Eq, Hash, PartialEq, Ord, PartialOrd)]
-pub struct Facet(Vec<u8>);
+pub struct Facet(String);
 
 impl Facet {
     /// Returns a new instance of the "root facet"
     /// Equivalent to `/`.
     pub fn root() -> Facet {
-        Facet(vec![])
+        Facet("".to_string())
     }
 
     /// Returns true iff the facet is the root facet `/`.
     pub fn is_root(&self) -> bool {
-        self.encoded_bytes().is_empty()
+        self.encoded_str().is_empty()
     }
 
     /// Returns a binary representation of the facet.
@@ -49,13 +54,19 @@ impl Facet {
     /// This representation has the benefit of making it possible to
     /// express "being a child of a given facet" as a range over
     /// the term ordinals.
-    pub fn encoded_bytes(&self) -> &[u8] {
+    pub fn encoded_str(&self) -> &str {
         &self.0
     }
 
+    pub(crate) fn from_encoded_string(facet_string: String) -> Facet {
+        Facet(facet_string)
+    }
+
     /// Creates a `Facet` from its binary representation.
-    pub(crate) unsafe fn from_encoded(encoded_bytes: Vec<u8>) -> Facet {
-        Facet(encoded_bytes)
+    pub fn from_encoded(encoded_bytes: Vec<u8>) -> Result<Facet, FromUtf8Error> {
+        // facet bytes validation. `0u8` is used a separator but that is still legal utf-8
+        //Ok(Facet(String::from_utf8(encoded_bytes)?))
+        String::from_utf8(encoded_bytes).map(Facet)
     }
 
     /// Parse a text representation of a facet.
@@ -79,36 +90,37 @@ impl Facet {
         Path: IntoIterator,
         Path::Item: ToString,
     {
-        let mut facet_bytes: Vec<u8> = Vec::with_capacity(100);
+        let mut facet_string: String = String::with_capacity(100);
         let mut step_it = path.into_iter();
         if let Some(step) = step_it.next() {
-            facet_bytes.extend_from_slice(step.to_string().as_bytes());
+            facet_string.push_str(&step.to_string());
         }
         for step in step_it {
-            facet_bytes.push(FACET_SEP_BYTE);
-            facet_bytes.extend_from_slice(step.to_string().as_bytes());
+            facet_string.push(FACET_SEP_CHAR);
+            facet_string.push_str(&step.to_string());
         }
-        Facet(facet_bytes)
+        Facet(facet_string)
     }
 
     /// Accessor for the inner buffer of the `Facet`.
-    pub(crate) fn inner_buffer_mut(&mut self) -> &mut Vec<u8> {
-        &mut self.0
+    pub(crate) fn set_facet_str(&mut self, facet_str: &str) {
+        self.0.clear();
+        self.0.push_str(facet_str);
     }
 
     /// Returns `true` iff other is a subfacet of `self`.
     pub fn is_prefix_of(&self, other: &Facet) -> bool {
-        let self_bytes: &[u8] = self.encoded_bytes();
-        let other_bytes: &[u8] = other.encoded_bytes();
-        self_bytes.len() < other_bytes.len()
-            && other_bytes.starts_with(self_bytes)
-            && other_bytes[self_bytes.len()] == 0u8
+        let self_str = self.encoded_str();
+        let other_str = other.encoded_str();
+        self_str.len() < other_str.len()
+            && other_str.starts_with(self_str)
+            && other_str.as_bytes()[self_str.len()] == FACET_SEP_BYTE
     }
 }
 
-impl Borrow<[u8]> for Facet {
-    fn borrow(&self) -> &[u8] {
-        self.encoded_bytes()
+impl Borrow<str> for Facet {
+    fn borrow(&self) -> &str {
+        self.encoded_str()
     }
 }
 
@@ -120,45 +132,51 @@ impl<'a, T: ?Sized + AsRef<str>> From<&'a T> for Facet {
             Idle,
         }
         let path: &str = path_asref.as_ref();
-        let mut facet_encoded = Vec::new();
+        assert!(!path.is_empty());
+        assert!(path.starts_with("/"));
+        let mut facet_encoded = String::new();
         let mut state = State::Idle;
         let path_bytes = path.as_bytes();
-        for &c in &path_bytes[1..] {
+        let mut last_offset = 1;
+        for i in 1..path_bytes.len() {
+            let c = path_bytes[i];
             match (state, c) {
-                (State::Idle, ESCAPE_BYTE) => state = State::Escaped,
+                (State::Idle, ESCAPE_BYTE) => {
+                    facet_encoded.push_str(&path[last_offset..i]);
+                    last_offset = i + 1;
+                    state = State::Escaped
+                }
                 (State::Idle, SLASH_BYTE) => {
-                    facet_encoded.push(FACET_SEP_BYTE);
+                    facet_encoded.push_str(&path[last_offset..i]);
+                    facet_encoded.push(FACET_SEP_CHAR);
+                    last_offset = i + 1;
                 }
-                (State::Escaped, any_char) => {
+                (State::Escaped, _escaped_char) => {
                     state = State::Idle;
-                    facet_encoded.push(any_char);
-                }
-                (State::Idle, other_char) => {
-                    facet_encoded.push(other_char);
                 }
+                (State::Idle, _any_char) => {}
             }
         }
+        facet_encoded.push_str(&path[last_offset..]);
         Facet(facet_encoded)
     }
 }
 
 impl BinarySerializable for Facet {
     fn serialize<W: Write>(&self, writer: &mut W) -> io::Result<()> {
-        <Vec<u8> as BinarySerializable>::serialize(&self.0, writer)
+        <String as BinarySerializable>::serialize(&self.0, writer)
     }
 
     fn deserialize<R: Read>(reader: &mut R) -> io::Result<Self> {
-        let bytes = <Vec<u8> as BinarySerializable>::deserialize(reader)?;
-        Ok(Facet(bytes))
+        Ok(Facet(<String as BinarySerializable>::deserialize(reader)?))
     }
 }
 
 impl Display for Facet {
     fn fmt(&self, f: &mut Formatter) -> fmt::Result {
-        for step in self.0.split(|&b| b == FACET_SEP_BYTE) {
+        for step in self.0.split(FACET_SEP_CHAR) {
             write!(f, "/")?;
-            let step_str = unsafe { str::from_utf8_unchecked(step) };
-            write!(f, "{}", escape_slashes(step_str))?;
+            write!(f, "{}", escape_slashes(step))?;
         }
         Ok(())
     }

src/schema/field_entry.rs

@@ -14,7 +14,7 @@ use std::fmt;
 /// - a field name
 /// - a field type, itself wrapping up options describing
 /// how the field should be indexed.
-#[derive(Clone, Debug)]
+#[derive(Clone, Debug, Eq, PartialEq)]
 pub struct FieldEntry {
     name: String,
     field_type: FieldType,

src/schema/mod.rs

@@ -27,7 +27,7 @@ directory.
 
 ```
 use tantivy::schema::*;
-let mut schema_builder = SchemaBuilder::default();
+let mut schema_builder = Schema::builder();
 let title_options = TextOptions::default()
     .set_stored()
     .set_indexing_options(TextFieldIndexing::default()
@@ -44,11 +44,11 @@ We can split the problem of generating a search result page into two phases :
   the search results page. (`doc_ids[] -> Document[]`)
 
 In the first phase, the ability to search for documents by the given field is determined by the
-[`TextIndexingOptions`](enum.TextIndexingOptions.html) of our [`TextOptions`]
-(struct.TextOptions.html).
+[`TextIndexingOptions`](enum.TextIndexingOptions.html) of our
+[`TextOptions`](struct.TextOptions.html).
 
-The effect of each possible setting is described more in detail [`TextIndexingOptions`]
-(enum.TextIndexingOptions.html).
+The effect of each possible setting is described more in detail
+[`TextIndexingOptions`](enum.TextIndexingOptions.html).
 
 On the other hand setting the field as stored or not determines whether the field should be returned
 when [`searcher.doc(doc_address)`](../struct.Searcher.html#method.doc) is called.
@@ -62,7 +62,7 @@ The example can be rewritten :
 
 ```
 use tantivy::schema::*;
-let mut schema_builder = SchemaBuilder::default();
+let mut schema_builder = Schema::builder();
 schema_builder.add_text_field("title_options", TEXT | STORED);
 let schema = schema_builder.build();
 ```
@@ -75,7 +75,7 @@ let schema = schema_builder.build();
 
 ```
 use tantivy::schema::*;
-let mut schema_builder = SchemaBuilder::default();
+let mut schema_builder = Schema::builder();
 let num_stars_options = IntOptions::default()
     .set_stored()
     .set_indexed();

src/schema/schema.rs

@@ -23,13 +23,14 @@ use std::fmt;
 /// ```
 /// use tantivy::schema::*;
 ///
-/// let mut schema_builder = SchemaBuilder::default();
+/// let mut schema_builder = Schema::builder();
 /// let id_field = schema_builder.add_text_field("id", STRING);
 /// let title_field = schema_builder.add_text_field("title", TEXT);
 /// let body_field = schema_builder.add_text_field("body", TEXT);
 /// let schema = schema_builder.build();
 ///
 /// ```
+#[derive(Default)]
 pub struct SchemaBuilder {
     fields: Vec<FieldEntry>,
     fields_map: HashMap<String, Field>,
@@ -120,20 +121,19 @@ impl SchemaBuilder {
     }
 }
 
-impl Default for SchemaBuilder {
-    fn default() -> SchemaBuilder {
-        SchemaBuilder {
-            fields: Vec::new(),
-            fields_map: HashMap::new(),
-        }
-    }
-}
-
 struct InnerSchema {
     fields: Vec<FieldEntry>,
     fields_map: HashMap<String, Field>, // transient
 }
 
+impl PartialEq for InnerSchema {
+    fn eq(&self, other: &InnerSchema) -> bool {
+        self.fields == other.fields
+    }
+}
+
+impl Eq for InnerSchema {}
+
 /// Tantivy has a very strict schema.
 /// You need to specify in advance, whether a field is indexed or not,
 /// stored or not, and RAM-based or not.
@@ -147,14 +147,14 @@ struct InnerSchema {
 /// ```
 /// use tantivy::schema::*;
 ///
-/// let mut schema_builder = SchemaBuilder::default();
+/// let mut schema_builder = Schema::builder();
 /// let id_field = schema_builder.add_text_field("id", STRING);
 /// let title_field = schema_builder.add_text_field("title", TEXT);
 /// let body_field = schema_builder.add_text_field("body", TEXT);
 /// let schema = schema_builder.build();
 ///
 /// ```
-#[derive(Clone)]
+#[derive(Clone, Eq, PartialEq)]
 pub struct Schema(Arc<InnerSchema>);
 
 impl Schema {
@@ -173,6 +173,11 @@ impl Schema {
         &self.0.fields
     }
 
+    /// Creates a new builder.
+    pub fn builder() -> SchemaBuilder {
+        SchemaBuilder::default()
+    }
+
     /// Returns the field options associated with a given name.
     ///
     /// # Panics
@@ -227,12 +232,14 @@ impl Schema {
                     let field_entry = self.get_field_entry(field);
                     let field_type = field_entry.field_type();
                     match *json_value {
-                        JsonValue::Array(ref json_items) => for json_item in json_items {
-                            let value = field_type
-                                .value_from_json(json_item)
-                                .map_err(|e| DocParsingError::ValueError(field_name.clone(), e))?;
-                            doc.add(FieldValue::new(field, value));
-                        },
+                        JsonValue::Array(ref json_items) => {
+                            for json_item in json_items {
+                                let value = field_type.value_from_json(json_item).map_err(|e| {
+                                    DocParsingError::ValueError(field_name.clone(), e)
+                                })?;
+                                doc.add(FieldValue::new(field, value));
+                            }
+                        }
                         _ => {
                             let value = field_type
                                 .value_from_json(json_value)
@@ -318,7 +325,7 @@ mod tests {
 
     #[test]
     pub fn is_indexed_test() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let field_str = schema_builder.add_text_field("field_str", STRING);
         let schema = schema_builder.build();
         assert!(schema.get_field_entry(field_str).is_indexed());
@@ -326,7 +333,7 @@ mod tests {
 
     #[test]
     pub fn test_schema_serialization() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let count_options = IntOptions::default()
             .set_stored()
             .set_fast(Cardinality::SingleValue);
@@ -395,7 +402,7 @@ mod tests {
 
     #[test]
     pub fn test_document_to_json() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let count_options = IntOptions::default()
             .set_stored()
             .set_fast(Cardinality::SingleValue);
@@ -416,7 +423,7 @@ mod tests {
 
     #[test]
     pub fn test_parse_document() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let count_options = IntOptions::default()
             .set_stored()
             .set_fast(Cardinality::SingleValue);
@@ -441,7 +448,8 @@ mod tests {
                 "count": 4,
                 "popularity": 10
             }"#,
-                ).unwrap();
+                )
+                .unwrap();
             assert_eq!(doc.get_first(title_field).unwrap().text(), Some("my title"));
             assert_eq!(
                 doc.get_first(author_field).unwrap().text(),

src/schema/term.rs

@@ -32,7 +32,7 @@ impl Term {
 
     /// Creates a `Term` given a facet.
     pub fn from_facet(field: Field, facet: &Facet) -> Term {
-        let bytes = facet.encoded_bytes();
+        let bytes = facet.encoded_str().as_bytes();
         let buffer = Vec::with_capacity(4 + bytes.len());
         let mut term = Term(buffer);
         term.set_field(field);
@@ -68,12 +68,7 @@ impl Term {
         term
     }
 
-    /// Creates a new Term with an empty buffer,
-    /// but with a given capacity.
-    ///
-    /// It is declared unsafe, as the term content
-    /// is not initialized, and a call to `.field()`
-    /// would panic.
+    /// Creates a new Term for a given field.
     pub(crate) fn for_field(field: Field) -> Term {
         let mut term = Term(Vec::with_capacity(100));
         term.set_field(field);
@@ -201,7 +196,7 @@ mod tests {
 
     #[test]
     pub fn test_term() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         schema_builder.add_text_field("text", STRING);
         let title_field = schema_builder.add_text_field("title", STRING);
         let count_field = schema_builder.add_text_field("count", STRING);

src/schema/text_options.rs

@@ -141,7 +141,7 @@ mod tests {
             assert!(field_options.get_indexing_options().is_some());
         }
         {
-            let mut schema_builder = SchemaBuilder::default();
+            let mut schema_builder = Schema::builder();
             schema_builder.add_text_field("body", TEXT);
             let schema = schema_builder.build();
             let field = schema.get_field("body").unwrap();

src/snippet/mod.rs

@@ -22,6 +22,11 @@ impl HighlightSection {
     fn new(start: usize, stop: usize) -> HighlightSection {
         HighlightSection { start, stop }
     }
+
+    /// Returns the bounds of the `HighlightSection`.
+    pub fn bounds(&self) -> (usize, usize) {
+        (self.start, self.stop)
+    }
 }
 
 #[derive(Debug)]
@@ -65,6 +70,8 @@ impl FragmentCandidate {
     }
 }
 
+/// `Snippet`
+/// Contains a fragment of a document, and some highlighed parts inside it.
 #[derive(Debug)]
 pub struct Snippet {
     fragments: String,
@@ -75,6 +82,7 @@ const HIGHLIGHTEN_PREFIX: &str = "<b>";
 const HIGHLIGHTEN_POSTFIX: &str = "</b>";
 
 impl Snippet {
+    /// Create a new, empty, `Snippet`
     pub fn empty() -> Snippet {
         Snippet {
             fragments: String::new(),
@@ -99,6 +107,16 @@ impl Snippet {
         ));
         html
     }
+
+    /// Returns a fragment from the `Snippet`.
+    pub fn fragments(&self) -> &str {
+        &self.fragments
+    }
+
+    /// Returns a list of higlighted positions from the `Snippet`.
+    pub fn highlighted(&self) -> &[HighlightSection] {
+        &self.highlighted
+    }
 }
 
 /// Returns a non-empty list of "good" fragments.
@@ -174,7 +192,8 @@ fn select_best_fragment_combination(fragments: &[FragmentCandidate], text: &str)
                     item.start - fragment.start_offset,
                     item.stop - fragment.start_offset,
                 )
-            }).collect();
+            })
+            .collect();
         Snippet {
             fragments: fragment_text.to_string(),
             highlighted,
@@ -197,12 +216,12 @@ fn select_best_fragment_combination(fragments: &[FragmentCandidate], text: &str)
 /// # #[macro_use]
 /// # extern crate tantivy;
 /// # use tantivy::Index;
-/// # use tantivy::schema::{SchemaBuilder, TEXT};
+/// # use tantivy::schema::{Schema, TEXT};
 /// # use tantivy::query::QueryParser;
 /// use tantivy::SnippetGenerator;
 ///
 /// # fn main() -> tantivy::Result<()> {
-/// #    let mut schema_builder = SchemaBuilder::default();
+/// #    let mut schema_builder = Schema::builder();
 /// #    let text_field = schema_builder.add_text_field("text", TEXT);
 /// #    let schema = schema_builder.build();
 /// #    let index = Index::create_in_ram(schema);
@@ -224,7 +243,7 @@ fn select_best_fragment_combination(fragments: &[FragmentCandidate], text: &str)
 /// let query = query_parser.parse_query("haleurs flamands").unwrap();
 /// # index.load_searchers()?;
 /// # let searcher = index.searcher();
-/// let mut snippet_generator = SnippetGenerator::new(&searcher, &*query, text_field)?;
+/// let mut snippet_generator = SnippetGenerator::create(&searcher, &*query, text_field)?;
 /// snippet_generator.set_max_num_chars(100);
 /// let snippet = snippet_generator.snippet_from_doc(&doc);
 /// let snippet_html: String = snippet.to_html();
@@ -241,7 +260,7 @@ pub struct SnippetGenerator {
 
 impl SnippetGenerator {
     /// Creates a new snippet generator
-    pub fn new(searcher: &Searcher, query: &Query, field: Field) -> Result<SnippetGenerator> {
+    pub fn create(searcher: &Searcher, query: &Query, field: Field) -> Result<SnippetGenerator> {
         let mut terms = BTreeSet::new();
         query.query_terms(&mut terms);
         let terms_text: BTreeMap<String, f32> = terms
@@ -306,7 +325,7 @@ impl SnippetGenerator {
 mod tests {
     use super::{search_fragments, select_best_fragment_combination};
     use query::QueryParser;
-    use schema::{IndexRecordOption, SchemaBuilder, TextFieldIndexing, TextOptions, TEXT};
+    use schema::{IndexRecordOption, Schema, TextFieldIndexing, TextOptions, TEXT};
     use std::collections::BTreeMap;
     use std::iter::Iterator;
     use tokenizer::{box_tokenizer, SimpleTokenizer};
@@ -328,8 +347,6 @@ to the project are from community members.[15]
 Rust won first place for "most loved programming language" in the Stack Overflow Developer
 Survey in 2016, 2017, and 2018."#;
 
-
-
     #[test]
     fn test_snippet() {
         let boxed_tokenizer = box_tokenizer(SimpleTokenizer);
@@ -345,13 +362,18 @@ Survey in 2016, 2017, and 2018."#;
             assert_eq!(first.stop_offset, 89);
         }
         let snippet = select_best_fragment_combination(&fragments[..], &TEST_TEXT);
-        assert_eq!(snippet.fragments, "Rust is a systems programming language sponsored by \
-         Mozilla which\ndescribes it as a \"safe");
-        assert_eq!(snippet.to_html(), "<b>Rust</b> is a systems programming <b>language</b> \
-         sponsored by Mozilla which\ndescribes it as a &quot;safe")
+        assert_eq!(
+            snippet.fragments,
+            "Rust is a systems programming language sponsored by \
+             Mozilla which\ndescribes it as a \"safe"
+        );
+        assert_eq!(
+            snippet.to_html(),
+            "<b>Rust</b> is a systems programming <b>language</b> \
+             sponsored by Mozilla which\ndescribes it as a &quot;safe"
+        )
     }
 
-
     #[test]
     fn test_snippet_scored_fragment() {
         let boxed_tokenizer = box_tokenizer(SimpleTokenizer);
@@ -385,10 +407,8 @@ Survey in 2016, 2017, and 2018."#;
             let snippet = select_best_fragment_combination(&fragments[..], &TEST_TEXT);
             assert_eq!(snippet.to_html(), "programming <b>language</b>")
         }
-
     }
 
-
     #[test]
     fn test_snippet_in_second_fragment() {
         let boxed_tokenizer = box_tokenizer(SimpleTokenizer);
@@ -495,10 +515,9 @@ Survey in 2016, 2017, and 2018."#;
         assert_eq!(snippet.to_html(), "");
     }
 
-
     #[test]
     fn test_snippet_generator_term_score() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -515,29 +534,42 @@ Survey in 2016, 2017, and 2018."#;
         let query_parser = QueryParser::for_index(&index, vec![text_field]);
         {
             let query = query_parser.parse_query("e").unwrap();
-            let snippet_generator = SnippetGenerator::new(&searcher, &*query, text_field).unwrap();
+            let snippet_generator =
+                SnippetGenerator::create(&searcher, &*query, text_field).unwrap();
             assert!(snippet_generator.terms_text().is_empty());
         }
         {
             let query = query_parser.parse_query("a").unwrap();
-            let snippet_generator = SnippetGenerator::new(&searcher, &*query, text_field).unwrap();
-            assert_eq!(&btreemap!("a".to_string() => 0.25f32), snippet_generator.terms_text());
+            let snippet_generator =
+                SnippetGenerator::create(&searcher, &*query, text_field).unwrap();
+            assert_eq!(
+                &btreemap!("a".to_string() => 0.25f32),
+                snippet_generator.terms_text()
+            );
         }
         {
             let query = query_parser.parse_query("a b").unwrap();
-            let snippet_generator = SnippetGenerator::new(&searcher, &*query, text_field).unwrap();
-            assert_eq!(&btreemap!("a".to_string() => 0.25f32, "b".to_string() => 0.5), snippet_generator.terms_text());
+            let snippet_generator =
+                SnippetGenerator::create(&searcher, &*query, text_field).unwrap();
+            assert_eq!(
+                &btreemap!("a".to_string() => 0.25f32, "b".to_string() => 0.5),
+                snippet_generator.terms_text()
+            );
         }
         {
             let query = query_parser.parse_query("a b c").unwrap();
-            let snippet_generator = SnippetGenerator::new(&searcher, &*query, text_field).unwrap();
-            assert_eq!(&btreemap!("a".to_string() => 0.25f32, "b".to_string() => 0.5), snippet_generator.terms_text());
+            let snippet_generator =
+                SnippetGenerator::create(&searcher, &*query, text_field).unwrap();
+            assert_eq!(
+                &btreemap!("a".to_string() => 0.25f32, "b".to_string() => 0.5),
+                snippet_generator.terms_text()
+            );
         }
     }
 
     #[test]
     fn test_snippet_generator() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_options = TextOptions::default().set_indexing_options(
             TextFieldIndexing::default()
                 .set_tokenizer("en_stem")
@@ -559,7 +591,8 @@ Survey in 2016, 2017, and 2018."#;
         let searcher = index.searcher();
         let query_parser = QueryParser::for_index(&index, vec![text_field]);
         let query = query_parser.parse_query("rust design").unwrap();
-        let mut snippet_generator = SnippetGenerator::new(&searcher, &*query, text_field).unwrap();
+        let mut snippet_generator =
+            SnippetGenerator::create(&searcher, &*query, text_field).unwrap();
         {
             let snippet = snippet_generator.snippet(TEST_TEXT);
             assert_eq!(snippet.to_html(), "imperative-procedural paradigms. <b>Rust</b> is syntactically similar to C++[according to whom?],\nbut its <b>designers</b> intend it to provide better memory safety");

src/space_usage/mod.rs

@@ -0,0 +1,497 @@
+/*!
+Representations for the space usage of various parts of a Tantivy index.
+
+This can be used programmatically, and will also be exposed in a human readable fashion in
+tantivy-cli.
+
+One important caveat for all of this functionality is that none of it currently takes storage-level
+details into consideration. For example, if your file system block size is 4096 bytes, we can
+under-count actual resultant space usage by up to 4095 bytes per file.
+*/
+
+use schema::Field;
+use std::collections::HashMap;
+use SegmentComponent;
+
+/// Indicates space usage in bytes
+pub type ByteCount = usize;
+
+/// Enum containing any of the possible space usage results for segment components.
+pub enum ComponentSpaceUsage {
+    /// Data is stored per field in a uniform way
+    PerField(PerFieldSpaceUsage),
+    /// Data is stored in separate pieces in the store
+    Store(StoreSpaceUsage),
+    /// Some sort of raw byte count
+    Basic(ByteCount),
+}
+
+/// Represents combined space usage of an entire searcher and its component segments.
+#[derive(Clone, Debug, Serialize, Deserialize)]
+pub struct SearcherSpaceUsage {
+    segments: Vec<SegmentSpaceUsage>,
+    total: ByteCount,
+}
+
+impl SearcherSpaceUsage {
+    pub(crate) fn new() -> SearcherSpaceUsage {
+        SearcherSpaceUsage {
+            segments: Vec::new(),
+            total: 0,
+        }
+    }
+
+    /// Add a segment, to `self`.
+    /// Performs no deduplication or other intelligence.
+    pub(crate) fn add_segment(&mut self, segment: SegmentSpaceUsage) {
+        self.total += segment.total();
+        self.segments.push(segment);
+    }
+
+    /// Per segment space usage
+    pub fn segments(&self) -> &[SegmentSpaceUsage] {
+        &self.segments[..]
+    }
+
+    /// Returns total byte usage of this searcher, including all large subcomponents.
+    /// Does not account for smaller things like `meta.json`.
+    pub fn total(&self) -> ByteCount {
+        self.total
+    }
+}
+
+/// Represents combined space usage for all of the large components comprising a segment.
+#[derive(Clone, Debug, Serialize, Deserialize)]
+pub struct SegmentSpaceUsage {
+    num_docs: u32,
+
+    termdict: PerFieldSpaceUsage,
+    postings: PerFieldSpaceUsage,
+    positions: PerFieldSpaceUsage,
+    positions_idx: PerFieldSpaceUsage,
+    fast_fields: PerFieldSpaceUsage,
+    fieldnorms: PerFieldSpaceUsage,
+
+    store: StoreSpaceUsage,
+
+    deletes: ByteCount,
+
+    total: ByteCount,
+}
+
+impl SegmentSpaceUsage {
+    #[allow(clippy::too_many_arguments)]
+    pub(crate) fn new(
+        num_docs: u32,
+        termdict: PerFieldSpaceUsage,
+        postings: PerFieldSpaceUsage,
+        positions: PerFieldSpaceUsage,
+        positions_idx: PerFieldSpaceUsage,
+        fast_fields: PerFieldSpaceUsage,
+        fieldnorms: PerFieldSpaceUsage,
+        store: StoreSpaceUsage,
+        deletes: ByteCount,
+    ) -> SegmentSpaceUsage {
+        let total = termdict.total()
+            + postings.total()
+            + positions.total()
+            + fast_fields.total()
+            + fieldnorms.total()
+            + store.total()
+            + deletes;
+        SegmentSpaceUsage {
+            num_docs,
+            termdict,
+            postings,
+            positions,
+            positions_idx,
+            fast_fields,
+            fieldnorms,
+            store,
+            deletes,
+            total,
+        }
+    }
+
+    /// Space usage for the given component
+    ///
+    /// Clones the underlying data.
+    /// Use the components directly if this is somehow in performance critical code.
+    pub fn component(&self, component: SegmentComponent) -> ComponentSpaceUsage {
+        use self::ComponentSpaceUsage::*;
+        use SegmentComponent::*;
+        match component {
+            POSTINGS => PerField(self.postings().clone()),
+            POSITIONS => PerField(self.positions().clone()),
+            POSITIONSSKIP => PerField(self.positions_skip_idx().clone()),
+            FASTFIELDS => PerField(self.fast_fields().clone()),
+            FIELDNORMS => PerField(self.fieldnorms().clone()),
+            TERMS => PerField(self.termdict().clone()),
+            STORE => Store(self.store().clone()),
+            DELETE => Basic(self.deletes()),
+        }
+    }
+
+    /// Num docs in segment
+    pub fn num_docs(&self) -> u32 {
+        self.num_docs
+    }
+
+    /// Space usage for term dictionary
+    pub fn termdict(&self) -> &PerFieldSpaceUsage {
+        &self.termdict
+    }
+
+    /// Space usage for postings list
+    pub fn postings(&self) -> &PerFieldSpaceUsage {
+        &self.postings
+    }
+
+    /// Space usage for positions
+    pub fn positions(&self) -> &PerFieldSpaceUsage {
+        &self.positions
+    }
+
+    /// Space usage for positions skip idx
+    pub fn positions_skip_idx(&self) -> &PerFieldSpaceUsage {
+        &self.positions_idx
+    }
+
+    /// Space usage for fast fields
+    pub fn fast_fields(&self) -> &PerFieldSpaceUsage {
+        &self.fast_fields
+    }
+
+    /// Space usage for field norms
+    pub fn fieldnorms(&self) -> &PerFieldSpaceUsage {
+        &self.fieldnorms
+    }
+
+    /// Space usage for stored documents
+    pub fn store(&self) -> &StoreSpaceUsage {
+        &self.store
+    }
+
+    /// Space usage for document deletions
+    pub fn deletes(&self) -> ByteCount {
+        self.deletes
+    }
+
+    /// Total space usage in bytes for this segment.
+    pub fn total(&self) -> ByteCount {
+        self.total
+    }
+}
+
+/// Represents space usage for the Store for this segment.
+///
+/// This is composed of two parts.
+/// `data` represents the compressed data itself.
+/// `offsets` represents a lookup to find the start of a block
+#[derive(Clone, Debug, Serialize, Deserialize)]
+pub struct StoreSpaceUsage {
+    data: ByteCount,
+    offsets: ByteCount,
+}
+
+impl StoreSpaceUsage {
+    pub(crate) fn new(data: ByteCount, offsets: ByteCount) -> StoreSpaceUsage {
+        StoreSpaceUsage { data, offsets }
+    }
+
+    /// Space usage for the data part of the store
+    pub fn data_usage(&self) -> ByteCount {
+        self.data
+    }
+
+    /// Space usage for the offsets part of the store (doc ID -> offset)
+    pub fn offsets_usage(&self) -> ByteCount {
+        self.offsets
+    }
+
+    /// Total space usage in bytes for this Store
+    pub fn total(&self) -> ByteCount {
+        self.data + self.offsets
+    }
+}
+
+/// Represents space usage for all of the (field, index) pairs that appear in a CompositeFile.
+///
+/// A field can appear with a single index (typically 0) or with multiple indexes.
+/// Multiple indexes are used to handle variable length things, where
+#[derive(Clone, Debug, Serialize, Deserialize)]
+pub struct PerFieldSpaceUsage {
+    fields: HashMap<Field, FieldUsage>,
+    total: ByteCount,
+}
+
+impl PerFieldSpaceUsage {
+    pub(crate) fn new(fields: HashMap<Field, FieldUsage>) -> PerFieldSpaceUsage {
+        let total = fields.values().map(|x| x.total()).sum();
+        PerFieldSpaceUsage { fields, total }
+    }
+
+    /// Per field space usage
+    pub fn fields(&self) -> impl Iterator<Item = (&Field, &FieldUsage)> {
+        self.fields.iter()
+    }
+
+    /// Bytes used by the represented file
+    pub fn total(&self) -> ByteCount {
+        self.total
+    }
+}
+
+/// Represents space usage of a given field, breaking it down into the (field, index) pairs that
+/// comprise it.
+///
+/// See documentation for PerFieldSpaceUsage for slightly more information.
+#[derive(Clone, Debug, Serialize, Deserialize)]
+pub struct FieldUsage {
+    field: Field,
+    num_bytes: ByteCount,
+    /// A field can be composed of more than one piece.
+    /// These pieces are indexed by arbitrary numbers starting at zero.
+    /// `self.num_bytes` includes all of `self.sub_num_bytes`.
+    sub_num_bytes: Vec<Option<ByteCount>>,
+}
+
+impl FieldUsage {
+    pub(crate) fn empty(field: Field) -> FieldUsage {
+        FieldUsage {
+            field,
+            num_bytes: 0,
+            sub_num_bytes: Vec::new(),
+        }
+    }
+
+    pub(crate) fn add_field_idx(&mut self, idx: usize, size: ByteCount) {
+        if self.sub_num_bytes.len() < idx + 1 {
+            self.sub_num_bytes.resize(idx + 1, None);
+        }
+        assert!(self.sub_num_bytes[idx].is_none());
+        self.sub_num_bytes[idx] = Some(size);
+        self.num_bytes += size
+    }
+
+    /// Field
+    pub fn field(&self) -> Field {
+        self.field
+    }
+
+    /// Space usage for each index
+    pub fn sub_num_bytes(&self) -> &[Option<ByteCount>] {
+        &self.sub_num_bytes[..]
+    }
+
+    /// Total bytes used for this field in this context
+    pub fn total(&self) -> ByteCount {
+        self.num_bytes
+    }
+}
+
+#[cfg(test)]
+mod test {
+    use core::Index;
+    use schema::Field;
+    use schema::Schema;
+    use schema::STORED;
+    use schema::{FAST, INT_INDEXED, TEXT};
+    use space_usage::ByteCount;
+    use space_usage::PerFieldSpaceUsage;
+    use Term;
+
+    #[test]
+    fn test_empty() {
+        let schema = Schema::builder().build();
+        let index = Index::create_in_ram(schema.clone());
+
+        index.load_searchers().unwrap();
+        let searcher = index.searcher();
+        let searcher_space_usage = searcher.space_usage();
+        assert_eq!(0, searcher_space_usage.total());
+    }
+
+    fn expect_single_field(
+        field_space: &PerFieldSpaceUsage,
+        field: &Field,
+        min_size: ByteCount,
+        max_size: ByteCount,
+    ) {
+        assert!(field_space.total() >= min_size);
+        assert!(field_space.total() <= max_size);
+        assert_eq!(
+            vec![(field, field_space.total())],
+            field_space
+                .fields()
+                .map(|(x, y)| (x, y.total()))
+                .collect::<Vec<_>>()
+        );
+    }
+
+    #[test]
+    fn test_fast_indexed() {
+        let mut schema_builder = Schema::builder();
+        let name = schema_builder.add_u64_field("name", FAST | INT_INDEXED);
+        let schema = schema_builder.build();
+        let index = Index::create_in_ram(schema.clone());
+
+        {
+            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            index_writer.add_document(doc!(name => 1u64));
+            index_writer.add_document(doc!(name => 2u64));
+            index_writer.add_document(doc!(name => 10u64));
+            index_writer.add_document(doc!(name => 20u64));
+            index_writer.commit().unwrap();
+        }
+
+        index.load_searchers().unwrap();
+        let searcher = index.searcher();
+        let searcher_space_usage = searcher.space_usage();
+        assert!(searcher_space_usage.total() > 0);
+        assert_eq!(1, searcher_space_usage.segments().len());
+
+        let segment = &searcher_space_usage.segments()[0];
+        assert!(segment.total() > 0);
+
+        assert_eq!(4, segment.num_docs());
+
+        expect_single_field(segment.termdict(), &name, 1, 512);
+        expect_single_field(segment.postings(), &name, 1, 512);
+        assert_eq!(0, segment.positions().total());
+        assert_eq!(0, segment.positions_skip_idx().total());
+        expect_single_field(segment.fast_fields(), &name, 1, 512);
+        expect_single_field(segment.fieldnorms(), &name, 1, 512);
+        // TODO: understand why the following fails
+        //        assert_eq!(0, segment.store().total());
+        assert_eq!(0, segment.deletes());
+    }
+
+    #[test]
+    fn test_text() {
+        let mut schema_builder = Schema::builder();
+        let name = schema_builder.add_text_field("name", TEXT);
+        let schema = schema_builder.build();
+        let index = Index::create_in_ram(schema.clone());
+
+        {
+            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            index_writer.add_document(doc!(name => "hi"));
+            index_writer.add_document(doc!(name => "this is a test"));
+            index_writer.add_document(
+                doc!(name => "some more documents with some word overlap with the other test"),
+            );
+            index_writer.add_document(doc!(name => "hello hi goodbye"));
+            index_writer.commit().unwrap();
+        }
+
+        index.load_searchers().unwrap();
+        let searcher = index.searcher();
+        let searcher_space_usage = searcher.space_usage();
+        assert!(searcher_space_usage.total() > 0);
+        assert_eq!(1, searcher_space_usage.segments().len());
+
+        let segment = &searcher_space_usage.segments()[0];
+        assert!(segment.total() > 0);
+
+        assert_eq!(4, segment.num_docs());
+
+        expect_single_field(segment.termdict(), &name, 1, 512);
+        expect_single_field(segment.postings(), &name, 1, 512);
+        expect_single_field(segment.positions(), &name, 1, 512);
+        expect_single_field(segment.positions_skip_idx(), &name, 1, 512);
+        assert_eq!(0, segment.fast_fields().total());
+        expect_single_field(segment.fieldnorms(), &name, 1, 512);
+        // TODO: understand why the following fails
+        //        assert_eq!(0, segment.store().total());
+        assert_eq!(0, segment.deletes());
+    }
+
+    #[test]
+    fn test_store() {
+        let mut schema_builder = Schema::builder();
+        let name = schema_builder.add_text_field("name", STORED);
+        let schema = schema_builder.build();
+        let index = Index::create_in_ram(schema.clone());
+
+        {
+            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            index_writer.add_document(doc!(name => "hi"));
+            index_writer.add_document(doc!(name => "this is a test"));
+            index_writer.add_document(
+                doc!(name => "some more documents with some word overlap with the other test"),
+            );
+            index_writer.add_document(doc!(name => "hello hi goodbye"));
+            index_writer.commit().unwrap();
+        }
+
+        index.load_searchers().unwrap();
+        let searcher = index.searcher();
+        let searcher_space_usage = searcher.space_usage();
+        assert!(searcher_space_usage.total() > 0);
+        assert_eq!(1, searcher_space_usage.segments().len());
+
+        let segment = &searcher_space_usage.segments()[0];
+        assert!(segment.total() > 0);
+
+        assert_eq!(4, segment.num_docs());
+
+        assert_eq!(0, segment.termdict().total());
+        assert_eq!(0, segment.postings().total());
+        assert_eq!(0, segment.positions().total());
+        assert_eq!(0, segment.positions_skip_idx().total());
+        assert_eq!(0, segment.fast_fields().total());
+        assert_eq!(0, segment.fieldnorms().total());
+        assert!(segment.store().total() > 0);
+        assert!(segment.store().total() < 512);
+        assert_eq!(0, segment.deletes());
+    }
+
+    #[test]
+    fn test_deletes() {
+        let mut schema_builder = Schema::builder();
+        let name = schema_builder.add_u64_field("name", INT_INDEXED);
+        let schema = schema_builder.build();
+        let index = Index::create_in_ram(schema.clone());
+
+        {
+            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            index_writer.add_document(doc!(name => 1u64));
+            index_writer.add_document(doc!(name => 2u64));
+            index_writer.add_document(doc!(name => 3u64));
+            index_writer.add_document(doc!(name => 4u64));
+            index_writer.commit().unwrap();
+        }
+
+        {
+            let mut index_writer2 = index.writer(50_000_000).unwrap();
+            index_writer2.delete_term(Term::from_field_u64(name, 2u64));
+            index_writer2.delete_term(Term::from_field_u64(name, 3u64));
+
+            // ok, now we should have a deleted doc
+            index_writer2.commit().unwrap();
+        }
+
+        index.load_searchers().unwrap();
+
+        let searcher = index.searcher();
+        let searcher_space_usage = searcher.space_usage();
+        assert!(searcher_space_usage.total() > 0);
+        assert_eq!(1, searcher_space_usage.segments().len());
+
+        let segment = &searcher_space_usage.segments()[0];
+        assert!(segment.total() > 0);
+
+        assert_eq!(2, segment.num_docs());
+
+        expect_single_field(segment.termdict(), &name, 1, 512);
+        expect_single_field(segment.postings(), &name, 1, 512);
+        assert_eq!(0, segment.positions().total());
+        assert_eq!(0, segment.positions_skip_idx().total());
+        assert_eq!(0, segment.fast_fields().total());
+        expect_single_field(segment.fieldnorms(), &name, 1, 512);
+        // TODO: understand why the following fails
+        //        assert_eq!(0, segment.store().total());
+        assert!(segment.deletes() > 0);
+    }
+}

src/store/mod.rs

@@ -56,12 +56,12 @@ pub mod tests {
     use directory::{Directory, RAMDirectory, WritePtr};
     use schema::Document;
     use schema::FieldValue;
+    use schema::Schema;
     use schema::TextOptions;
-    use schema::{Schema, SchemaBuilder};
     use std::path::Path;
 
     pub fn write_lorem_ipsum_store(writer: WritePtr, num_docs: usize) -> Schema {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let field_body = schema_builder.add_text_field("body", TextOptions::default().set_stored());
         let field_title =
             schema_builder.add_text_field("title", TextOptions::default().set_stored());

src/store/reader.rs

@@ -6,6 +6,7 @@ use common::BinarySerializable;
 use common::VInt;
 use directory::ReadOnlySource;
 use schema::Document;
+use space_usage::StoreSpaceUsage;
 use std::cell::RefCell;
 use std::io;
 use std::mem::size_of;
@@ -87,12 +88,14 @@ impl StoreReader {
         cursor = &cursor[..doc_length];
         Ok(Document::deserialize(&mut cursor)?)
     }
+
+    /// Summarize total space usage of this store reader.
+    pub fn space_usage(&self) -> StoreSpaceUsage {
+        StoreSpaceUsage::new(self.data.len(), self.offset_index_source.len())
+    }
 }
 
-#[cfg_attr(
-    feature = "cargo-clippy",
-    allow(clippy::needless_pass_by_value)
-)]
+#[cfg_attr(feature = "cargo-clippy", allow(clippy::needless_pass_by_value))]
 fn split_source(data: ReadOnlySource) -> (ReadOnlySource, ReadOnlySource, DocId) {
     let data_len = data.len();
     let footer_offset = data_len - size_of::<u64>() - size_of::<u32>();

src/termdict/merger.rs

@@ -53,7 +53,8 @@ impl<'a> TermMerger<'a> {
                 .map(|(ord, streamer)| HeapItem {
                     streamer,
                     segment_ord: ord,
-                }).collect(),
+                })
+                .collect(),
         }
     }
 
@@ -122,10 +123,7 @@ impl<'a> TermMerger<'a> {
     }
 
     /// Iterates through terms
-    #[cfg_attr(
-        feature = "cargo-clippy",
-        allow(clippy::should_implement_trait)
-    )]
+    #[cfg_attr(feature = "cargo-clippy", allow(clippy::should_implement_trait))]
     pub fn next(&mut self) -> Option<Term<&[u8]>> {
         if self.advance() {
             Some(Term::wrap(self.current_streamers[0].streamer.key()))

src/termdict/mod.rs

@@ -35,7 +35,7 @@ mod tests {
     use core::Index;
     use directory::{Directory, RAMDirectory, ReadOnlySource};
     use postings::TermInfo;
-    use schema::{Document, FieldType, SchemaBuilder, TEXT};
+    use schema::{Document, FieldType, Schema, TEXT};
     use std::path::PathBuf;
     use std::str;
 
@@ -66,7 +66,7 @@ mod tests {
             let write = directory.open_write(&path).unwrap();
             let field_type = FieldType::Str(TEXT);
             let mut term_dictionary_builder =
-                TermDictionaryBuilder::new(write, &field_type).unwrap();
+                TermDictionaryBuilder::create(write, &field_type).unwrap();
             for term in COUNTRIES.iter() {
                 term_dictionary_builder
                     .insert(term.as_bytes(), &make_term_info(0u64))
@@ -92,7 +92,7 @@ mod tests {
             let write = directory.open_write(&path).unwrap();
             let field_type = FieldType::Str(TEXT);
             let mut term_dictionary_builder =
-                TermDictionaryBuilder::new(write, &field_type).unwrap();
+                TermDictionaryBuilder::create(write, &field_type).unwrap();
             term_dictionary_builder
                 .insert("abc".as_bytes(), &make_term_info(34u64))
                 .unwrap();
@@ -129,7 +129,7 @@ mod tests {
 
     #[test]
     fn test_term_iterator() {
-        let mut schema_builder = SchemaBuilder::default();
+        let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let index = Index::create_in_ram(schema_builder.build());
         {
@@ -167,7 +167,7 @@ mod tests {
         let mut term_string = String::new();
         while term_it.advance() {
             //let term = Term::from_bytes(term_it.key());
-            term_string.push_str(unsafe { str::from_utf8_unchecked(term_it.key()) }); // ok test
+            term_string.push_str(str::from_utf8(term_it.key()).expect("test"));
         }
         assert_eq!(&*term_string, "abcdef");
     }
@@ -180,7 +180,7 @@ mod tests {
         let field_type = FieldType::Str(TEXT);
         let buffer: Vec<u8> = {
             let mut term_dictionary_builder =
-                TermDictionaryBuilder::new(vec![], &field_type).unwrap();
+                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
             for &(ref id, ref i) in &ids {
                 term_dictionary_builder
                     .insert(id.as_bytes(), &make_term_info(*i as u64))
@@ -210,7 +210,7 @@ mod tests {
         let field_type = FieldType::Str(TEXT);
         let buffer: Vec<u8> = {
             let mut term_dictionary_builder =
-                TermDictionaryBuilder::new(vec![], &field_type).unwrap();
+                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
             // term requires more than 16bits
             term_dictionary_builder
                 .insert("abcdefghijklmnopqrstuvwxy", &make_term_info(1))
@@ -245,7 +245,7 @@ mod tests {
         let field_type = FieldType::Str(TEXT);
         let buffer: Vec<u8> = {
             let mut term_dictionary_builder =
-                TermDictionaryBuilder::new(vec![], &field_type).unwrap();
+                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
             for &(ref id, ref i) in &ids {
                 term_dictionary_builder
                     .insert(id.as_bytes(), &make_term_info(*i as u64))
@@ -314,7 +314,7 @@ mod tests {
         let field_type = FieldType::Str(TEXT);
         let buffer: Vec<u8> = {
             let mut term_dictionary_builder =
-                TermDictionaryBuilder::new(vec![], &field_type).unwrap();
+                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
             term_dictionary_builder
                 .insert(&[], &make_term_info(1 as u64))
                 .unwrap();
@@ -338,7 +338,7 @@ mod tests {
         let field_type = FieldType::Str(TEXT);
         let buffer: Vec<u8> = {
             let mut term_dictionary_builder =
-                TermDictionaryBuilder::new(vec![], &field_type).unwrap();
+                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
             for i in 0u8..10u8 {
                 let number_arr = [i; 1];
                 term_dictionary_builder
@@ -408,7 +408,7 @@ mod tests {
             let write = directory.open_write(&path).unwrap();
             let field_type = FieldType::Str(TEXT);
             let mut term_dictionary_builder =
-                TermDictionaryBuilder::new(write, &field_type).unwrap();
+                TermDictionaryBuilder::create(write, &field_type).unwrap();
             for term in COUNTRIES.iter() {
                 term_dictionary_builder
                     .insert(term.as_bytes(), &make_term_info(0u64))

src/termdict/streamer.rs

@@ -132,10 +132,7 @@ where
     }
 
     /// Return the next `(key, value)` pair.
-    #[cfg_attr(
-        feature = "cargo-clippy",
-        allow(clippy::should_implement_trait)
-    )]
+    #[cfg_attr(feature = "cargo-clippy", allow(clippy::should_implement_trait))]
     pub fn next(&mut self) -> Option<(&[u8], &TermInfo)> {
         if self.advance() {
             Some((self.key(), self.value()))

src/termdict/termdict.rs

@@ -29,7 +29,7 @@ where
     W: Write,
 {
     /// Creates a new `TermDictionaryBuilder`
-    pub fn new(w: W, _field_type: &FieldType) -> io::Result<Self> {
+    pub fn create(w: W, _field_type: &FieldType) -> io::Result<Self> {
         let fst_builder = fst::MapBuilder::new(w).map_err(convert_fst_error)?;
         Ok(TermDictionaryBuilder {
             fst_builder,
@@ -96,9 +96,6 @@ fn open_fst_index(source: ReadOnlySource) -> fst::Map {
         ReadOnlySource::Mmap(mmap_readonly) => {
             Fst::from_mmap(mmap_readonly).expect("FST data is corrupted")
         }
-        ReadOnlySource::Static(data) => {
-            Fst::from_static_slice(data).expect("FST data is corrupted")
-        }
     };
     fst::Map::from(fst)
 }
@@ -135,7 +132,7 @@ impl TermDictionary {
     /// Creates an empty term dictionary which contains no terms.
     pub fn empty(field_type: &FieldType) -> Self {
         let term_dictionary_data: Vec<u8> =
-            TermDictionaryBuilder::new(Vec::<u8>::new(), &field_type)
+            TermDictionaryBuilder::create(Vec::<u8>::new(), &field_type)
                 .expect("Creating a TermDictionaryBuilder in a Vec<u8> should never fail")
                 .finish()
                 .expect("Writing in a Vec<u8> should never fail");

src/tokenizer/facet_tokenizer.rs

@@ -1,6 +1,5 @@
 use super::{Token, TokenStream, Tokenizer};
 use schema::FACET_SEP_BYTE;
-use std::str;
 
 /// The `FacetTokenizer` process a `Facet` binary representation
 /// and emits a token for all of its parent.
@@ -57,12 +56,11 @@ impl<'a> TokenStream for FacetTokenStream<'a> {
                     .position(|b| b == FACET_SEP_BYTE)
                     .map(|pos| cursor + 1 + pos)
                 {
-                    let facet_part =
-                        unsafe { str::from_utf8_unchecked(&bytes[cursor..next_sep_pos]) };
+                    let facet_part = &self.text[cursor..next_sep_pos];
                     self.token.text.push_str(facet_part);
                     self.state = State::UpToPosition(next_sep_pos);
                 } else {
-                    let facet_part = unsafe { str::from_utf8_unchecked(&bytes[cursor..]) };
+                    let facet_part = &self.text[cursor..];
                     self.token.text.push_str(facet_part);
                     self.state = State::Terminated;
                 }
@@ -86,7 +84,6 @@ mod tests {
 
     use super::FacetTokenizer;
     use schema::Facet;
-    use std::str;
     use tokenizer::{Token, TokenStream, Tokenizer};
 
     #[test]
@@ -95,11 +92,11 @@ mod tests {
         let mut tokens = vec![];
         {
             let mut add_token = |token: &Token| {
-                let facet = unsafe { Facet::from_encoded(token.text.as_bytes().to_owned()) }; // ok test
+                let facet = Facet::from_encoded(token.text.as_bytes().to_owned()).unwrap();
                 tokens.push(format!("{}", facet));
             };
             FacetTokenizer
-                .token_stream(unsafe { str::from_utf8_unchecked(facet.encoded_bytes()) })
+                .token_stream(facet.encoded_str())
                 .process(&mut add_token);
         }
         assert_eq!(tokens.len(), 4);
@@ -115,11 +112,11 @@ mod tests {
         let mut tokens = vec![];
         {
             let mut add_token = |token: &Token| {
-                let facet = unsafe { Facet::from_encoded(token.text.as_bytes().to_owned()) }; // ok test
+                let facet = Facet::from_encoded(token.text.as_bytes().to_owned()).unwrap(); // ok test
                 tokens.push(format!("{}", facet));
             };
             FacetTokenizer
-                .token_stream(unsafe { str::from_utf8_unchecked(facet.encoded_bytes()) }) // ok test
+                .token_stream(facet.encoded_str()) // ok test
                 .process(&mut add_token);
         }
         assert_eq!(tokens.len(), 1);

src/tokenizer/mod.rs

@@ -9,7 +9,7 @@
 //! use tantivy::schema::*;
 //!
 //! # fn main() {
-//! let mut schema_builder = SchemaBuilder::new();
+//! let mut schema_builder = Schema::builder();
 //!
 //! let text_options = TextOptions::default()
 //!     .set_indexing_options(
@@ -82,12 +82,12 @@
 //!
 //! ```
 //! # extern crate tantivy;
-//! # use tantivy::schema::SchemaBuilder;
+//! # use tantivy::schema::Schema;
 //! # use tantivy::tokenizer::*;
 //! # use tantivy::Index;
 //! # fn main() {
 //! # let custom_en_tokenizer = SimpleTokenizer;
-//! # let schema = SchemaBuilder::new().build();
+//! # let schema = Schema::builder().build();
 //! let index = Index::create_in_ram(schema);
 //! index.tokenizers()
 //!      .register("custom_en", custom_en_tokenizer);
@@ -101,12 +101,12 @@
 //!
 //! ```
 //! extern crate tantivy;
-//! use tantivy::schema::{SchemaBuilder, IndexRecordOption, TextOptions, TextFieldIndexing};
+//! use tantivy::schema::{Schema, IndexRecordOption, TextOptions, TextFieldIndexing};
 //! use tantivy::tokenizer::*;
 //! use tantivy::Index;
 //!
 //! # fn main() {
-//! let mut schema_builder = SchemaBuilder::new();
+//! let mut schema_builder = Schema::builder();
 //! let text_field_indexing = TextFieldIndexing::default()
 //!     .set_tokenizer("custom_en")
 //!     .set_index_option(IndexRecordOption::WithFreqsAndPositions);
@@ -157,35 +157,32 @@ pub use self::tokenizer::BoxedTokenizer;
 pub use self::tokenizer::{Token, TokenFilter, TokenStream, Tokenizer};
 pub use self::tokenizer_manager::TokenizerManager;
 
-/// This is a function that can be used in tests and doc tests
-/// to assert a token's correctness.
-/// TODO: can this be wrapped in #[cfg(test)] so as not to be in the
-/// public api?
-pub fn assert_token(token: &Token, position: usize, text: &str, from: usize, to: usize) {
-    assert_eq!(
-        token.position, position,
-        "expected position {} but {:?}",
-        position, token
-    );
-    assert_eq!(token.text, text, "expected text {} but {:?}", text, token);
-    assert_eq!(
-        token.offset_from, from,
-        "expected offset_from {} but {:?}",
-        from, token
-    );
-    assert_eq!(
-        token.offset_to, to,
-        "expected offset_to {} but {:?}",
-        to, token
-    );
-}
-
 #[cfg(test)]
-pub mod test {
-    use super::assert_token;
+pub mod tests {
     use super::Token;
     use super::TokenizerManager;
 
+    /// This is a function that can be used in tests and doc tests
+    /// to assert a token's correctness.
+    pub fn assert_token(token: &Token, position: usize, text: &str, from: usize, to: usize) {
+        assert_eq!(
+            token.position, position,
+            "expected position {} but {:?}",
+            position, token
+        );
+        assert_eq!(token.text, text, "expected text {} but {:?}", text, token);
+        assert_eq!(
+            token.offset_from, from,
+            "expected offset_from {} but {:?}",
+            from, token
+        );
+        assert_eq!(
+            token.offset_to, to,
+            "expected offset_to {} but {:?}",
+            to, token
+        );
+    }
+
     #[test]
     fn test_raw_tokenizer() {
         let tokenizer_manager = TokenizerManager::default();
@@ -224,72 +221,6 @@ pub mod test {
         assert_token(&tokens[3], 3, "payer", 17, 22);
     }
 
-    #[test]
-    fn test_ngram_tokenizer() {
-        use super::{LowerCaser, NgramTokenizer};
-        use tokenizer::tokenizer::TokenStream;
-        use tokenizer::tokenizer::Tokenizer;
-
-        let tokenizer_manager = TokenizerManager::default();
-        tokenizer_manager.register("ngram12", NgramTokenizer::new(1, 2, false));
-        tokenizer_manager.register(
-            "ngram3",
-            NgramTokenizer::new(3, 3, false).filter(LowerCaser),
-        );
-        tokenizer_manager.register(
-            "edgegram5",
-            NgramTokenizer::new(2, 5, true).filter(LowerCaser),
-        );
-
-        let tokenizer = NgramTokenizer::new(1, 2, false);
-        let mut tokens: Vec<Token> = vec![];
-        {
-            let mut add_token = |token: &Token| {
-                tokens.push(token.clone());
-            };
-            tokenizer.token_stream("hello").process(&mut add_token);
-        }
-        assert_eq!(tokens.len(), 9);
-        assert_token(&tokens[0], 0, "h", 0, 1);
-        assert_token(&tokens[1], 0, "he", 0, 2);
-        assert_token(&tokens[2], 1, "e", 1, 2);
-        assert_token(&tokens[3], 1, "el", 1, 3);
-        assert_token(&tokens[4], 2, "l", 2, 3);
-        assert_token(&tokens[5], 2, "ll", 2, 4);
-        assert_token(&tokens[6], 3, "l", 3, 4);
-        assert_token(&tokens[7], 3, "lo", 3, 5);
-        assert_token(&tokens[8], 4, "o", 4, 5);
-
-        let tokenizer = tokenizer_manager.get("ngram3").unwrap();
-        let mut tokens: Vec<Token> = vec![];
-        {
-            let mut add_token = |token: &Token| {
-                tokens.push(token.clone());
-            };
-            tokenizer.token_stream("Hello").process(&mut add_token);
-        }
-        assert_eq!(tokens.len(), 3);
-        assert_token(&tokens[0], 0, "hel", 0, 3);
-        assert_token(&tokens[1], 1, "ell", 1, 4);
-        assert_token(&tokens[2], 2, "llo", 2, 5);
-
-        let tokenizer = tokenizer_manager.get("edgegram5").unwrap();
-        let mut tokens: Vec<Token> = vec![];
-        {
-            let mut add_token = |token: &Token| {
-                tokens.push(token.clone());
-            };
-            tokenizer
-                .token_stream("Frankenstein")
-                .process(&mut add_token);
-        }
-        assert_eq!(tokens.len(), 4);
-        assert_token(&tokens[0], 0, "fr", 0, 2);
-        assert_token(&tokens[1], 0, "fra", 0, 3);
-        assert_token(&tokens[2], 0, "fran", 0, 4);
-        assert_token(&tokens[3], 0, "frank", 0, 5);
-    }
-
     #[test]
     fn test_tokenizer_empty() {
         let tokenizer_manager = TokenizerManager::default();

src/tokenizer/ngram_tokenizer.rs

@@ -2,14 +2,15 @@ use super::{Token, TokenStream, Tokenizer};
 
 /// Tokenize the text by splitting words into n-grams of the given size(s)
 ///
-/// With this tokenizer, the `position` field expresses the starting offset of the ngram
-/// rather than the `token` offset.
+/// With this tokenizer, the `position` is always 0.
+/// Beware however, in presence of multiple value for the same field,
+/// the position will be `POSITION_GAP * index of value`.
 ///
 /// Example 1: `hello` would be tokenized as (min_gram: 2, max_gram: 3, prefix_only: false)
 ///
 /// | Term     | he  | hel | el  | ell | ll  | llo | lo |
 /// |----------|-----|-----|-----|-----|-----|-----|----|
-/// | Position | 0   | 0   | 1   | 1   | 2   | 2   | 3  |
+/// | Position | 0   | 0   | 0   | 0   | 0   | 0   | 0  |
 /// | Offsets  | 0,2 | 0,3 | 1,3 | 1,4 | 2,4 | 2,5 | 3,5|
 ///
 /// Example 2: `hello` would be tokenized as (min_gram: 2, max_gram: 5, prefix_only: **true**)
@@ -19,24 +20,63 @@ use super::{Token, TokenStream, Tokenizer};
 /// | Position | 0   | 0   | 0     | 0     |
 /// | Offsets  | 0,2 | 0,3 | 0,4   | 0,5   |
 ///
+/// Example 3: `hεllo` (non-ascii) would be tokenized as (min_gram: 2, max_gram: 5, prefix_only: **true**)
+///
+/// | Term     | hε  | hεl | hεll  | hεllo |
+/// |----------|-----|-----|-------|-------|
+/// | Position | 0   | 0   | 0     | 0     |
+/// | Offsets  | 0,3 | 0,4 | 0,5   | 0,6   |
+///
 /// # Example
 ///
 /// ```
-/// extern crate tantivy;
+/// # extern crate tantivy;
 /// use tantivy::tokenizer::*;
-/// use tantivy::tokenizer::assert_token;
-///
 /// # fn main() {
 /// let tokenizer = NgramTokenizer::new(2, 3, false);
 /// let mut stream = tokenizer.token_stream("hello");
-///
-/// assert_token(stream.next().unwrap(), 0, "he", 0, 2);
-/// assert_token(stream.next().unwrap(), 0, "hel", 0, 3);
-/// assert_token(stream.next().unwrap(), 1, "el", 1, 3);
-/// assert_token(stream.next().unwrap(), 1, "ell", 1, 4);
-/// assert_token(stream.next().unwrap(), 2, "ll", 2, 4);
-/// assert_token(stream.next().unwrap(), 2, "llo", 2, 5);
-/// assert_token(stream.next().unwrap(), 3, "lo", 3, 5);
+/// {
+///     let token = stream.next().unwrap();
+///     assert_eq!(token.text, "he");
+///     assert_eq!(token.offset_from, 0);
+///     assert_eq!(token.offset_to, 2);
+/// }
+/// {
+///   let token = stream.next().unwrap();
+///     assert_eq!(token.text, "hel");
+///     assert_eq!(token.offset_from, 0);
+///     assert_eq!(token.offset_to, 3);
+/// }
+/// {
+///   let token = stream.next().unwrap();
+///     assert_eq!(token.text, "el");
+///     assert_eq!(token.offset_from, 1);
+///     assert_eq!(token.offset_to, 3);
+/// }
+/// {
+///   let token = stream.next().unwrap();
+///     assert_eq!(token.text, "ell");
+///     assert_eq!(token.offset_from, 1);
+///     assert_eq!(token.offset_to, 4);
+/// }
+/// {
+///   let token = stream.next().unwrap();
+///     assert_eq!(token.text, "ll");
+///     assert_eq!(token.offset_from, 2);
+///     assert_eq!(token.offset_to, 4);
+/// }
+/// {
+///   let token = stream.next().unwrap();
+///     assert_eq!(token.text, "llo");
+///     assert_eq!(token.offset_from, 2);
+///     assert_eq!(token.offset_to, 5);
+/// }
+/// {
+///   let token = stream.next().unwrap();
+///   assert_eq!(token.text, "lo");
+///   assert_eq!(token.offset_from, 3);
+///   assert_eq!(token.offset_to, 5);
+/// }
 /// assert!(stream.next().is_none());
 /// # }
 /// ```
@@ -58,23 +98,37 @@ impl NgramTokenizer {
             min_gram <= max_gram,
             "min_gram must not be greater than max_gram"
         );
-
         NgramTokenizer {
             min_gram,
             max_gram,
             prefix_only,
         }
     }
+
+    /// Create a `NGramTokenizer` which generates tokens for all inner ngrams.
+    ///
+    /// This is as opposed to only prefix ngrams    .
+    pub fn all_ngrams(min_gram: usize, max_gram: usize) -> NgramTokenizer {
+        Self::new(min_gram, max_gram, false)
+    }
+
+    /// Create a `NGramTokenizer` which only generates tokens for the
+    /// prefix ngrams.
+    pub fn prefix_only(min_gram: usize, max_gram: usize) -> NgramTokenizer {
+        Self::new(min_gram, max_gram, true)
+    }
 }
+
+/// TokenStream associate to the `NgramTokenizer`
 pub struct NgramTokenStream<'a> {
-    text: &'a str,
-    position: usize,
-    text_length: usize,
-    token: Token,
-    min_gram: usize,
-    max_gram: usize,
-    gram_size: usize,
+    /// parameters
+    ngram_charidx_iterator: StutteringIterator<CodepointFrontiers<'a>>,
+    /// true if the NgramTokenStream is in prefix mode.
     prefix_only: bool,
+    /// input
+    text: &'a str,
+    /// output
+    token: Token,
 }
 
 impl<'a> Tokenizer<'a> for NgramTokenizer {
@@ -82,65 +136,29 @@ impl<'a> Tokenizer<'a> for NgramTokenizer {
 
     fn token_stream(&self, text: &'a str) -> Self::TokenStreamImpl {
         NgramTokenStream {
-            text,
-            position: 0,
-            text_length: text.len(),
-            token: Token::default(),
-            min_gram: self.min_gram,
-            max_gram: self.max_gram,
+            ngram_charidx_iterator: StutteringIterator::new(
+                CodepointFrontiers::for_str(text),
+                self.min_gram,
+                self.max_gram,
+            ),
             prefix_only: self.prefix_only,
-            gram_size: self.min_gram,
+            text,
+            token: Token::default(),
         }
     }
 }
 
-impl<'a> NgramTokenStream<'a> {
-    /// Get the next set of token options
-    /// cycle through 1,2 (min..=max)
-    /// returning None if processing should stop
-    fn chomp(&mut self) -> Option<(usize, usize)> {
-        // Have we exceeded the bounds of the text we are indexing?
-        if self.gram_size > self.max_gram {
-            if self.prefix_only {
-                return None;
-            }
-
-            // since we aren't just processing edges
-            // we need to reset the gram size
-            self.gram_size = self.min_gram;
-
-            // and move down the chain of letters
-            self.position += 1;
-        }
-
-        let result = if (self.position + self.gram_size) <= self.text_length {
-            Some((self.position, self.gram_size))
-        } else {
-            None
-        };
-
-        // increase the gram size for the next pass
-        self.gram_size += 1;
-
-        result
-    }
-}
-
 impl<'a> TokenStream for NgramTokenStream<'a> {
     fn advance(&mut self) -> bool {
-        // clear out working token text
-        self.token.text.clear();
-
-        if let Some((position, size)) = self.chomp() {
-            self.token.position = position;
-            let offset_from = position;
-            let offset_to = offset_from + size;
-
+        if let Some((offset_from, offset_to)) = self.ngram_charidx_iterator.next() {
+            if self.prefix_only && offset_from > 0 {
+                return false;
+            }
+            self.token.position = 0;
             self.token.offset_from = offset_from;
             self.token.offset_to = offset_to;
-
+            self.token.text.clear();
             self.token.text.push_str(&self.text[offset_from..offset_to]);
-
             true
         } else {
             false
@@ -150,8 +168,298 @@ impl<'a> TokenStream for NgramTokenStream<'a> {
     fn token(&self) -> &Token {
         &self.token
     }
-
     fn token_mut(&mut self) -> &mut Token {
         &mut self.token
     }
 }
+
+/// This iterator takes an underlying Iterator
+/// and emits all of the pairs `(a,b)` such that
+/// a and b are items emitted by the iterator at
+/// an interval between `min_gram` and `max_gram`.
+///
+/// The elements are emitted in the order of appearance
+/// of `a` first, `b` then.
+///
+/// See `test_stutterring_iterator` for an example of its
+/// output.
+struct StutteringIterator<T> {
+    underlying: T,
+    min_gram: usize,
+    max_gram: usize,
+
+    memory: Vec<usize>,
+    cursor: usize,
+    gram_len: usize,
+}
+
+impl<T> StutteringIterator<T>
+where
+    T: Iterator<Item = usize>,
+{
+    pub fn new(mut underlying: T, min_gram: usize, max_gram: usize) -> StutteringIterator<T> {
+        assert!(min_gram > 0);
+        let memory: Vec<usize> = (&mut underlying).take(max_gram + 1).collect();
+        if memory.len() <= min_gram {
+            // returns an empty iterator
+            StutteringIterator {
+                underlying,
+                min_gram: 1,
+                max_gram: 0,
+                memory,
+                cursor: 0,
+                gram_len: 0,
+            }
+        } else {
+            StutteringIterator {
+                underlying,
+                min_gram,
+                max_gram: memory.len() - 1,
+                memory,
+                cursor: 0,
+                gram_len: min_gram,
+            }
+        }
+    }
+}
+
+impl<T> Iterator for StutteringIterator<T>
+where
+    T: Iterator<Item = usize>,
+{
+    type Item = (usize, usize);
+
+    fn next(&mut self) -> Option<(usize, usize)> {
+        if self.gram_len > self.max_gram {
+            // we have exhausted all options
+            // starting at `self.memory[self.cursor]`.
+            //
+            // Time to advance.
+            self.gram_len = self.min_gram;
+            if let Some(next_val) = self.underlying.next() {
+                self.memory[self.cursor] = next_val;
+            } else {
+                self.max_gram -= 1;
+            }
+            self.cursor += 1;
+            if self.cursor >= self.memory.len() {
+                self.cursor = 0;
+            }
+        }
+        if self.max_gram < self.min_gram {
+            return None;
+        }
+        let start = self.memory[self.cursor % self.memory.len()];
+        let stop = self.memory[(self.cursor + self.gram_len) % self.memory.len()];
+        self.gram_len += 1;
+        Some((start, stop))
+    }
+}
+
+/// Emits all of the offsets where a codepoint starts
+/// or a codepoint ends.
+///
+/// By convention, we emit [0] for the empty string.
+struct CodepointFrontiers<'a> {
+    s: &'a str,
+    next_el: Option<usize>,
+}
+
+impl<'a> CodepointFrontiers<'a> {
+    fn for_str(s: &'a str) -> Self {
+        CodepointFrontiers {
+            s,
+            next_el: Some(0),
+        }
+    }
+}
+
+impl<'a> Iterator for CodepointFrontiers<'a> {
+    type Item = usize;
+
+    fn next(&mut self) -> Option<usize> {
+        self.next_el.map(|offset| {
+            if self.s.is_empty() {
+                self.next_el = None;
+            } else {
+                let first_codepoint_width = utf8_codepoint_width(self.s.as_bytes()[0]);
+                self.s = &self.s[first_codepoint_width..];
+                self.next_el = Some(offset + first_codepoint_width);
+            }
+            offset
+        })
+    }
+}
+
+const CODEPOINT_UTF8_WIDTH: [u8; 16] = [1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 3, 4];
+
+// Number of bytes to encode a codepoint in UTF-8 given
+// the first byte.
+//
+// To do that we count the number of higher significant bits set to `1`.
+fn utf8_codepoint_width(b: u8) -> usize {
+    let higher_4_bits = (b as usize) >> 4;
+    CODEPOINT_UTF8_WIDTH[higher_4_bits] as usize
+}
+
+#[cfg(test)]
+mod tests {
+
+    use super::utf8_codepoint_width;
+    use super::CodepointFrontiers;
+    use super::NgramTokenizer;
+    use super::StutteringIterator;
+    use tokenizer::tests::assert_token;
+    use tokenizer::tokenizer::{TokenStream, Tokenizer};
+    use tokenizer::Token;
+
+    fn test_helper<T: TokenStream>(mut tokenizer: T) -> Vec<Token> {
+        let mut tokens: Vec<Token> = vec![];
+        tokenizer.process(&mut |token: &Token| tokens.push(token.clone()));
+        tokens
+    }
+
+    #[test]
+    fn test_utf8_codepoint_width() {
+        // 0xxx
+        for i in 0..128 {
+            assert_eq!(utf8_codepoint_width(i), 1);
+        }
+        // 110xx
+        for i in (128 | 64)..(128 | 64 | 32) {
+            assert_eq!(utf8_codepoint_width(i), 2);
+        }
+        // 1110xx
+        for i in (128 | 64 | 32)..(128 | 64 | 32 | 16) {
+            assert_eq!(utf8_codepoint_width(i), 3);
+        }
+        // 1111xx
+        for i in (128 | 64 | 32 | 16)..256 {
+            assert_eq!(utf8_codepoint_width(i as u8), 4);
+        }
+    }
+
+    #[test]
+    fn test_codepoint_frontiers() {
+        assert_eq!(CodepointFrontiers::for_str("").collect::<Vec<_>>(), vec![0]);
+        assert_eq!(
+            CodepointFrontiers::for_str("abcd").collect::<Vec<_>>(),
+            vec![0, 1, 2, 3, 4]
+        );
+        assert_eq!(
+            CodepointFrontiers::for_str("aあ").collect::<Vec<_>>(),
+            vec![0, 1, 4]
+        );
+    }
+
+    #[test]
+    fn test_ngram_tokenizer_1_2_false() {
+        let tokens = test_helper(NgramTokenizer::all_ngrams(1, 2).token_stream("hello"));
+        assert_eq!(tokens.len(), 9);
+        assert_token(&tokens[0], 0, "h", 0, 1);
+        assert_token(&tokens[1], 0, "he", 0, 2);
+        assert_token(&tokens[2], 0, "e", 1, 2);
+        assert_token(&tokens[3], 0, "el", 1, 3);
+        assert_token(&tokens[4], 0, "l", 2, 3);
+        assert_token(&tokens[5], 0, "ll", 2, 4);
+        assert_token(&tokens[6], 0, "l", 3, 4);
+        assert_token(&tokens[7], 0, "lo", 3, 5);
+        assert_token(&tokens[8], 0, "o", 4, 5);
+    }
+
+    #[test]
+    fn test_ngram_tokenizer_min_max_equal() {
+        let tokens = test_helper(NgramTokenizer::all_ngrams(3, 3).token_stream("hello"));
+        assert_eq!(tokens.len(), 3);
+        assert_token(&tokens[0], 0, "hel", 0, 3);
+        assert_token(&tokens[1], 0, "ell", 1, 4);
+        assert_token(&tokens[2], 0, "llo", 2, 5);
+    }
+
+    #[test]
+    fn test_ngram_tokenizer_2_5_prefix() {
+        let tokens = test_helper(NgramTokenizer::prefix_only(2, 5).token_stream("frankenstein"));
+        assert_eq!(tokens.len(), 4);
+        assert_token(&tokens[0], 0, "fr", 0, 2);
+        assert_token(&tokens[1], 0, "fra", 0, 3);
+        assert_token(&tokens[2], 0, "fran", 0, 4);
+        assert_token(&tokens[3], 0, "frank", 0, 5);
+    }
+
+    #[test]
+    fn test_ngram_non_ascii_1_2() {
+        let tokens = test_helper(NgramTokenizer::all_ngrams(1, 2).token_stream("hεllo"));
+        assert_eq!(tokens.len(), 9);
+        assert_token(&tokens[0], 0, "h", 0, 1);
+        assert_token(&tokens[1], 0, "hε", 0, 3);
+        assert_token(&tokens[2], 0, "ε", 1, 3);
+        assert_token(&tokens[3], 0, "εl", 1, 4);
+        assert_token(&tokens[4], 0, "l", 3, 4);
+        assert_token(&tokens[5], 0, "ll", 3, 5);
+        assert_token(&tokens[6], 0, "l", 4, 5);
+        assert_token(&tokens[7], 0, "lo", 4, 6);
+        assert_token(&tokens[8], 0, "o", 5, 6);
+    }
+
+    #[test]
+    fn test_ngram_non_ascii_2_5_prefix() {
+        let tokens = test_helper(NgramTokenizer::prefix_only(2, 5).token_stream("hεllo"));
+        assert_eq!(tokens.len(), 4);
+        assert_token(&tokens[0], 0, "hε", 0, 3);
+        assert_token(&tokens[1], 0, "hεl", 0, 4);
+        assert_token(&tokens[2], 0, "hεll", 0, 5);
+        assert_token(&tokens[3], 0, "hεllo", 0, 6);
+    }
+
+    #[test]
+    fn test_ngram_empty() {
+        let tokens = test_helper(NgramTokenizer::all_ngrams(1, 5).token_stream(""));
+        assert!(tokens.is_empty());
+        let tokens = test_helper(NgramTokenizer::all_ngrams(2, 5).token_stream(""));
+        assert!(tokens.is_empty());
+    }
+
+    #[test]
+    #[should_panic(expected = "min_gram must be greater than 0")]
+    fn test_ngram_min_max_interval_empty() {
+        test_helper(NgramTokenizer::all_ngrams(0, 2).token_stream("hellossss"));
+    }
+
+    #[test]
+    #[should_panic(expected = "min_gram must not be greater than max_gram")]
+    fn test_invalid_interval_should_panic_if_smaller() {
+        NgramTokenizer::all_ngrams(2, 1);
+    }
+
+    #[test]
+    fn test_stutterring_iterator_empty() {
+        let rg: Vec<usize> = vec![0];
+        let mut it = StutteringIterator::new(rg.into_iter(), 1, 2);
+        assert_eq!(it.next(), None);
+    }
+
+    #[test]
+    fn test_stutterring_iterator() {
+        let rg: Vec<usize> = (0..10).collect();
+        let mut it = StutteringIterator::new(rg.into_iter(), 1, 2);
+        assert_eq!(it.next(), Some((0, 1)));
+        assert_eq!(it.next(), Some((0, 2)));
+        assert_eq!(it.next(), Some((1, 2)));
+        assert_eq!(it.next(), Some((1, 3)));
+        assert_eq!(it.next(), Some((2, 3)));
+        assert_eq!(it.next(), Some((2, 4)));
+        assert_eq!(it.next(), Some((3, 4)));
+        assert_eq!(it.next(), Some((3, 5)));
+        assert_eq!(it.next(), Some((4, 5)));
+        assert_eq!(it.next(), Some((4, 6)));
+        assert_eq!(it.next(), Some((5, 6)));
+        assert_eq!(it.next(), Some((5, 7)));
+        assert_eq!(it.next(), Some((6, 7)));
+        assert_eq!(it.next(), Some((6, 8)));
+        assert_eq!(it.next(), Some((7, 8)));
+        assert_eq!(it.next(), Some((7, 9)));
+        assert_eq!(it.next(), Some((8, 9)));
+        assert_eq!(it.next(), None);
+    }
+
+}