Merge branch 'query-boost' of https://github.com/audunhalland/tantivy into audunhalland-query-boost

* Make Executor public so Searcher::search_in_executor method now can be used

* Fixed cargo fmt

CHANGELOG.md

@@ -1,3 +1,30 @@
+Tantivy 0.12.0
+======================
+- Removing static dispatch in tokenizers for simplicity. (#762)
+- Added backward iteration for `TermDictionary` stream. (@halvorboe)
+- Fixed a performance issue when searching for the posting lists of a missing term (@audunhalland)
+- Added a configurable maximum number of docs (10M by default) for a segment to be considered for merge (@hntd187, landed by @halvorboe #713) 
+
+## How to update?
+
+Crates relying on custom tokenizer, or registering tokenizer in the manager will require some 
+minor changes. Check https://github.com/tantivy-search/tantivy/blob/master/examples/custom_tokenizer.rs
+to check for some code sample.
+
+Tantivy 0.11.3
+=======================
+- Fixed DateTime as a fast field (#735)
+
+Tantivy 0.11.2
+=======================
+- The future returned by `IndexWriter::merge` does not borrow `self` mutably anymore (#732)
+- Exposing a constructor for `WatchHandle` (#731)
+
+Tantivy 0.11.1
+=====================
+- Bug fix #729
+
+
 Tantivy 0.11.0
 =====================
 

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "tantivy"
-version = "0.11.0"
+version = "0.12.0"
 authors = ["Paul Masurel <paul.masurel@gmail.com>"]
 license = "MIT"
 categories = ["database-implementations", "data-structures"]
@@ -18,7 +18,7 @@ byteorder = "1.0"
 crc32fast = "1.2.0"
 once_cell = "1.0"
 regex ={version = "1.3.0", default-features = false, features = ["std"]}
-tantivy-fst = "0.1"
+tantivy-fst = "0.2.1"
 memmap = {version = "0.7", optional=true}
 lz4 = {version="1.20", optional=true}
 snap = {version="0.2"}
@@ -60,7 +60,6 @@ winapi = "0.3"
 rand = "0.7"
 maplit = "1"
 matches = "0.1.8"
-time = "0.1.42"
 
 [dev-dependencies.fail]
 version = "0.3"

examples/faceted_search.rs

@@ -13,63 +13,100 @@
 // ---
 // Importing tantivy...
 use tantivy::collector::FacetCollector;
-use tantivy::query::AllQuery;
+use tantivy::query::{AllQuery, TermQuery};
 use tantivy::schema::*;
 use tantivy::{doc, Index};
-use tempfile::TempDir;
 
 fn main() -> tantivy::Result<()> {
-    // Let's create a temporary directory for the
-    // sake of this example
-    let index_path = TempDir::new()?;
+    // Let's create a temporary directory for the sake of this example
     let mut schema_builder = Schema::builder();
 
-    schema_builder.add_text_field("name", TEXT | STORED);
-
-    // this is our faceted field
-    schema_builder.add_facet_field("tags");
+    let name = schema_builder.add_text_field("felin_name", TEXT | STORED);
+    // this is our faceted field: its scientific classification
+    let classification = schema_builder.add_facet_field("classification");
 
     let schema = schema_builder.build();
+    let index = Index::create_in_ram(schema);
 
-    let index = Index::create_in_dir(&index_path, schema.clone())?;
-
-    let mut index_writer = index.writer(50_000_000)?;
-
-    let name = schema.get_field("name").unwrap();
-    let tags = schema.get_field("tags").unwrap();
+    let mut index_writer = index.writer(30_000_000)?;
 
     // For convenience, tantivy also comes with a macro to
     // reduce the boilerplate above.
     index_writer.add_document(doc!(
-        name => "the ditch",
-        tags => Facet::from("/pools/north")
+        name => "Cat",
+        classification => Facet::from("/Felidae/Felinae/Felis")
     ));
-
     index_writer.add_document(doc!(
-        name => "little stacey",
-        tags => Facet::from("/pools/south")
+        name => "Canada lynx",
+        classification => Facet::from("/Felidae/Felinae/Lynx")
+    ));
+    index_writer.add_document(doc!(
+        name => "Cheetah",
+        classification => Facet::from("/Felidae/Felinae/Acinonyx")
+    ));
+    index_writer.add_document(doc!(
+        name => "Tiger",
+        classification => Facet::from("/Felidae/Pantherinae/Panthera")
+    ));
+    index_writer.add_document(doc!(
+        name => "Lion",
+        classification => Facet::from("/Felidae/Pantherinae/Panthera")
+    ));
+    index_writer.add_document(doc!(
+        name => "Jaguar",
+        classification => Facet::from("/Felidae/Pantherinae/Panthera")
+    ));
+    index_writer.add_document(doc!(
+        name => "Sunda clouded leopard",
+        classification => Facet::from("/Felidae/Pantherinae/Neofelis")
+    ));
+    index_writer.add_document(doc!(
+        name => "Fossa",
+        classification => Facet::from("/Eupleridae/Cryptoprocta")
     ));
-
     index_writer.commit()?;
 
     let reader = index.reader()?;
-
     let searcher = reader.searcher();
+    {
+        let mut facet_collector = FacetCollector::for_field(classification);
+        facet_collector.add_facet("/Felidae");
+        let facet_counts = searcher.search(&AllQuery, &facet_collector)?;
+        // This lists all of the facet counts, right below "/Felidae".
+        let facets: Vec<(&Facet, u64)> = facet_counts.get("/Felidae").collect();
+        assert_eq!(
+            facets,
+            vec![
+                (&Facet::from("/Felidae/Felinae"), 3),
+                (&Facet::from("/Felidae/Pantherinae"), 4),
+            ]
+        );
+    }
 
-    let mut facet_collector = FacetCollector::for_field(tags);
-    facet_collector.add_facet("/pools");
+    // Facets are also searchable.
+    //
+    // For instance a common UI pattern is to allow the user someone to click on a facet link
+    // (e.g: `Pantherinae`) to drill down and filter the current result set with this subfacet.
+    //
+    // The search would then look as follows.
 
-    let facet_counts = searcher.search(&AllQuery, &facet_collector).unwrap();
-
-    // This lists all of the facet counts
-    let facets: Vec<(&Facet, u64)> = facet_counts.get("/pools").collect();
-    assert_eq!(
-        facets,
-        vec![
-            (&Facet::from("/pools/north"), 1),
-            (&Facet::from("/pools/south"), 1),
-        ]
-    );
+    // Check the reference doc for different ways to create a `Facet` object.
+    {
+        let facet = Facet::from_text("/Felidae/Pantherinae");
+        let facet_term = Term::from_facet(classification, &facet);
+        let facet_term_query = TermQuery::new(facet_term, IndexRecordOption::Basic);
+        let mut facet_collector = FacetCollector::for_field(classification);
+        facet_collector.add_facet("/Felidae/Pantherinae");
+        let facet_counts = searcher.search(&facet_term_query, &facet_collector)?;
+        let facets: Vec<(&Facet, u64)> = facet_counts.get("/Felidae/Pantherinae").collect();
+        assert_eq!(
+            facets,
+            vec![
+                (&Facet::from("/Felidae/Pantherinae/Neofelis"), 1),
+                (&Facet::from("/Felidae/Pantherinae/Panthera"), 3),
+            ]
+        );
+    }
 
     Ok(())
 }

examples/pre_tokenized_text.rs

@@ -9,7 +9,7 @@
 // - import tokenized text straight from json,
 // - perform a search on documents with pre-tokenized text
 
-use tantivy::tokenizer::{PreTokenizedString, SimpleTokenizer, Token, TokenStream, Tokenizer};
+use tantivy::tokenizer::{PreTokenizedString, SimpleTokenizer, Token, Tokenizer};
 
 use tantivy::collector::{Count, TopDocs};
 use tantivy::query::TermQuery;

examples/stop_words.rs

@@ -50,7 +50,7 @@ fn main() -> tantivy::Result<()> {
 
     // This tokenizer lowers all of the text (to help with stop word matching)
     // then removes all instances of `the` and `and` from the corpus
-    let tokenizer = SimpleTokenizer
+    let tokenizer = TextAnalyzer::from(SimpleTokenizer)
         .filter(LowerCaser)
         .filter(StopWordFilter::remove(vec![
             "the".to_string(),

src/collector/facet_collector.rs

@@ -452,9 +452,11 @@ impl FacetCounts {
 #[cfg(test)]
 mod tests {
     use super::{FacetCollector, FacetCounts};
+    use crate::collector::Count;
     use crate::core::Index;
-    use crate::query::AllQuery;
-    use crate::schema::{Document, Facet, Field, Schema};
+    use crate::query::{AllQuery, QueryParser, TermQuery};
+    use crate::schema::{Document, Facet, Field, IndexRecordOption, Schema};
+    use crate::Term;
     use rand::distributions::Uniform;
     use rand::prelude::SliceRandom;
     use rand::{thread_rng, Rng};
@@ -544,6 +546,56 @@ mod tests {
         assert_eq!(facets[0].1, 1);
     }
 
+    #[test]
+    fn test_doc_search_by_facet() {
+        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);
+        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        index_writer.add_document(doc!(
+            facet_field => Facet::from_text(&"/A/A"),
+        ));
+        index_writer.add_document(doc!(
+            facet_field => Facet::from_text(&"/A/B"),
+        ));
+        index_writer.add_document(doc!(
+            facet_field => Facet::from_text(&"/A/C/A"),
+        ));
+        index_writer.add_document(doc!(
+            facet_field => Facet::from_text(&"/D/C/A"),
+        ));
+        index_writer.commit().unwrap();
+        let reader = index.reader().unwrap();
+        let searcher = reader.searcher();
+        assert_eq!(searcher.num_docs(), 4);
+
+        let count_facet = |facet_str: &str| {
+            let term = Term::from_facet(facet_field, &Facet::from_text(facet_str));
+            searcher
+                .search(&TermQuery::new(term, IndexRecordOption::Basic), &Count)
+                .unwrap()
+        };
+
+        assert_eq!(count_facet("/"), 4);
+        assert_eq!(count_facet("/A"), 3);
+        assert_eq!(count_facet("/A/B"), 1);
+        assert_eq!(count_facet("/A/C"), 1);
+        assert_eq!(count_facet("/A/C/A"), 1);
+        assert_eq!(count_facet("/C/A"), 0);
+        {
+            let query_parser = QueryParser::for_index(&index, vec![]);
+            {
+                let query = query_parser.parse_query("facet:/A/B").unwrap();
+                assert_eq!(1, searcher.search(&query, &Count).unwrap());
+            }
+            {
+                let query = query_parser.parse_query("facet:/A").unwrap();
+                assert_eq!(3, searcher.search(&query, &Count).unwrap());
+            }
+        }
+    }
+
     #[test]
     fn test_non_used_facet_collector() {
         let mut facet_collector = FacetCollector::for_field(Field::from_field_id(0));

src/collector/top_score_collector.rs

@@ -6,6 +6,7 @@ use crate::collector::tweak_score_top_collector::TweakedScoreTopCollector;
 use crate::collector::{
     CustomScorer, CustomSegmentScorer, ScoreSegmentTweaker, ScoreTweaker, SegmentCollector,
 };
+use crate::fastfield::FastFieldReader;
 use crate::schema::Field;
 use crate::DocAddress;
 use crate::DocId;
@@ -61,6 +62,34 @@ impl fmt::Debug for TopDocs {
     }
 }
 
+struct ScorerByFastFieldReader {
+    ff_reader: FastFieldReader<u64>,
+}
+
+impl CustomSegmentScorer<u64> for ScorerByFastFieldReader {
+    fn score(&self, doc: DocId) -> u64 {
+        self.ff_reader.get_u64(u64::from(doc))
+    }
+}
+
+struct ScorerByField {
+    field: Field,
+}
+
+impl CustomScorer<u64> for ScorerByField {
+    type Child = ScorerByFastFieldReader;
+
+    fn segment_scorer(&self, segment_reader: &SegmentReader) -> crate::Result<Self::Child> {
+        let ff_reader = segment_reader
+            .fast_fields()
+            .u64(self.field)
+            .ok_or_else(|| {
+                crate::Error::SchemaError(format!("Field requested is not a i64/u64 fast field."))
+            })?;
+        Ok(ScorerByFastFieldReader { ff_reader })
+    }
+}
+
 impl TopDocs {
     /// Creates a top score collector, with a number of documents equal to "limit".
     ///
@@ -143,14 +172,7 @@ impl TopDocs {
         self,
         field: Field,
     ) -> impl Collector<Fruit = Vec<(u64, DocAddress)>> {
-        self.custom_score(move |segment_reader: &SegmentReader| {
-            let ff_reader = segment_reader
-                .fast_fields()
-                .u64(field)
-                .expect("Field requested is not a i64/u64 fast field.");
-            //TODO error message missmatch actual behavior for i64
-            move |doc: DocId| ff_reader.get(doc)
-        })
+        self.custom_score(ScorerByField { field })
     }
 
     /// Ranks the documents using a custom score.
@@ -572,7 +594,6 @@ mod tests {
     }
 
     #[test]
-    #[should_panic(expected = "Field requested is not a i64/u64 fast field")]
     fn test_field_not_fast_field() {
         let mut schema_builder = Schema::builder();
         let title = schema_builder.add_text_field(TITLE, TEXT);
@@ -587,7 +608,12 @@ mod tests {
         let searcher = index.reader().unwrap().searcher();
         let segment = searcher.segment_reader(0);
         let top_collector = TopDocs::with_limit(4).order_by_u64_field(size);
-        assert!(top_collector.for_segment(0, segment).is_ok());
+        let err = top_collector.for_segment(0, segment);
+        if let Err(crate::Error::SchemaError(msg)) = err {
+            assert_eq!(msg, "Field requested is not a i64/u64 fast field.");
+        } else {
+            assert!(false);
+        }
     }
 
     fn index(

src/core/executor.rs

@@ -10,7 +10,9 @@ use rayon::{ThreadPool, ThreadPoolBuilder};
 /// 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 {
+    /// Single thread variant of an Executor
     SingleThread,
+    /// Thread pool variant of an Executor
     ThreadPool(ThreadPool),
 }
 
@@ -20,7 +22,7 @@ impl Executor {
         Executor::SingleThread
     }
 
-    // Creates an Executor that dispatches the tasks in a thread pool.
+    /// Creates an Executor that dispatches the tasks in a thread pool.
     pub fn multi_thread(num_threads: usize, prefix: &'static str) -> Result<Executor> {
         let pool = ThreadPoolBuilder::new()
             .num_threads(num_threads)
@@ -29,10 +31,10 @@ impl Executor {
         Ok(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.
+    /// 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,

src/core/index.rs

@@ -20,8 +20,7 @@ use crate::reader::IndexReaderBuilder;
 use crate::schema::Field;
 use crate::schema::FieldType;
 use crate::schema::Schema;
-use crate::tokenizer::BoxedTokenizer;
-use crate::tokenizer::TokenizerManager;
+use crate::tokenizer::{TextAnalyzer, TokenizerManager};
 use crate::IndexWriter;
 use crate::Result;
 use num_cpus;
@@ -173,11 +172,11 @@ impl Index {
     }
 
     /// Helper to access the tokenizer associated to a specific field.
-    pub fn tokenizer_for_field(&self, field: Field) -> Result<BoxedTokenizer> {
+    pub fn tokenizer_for_field(&self, field: Field) -> Result<TextAnalyzer> {
         let field_entry = self.schema.get_field_entry(field);
         let field_type = field_entry.field_type();
         let tokenizer_manager: &TokenizerManager = self.tokenizers();
-        let tokenizer_name_opt: Option<BoxedTokenizer> = match field_type {
+        let tokenizer_name_opt: Option<TextAnalyzer> = match field_type {
             FieldType::Str(text_options) => text_options
                 .get_indexing_options()
                 .map(|text_indexing_options| text_indexing_options.tokenizer().to_string())

src/core/inverted_index_reader.rs

@@ -60,7 +60,7 @@ impl InvertedIndexReader {
             .get_index_record_option()
             .unwrap_or(IndexRecordOption::Basic);
         InvertedIndexReader {
-            termdict: TermDictionary::empty(&field_type),
+            termdict: TermDictionary::empty(),
             postings_source: ReadOnlySource::empty(),
             positions_source: ReadOnlySource::empty(),
             positions_idx_source: ReadOnlySource::empty(),

src/directory/directory.rs

@@ -119,7 +119,7 @@ pub trait Directory: DirectoryClone + fmt::Debug + Send + Sync + 'static {
     /// Specifically, subsequent writes or flushes should
     /// have no effect on the returned `ReadOnlySource` object.
     ///
-    /// You should only use this to read files create with [`open_write`]
+    /// You should only use this to read files create with [Directory::open_write].
     fn open_read(&self, path: &Path) -> result::Result<ReadOnlySource, OpenReadError>;
 
     /// Removes a file
@@ -160,7 +160,7 @@ pub trait Directory: DirectoryClone + fmt::Debug + Send + Sync + 'static {
     ///
     /// This should only be used for small files.
     ///
-    /// You should only use this to read files create with [`atomic_write`]
+    /// You should only use this to read files create with [Directory::atomic_write].
     fn atomic_read(&self, path: &Path) -> Result<Vec<u8>, OpenReadError>;
 
     /// Atomically replace the content of a file with data.
@@ -197,7 +197,7 @@ pub trait Directory: DirectoryClone + fmt::Debug + Send + Sync + 'static {
     /// Registers a callback that will be called whenever a change on the `meta.json`
     /// using the `atomic_write` API is detected.
     ///
-    /// The behavior when using `.watch()` on a file using `.open_write(...)` is, on the other
+    /// The behavior when using `.watch()` on a file using [Directory::open_write] is, on the other
     /// hand, undefined.
     ///
     /// The file will be watched for the lifetime of the returned `WatchHandle`. The caller is

src/directory/mmap_directory.rs

@@ -131,14 +131,13 @@ impl MmapCache {
         }
         self.cache.remove(full_path);
         self.counters.miss += 1;
-        Ok(if let Some(mmap) = open_mmap(full_path)? {
+        let mmap_opt = open_mmap(full_path)?;
+        Ok(mmap_opt.map(|mmap| {
             let mmap_arc: Arc<BoxedData> = Arc::new(Box::new(mmap));
             let mmap_weak = Arc::downgrade(&mmap_arc);
             self.cache.insert(full_path.to_owned(), mmap_weak);
-            Some(mmap_arc)
-        } else {
-            None
-        })
+            mmap_arc
+        }))
     }
 }
 

src/directory/mod.rs

@@ -23,11 +23,9 @@ pub use self::directory::{Directory, DirectoryClone};
 pub use self::directory_lock::{Lock, INDEX_WRITER_LOCK, META_LOCK};
 pub use self::ram_directory::RAMDirectory;
 pub use self::read_only_source::ReadOnlySource;
-pub(crate) use self::watch_event_router::WatchCallbackList;
-pub use self::watch_event_router::{WatchCallback, WatchHandle};
+pub use self::watch_event_router::{WatchCallback, WatchCallbackList, WatchHandle};
 use std::io::{self, BufWriter, Write};
 use std::path::PathBuf;
-
 /// Outcome of the Garbage collection
 pub struct GarbageCollectionResult {
     /// List of files that were deleted in this cycle

src/directory/watch_event_router.rs

@@ -24,13 +24,20 @@ pub struct WatchCallbackList {
 #[derive(Clone)]
 pub struct WatchHandle(Arc<WatchCallback>);
 
+impl WatchHandle {
+    /// Create a WatchHandle handle.
+    pub fn new(watch_callback: Arc<WatchCallback>) -> WatchHandle {
+        WatchHandle(watch_callback)
+    }
+}
+
 impl WatchCallbackList {
     /// Suscribes a new callback and returns a handle that controls the lifetime of the callback.
     pub fn subscribe(&self, watch_callback: WatchCallback) -> WatchHandle {
         let watch_callback_arc = Arc::new(watch_callback);
         let watch_callback_weak = Arc::downgrade(&watch_callback_arc);
         self.router.write().unwrap().push(watch_callback_weak);
-        WatchHandle(watch_callback_arc)
+        WatchHandle::new(watch_callback_arc)
     }
 
     fn list_callback(&self) -> Vec<Arc<WatchCallback>> {

src/fastfield/mod.rs

@@ -33,6 +33,7 @@ pub use self::reader::FastFieldReader;
 pub use self::readers::FastFieldReaders;
 pub use self::serializer::FastFieldSerializer;
 pub use self::writer::{FastFieldsWriter, IntFastFieldWriter};
+use crate::chrono::{NaiveDateTime, Utc};
 use crate::common;
 use crate::schema::Cardinality;
 use crate::schema::FieldType;
@@ -49,7 +50,7 @@ mod serializer;
 mod writer;
 
 /// Trait for types that are allowed for fast fields: (u64, i64 and f64).
-pub trait FastValue: Default + Clone + Copy + Send + Sync + PartialOrd {
+pub trait FastValue: Clone + Copy + Send + Sync + PartialOrd {
     /// Converts a value from u64
     ///
     /// Internally all fast field values are encoded as u64.
@@ -69,6 +70,12 @@ pub trait FastValue: Default + Clone + Copy + Send + Sync + PartialOrd {
     /// Cast value to `u64`.
     /// The value is just reinterpreted in memory.
     fn as_u64(&self) -> u64;
+
+    /// Build a default value. This default value is never used, so the value does not
+    /// really matter.
+    fn make_zero() -> Self {
+        Self::from_u64(0i64.to_u64())
+    }
 }
 
 impl FastValue for u64 {
@@ -135,11 +142,34 @@ impl FastValue for f64 {
     }
 }
 
+impl FastValue for crate::DateTime {
+    fn from_u64(timestamp_u64: u64) -> Self {
+        let timestamp_i64 = i64::from_u64(timestamp_u64);
+        crate::DateTime::from_utc(NaiveDateTime::from_timestamp(timestamp_i64, 0), Utc)
+    }
+
+    fn to_u64(&self) -> u64 {
+        self.timestamp().to_u64()
+    }
+
+    fn fast_field_cardinality(field_type: &FieldType) -> Option<Cardinality> {
+        match *field_type {
+            FieldType::Date(ref integer_options) => integer_options.get_fastfield_cardinality(),
+            _ => None,
+        }
+    }
+
+    fn as_u64(&self) -> u64 {
+        self.timestamp().as_u64()
+    }
+}
+
 fn value_to_u64(value: &Value) -> u64 {
     match *value {
         Value::U64(ref val) => *val,
         Value::I64(ref val) => common::i64_to_u64(*val),
         Value::F64(ref val) => common::f64_to_u64(*val),
+        Value::Date(ref datetime) => common::i64_to_u64(datetime.timestamp()),
         _ => panic!("Expected a u64/i64/f64 field, got {:?} ", value),
     }
 }
@@ -151,10 +181,12 @@ mod tests {
     use crate::common::CompositeFile;
     use crate::directory::{Directory, RAMDirectory, WritePtr};
     use crate::fastfield::FastFieldReader;
-    use crate::schema::Document;
+    use crate::merge_policy::NoMergePolicy;
     use crate::schema::Field;
     use crate::schema::Schema;
     use crate::schema::FAST;
+    use crate::schema::{Document, IntOptions};
+    use crate::{Index, SegmentId, SegmentReader};
     use once_cell::sync::Lazy;
     use rand::prelude::SliceRandom;
     use rand::rngs::StdRng;
@@ -178,6 +210,12 @@ mod tests {
         assert_eq!(test_fastfield.get(2), 300);
     }
 
+    #[test]
+    pub fn test_fastfield_i64_u64() {
+        let datetime = crate::DateTime::from_utc(NaiveDateTime::from_timestamp(0i64, 0), Utc);
+        assert_eq!(i64::from_u64(datetime.to_u64()), 0i64);
+    }
+
     #[test]
     fn test_intfastfield_small() {
         let path = Path::new("test");
@@ -429,6 +467,93 @@ mod tests {
             }
         }
     }
+
+    #[test]
+    fn test_merge_missing_date_fast_field() {
+        let mut schema_builder = Schema::builder();
+        let date_field = schema_builder.add_date_field("date", 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.set_merge_policy(Box::new(NoMergePolicy));
+        index_writer.add_document(doc!(date_field =>crate::chrono::prelude::Utc::now()));
+        index_writer.commit().unwrap();
+        index_writer.add_document(doc!());
+        index_writer.commit().unwrap();
+        let reader = index.reader().unwrap();
+        let segment_ids: Vec<SegmentId> = reader
+            .searcher()
+            .segment_readers()
+            .iter()
+            .map(SegmentReader::segment_id)
+            .collect();
+        assert_eq!(segment_ids.len(), 2);
+        let merge_future = index_writer.merge(&segment_ids[..]);
+        let merge_res = futures::executor::block_on(merge_future);
+        assert!(merge_res.is_ok());
+        assert!(reader.reload().is_ok());
+        assert_eq!(reader.searcher().segment_readers().len(), 1);
+    }
+
+    #[test]
+    fn test_default_datetime() {
+        assert_eq!(crate::DateTime::make_zero().timestamp(), 0i64);
+    }
+
+    #[test]
+    fn test_datefastfield() {
+        use crate::fastfield::FastValue;
+        let mut schema_builder = Schema::builder();
+        let date_field = schema_builder.add_date_field("date", FAST);
+        let multi_date_field = schema_builder.add_date_field(
+            "multi_date",
+            IntOptions::default().set_fast(Cardinality::MultiValues),
+        );
+        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.set_merge_policy(Box::new(NoMergePolicy));
+        index_writer.add_document(doc!(
+            date_field => crate::DateTime::from_u64(1i64.to_u64()),
+            multi_date_field => crate::DateTime::from_u64(2i64.to_u64()),
+            multi_date_field => crate::DateTime::from_u64(3i64.to_u64())
+        ));
+        index_writer.add_document(doc!(
+            date_field => crate::DateTime::from_u64(4i64.to_u64())
+        ));
+        index_writer.add_document(doc!(
+            multi_date_field => crate::DateTime::from_u64(5i64.to_u64()),
+            multi_date_field => crate::DateTime::from_u64(6i64.to_u64())
+        ));
+        index_writer.commit().unwrap();
+        let reader = index.reader().unwrap();
+        let searcher = reader.searcher();
+        assert_eq!(searcher.segment_readers().len(), 1);
+        let segment_reader = searcher.segment_reader(0);
+        let fast_fields = segment_reader.fast_fields();
+        let date_fast_field = fast_fields.date(date_field).unwrap();
+        let dates_fast_field = fast_fields.dates(multi_date_field).unwrap();
+        let mut dates = vec![];
+        {
+            assert_eq!(date_fast_field.get(0u32).timestamp(), 1i64);
+            dates_fast_field.get_vals(0u32, &mut dates);
+            assert_eq!(dates.len(), 2);
+            assert_eq!(dates[0].timestamp(), 2i64);
+            assert_eq!(dates[1].timestamp(), 3i64);
+        }
+        {
+            assert_eq!(date_fast_field.get(1u32).timestamp(), 4i64);
+            dates_fast_field.get_vals(1u32, &mut dates);
+            assert!(dates.is_empty());
+        }
+        {
+            assert_eq!(date_fast_field.get(2u32).timestamp(), 0i64);
+            dates_fast_field.get_vals(2u32, &mut dates);
+            assert_eq!(dates.len(), 2);
+            assert_eq!(dates[0].timestamp(), 5i64);
+            assert_eq!(dates[1].timestamp(), 6i64);
+        }
+    }
 }
 
 #[cfg(all(test, feature = "unstable"))]

src/fastfield/multivalued/mod.rs

@@ -7,9 +7,6 @@ pub use self::writer::MultiValueIntFastFieldWriter;
 #[cfg(test)]
 mod tests {
 
-    use time;
-
-    use self::time::Duration;
     use crate::collector::TopDocs;
     use crate::query::QueryParser;
     use crate::schema::Cardinality;
@@ -17,6 +14,7 @@ mod tests {
     use crate::schema::IntOptions;
     use crate::schema::Schema;
     use crate::Index;
+    use chrono::Duration;
 
     #[test]
     fn test_multivalued_u64() {

src/fastfield/multivalued/reader.rs

@@ -45,7 +45,7 @@ impl<Item: FastValue> MultiValueIntFastFieldReader<Item> {
     pub fn get_vals(&self, doc: DocId, vals: &mut Vec<Item>) {
         let (start, stop) = self.range(doc);
         let len = (stop - start) as usize;
-        vals.resize(len, Item::default());
+        vals.resize(len, Item::make_zero());
         self.vals_reader.get_range_u64(start, &mut vals[..]);
     }
 

src/fastfield/readers.rs

@@ -15,9 +15,11 @@ pub struct FastFieldReaders {
     fast_field_i64: HashMap<Field, FastFieldReader<i64>>,
     fast_field_u64: HashMap<Field, FastFieldReader<u64>>,
     fast_field_f64: HashMap<Field, FastFieldReader<f64>>,
+    fast_field_date: HashMap<Field, FastFieldReader<crate::DateTime>>,
     fast_field_i64s: HashMap<Field, MultiValueIntFastFieldReader<i64>>,
     fast_field_u64s: HashMap<Field, MultiValueIntFastFieldReader<u64>>,
     fast_field_f64s: HashMap<Field, MultiValueIntFastFieldReader<f64>>,
+    fast_field_dates: HashMap<Field, MultiValueIntFastFieldReader<crate::DateTime>>,
     fast_bytes: HashMap<Field, BytesFastFieldReader>,
     fast_fields_composite: CompositeFile,
 }
@@ -26,6 +28,7 @@ enum FastType {
     I64,
     U64,
     F64,
+    Date,
 }
 
 fn type_and_cardinality(field_type: &FieldType) -> Option<(FastType, Cardinality)> {
@@ -39,6 +42,9 @@ fn type_and_cardinality(field_type: &FieldType) -> Option<(FastType, Cardinality
         FieldType::F64(options) => options
             .get_fastfield_cardinality()
             .map(|cardinality| (FastType::F64, cardinality)),
+        FieldType::Date(options) => options
+            .get_fastfield_cardinality()
+            .map(|cardinality| (FastType::Date, cardinality)),
         FieldType::HierarchicalFacet => Some((FastType::U64, Cardinality::MultiValues)),
         _ => None,
     }
@@ -53,9 +59,11 @@ impl FastFieldReaders {
             fast_field_i64: Default::default(),
             fast_field_u64: Default::default(),
             fast_field_f64: Default::default(),
+            fast_field_date: Default::default(),
             fast_field_i64s: Default::default(),
             fast_field_u64s: Default::default(),
             fast_field_f64s: Default::default(),
+            fast_field_dates: Default::default(),
             fast_bytes: Default::default(),
             fast_fields_composite: fast_fields_composite.clone(),
         };
@@ -95,6 +103,12 @@ impl FastFieldReaders {
                                         FastFieldReader::open(fast_field_data.clone()),
                                     );
                                 }
+                                FastType::Date => {
+                                    fast_field_readers.fast_field_date.insert(
+                                        field,
+                                        FastFieldReader::open(fast_field_data.clone()),
+                                    );
+                                }
                             }
                         } else {
                             return Err(From::from(FastFieldNotAvailableError::new(field_entry)));
@@ -130,6 +144,14 @@ impl FastFieldReaders {
                                         .fast_field_f64s
                                         .insert(field, multivalued_int_fast_field);
                                 }
+                                FastType::Date => {
+                                    let vals_reader = FastFieldReader::open(fast_field_data);
+                                    let multivalued_int_fast_field =
+                                        MultiValueIntFastFieldReader::open(idx_reader, vals_reader);
+                                    fast_field_readers
+                                        .fast_field_dates
+                                        .insert(field, multivalued_int_fast_field);
+                                }
                             }
                         } else {
                             return Err(From::from(FastFieldNotAvailableError::new(field_entry)));
@@ -167,6 +189,9 @@ impl FastFieldReaders {
         if let Some(f64_ff_reader) = self.f64(field) {
             return Some(f64_ff_reader.into_u64_reader());
         }
+        if let Some(date_ff_reader) = self.date(field) {
+            return Some(date_ff_reader.into_u64_reader());
+        }
         None
     }
 
@@ -177,6 +202,13 @@ impl FastFieldReaders {
         self.fast_field_i64.get(&field).cloned()
     }
 
+    /// Returns the `i64` fast field reader reader associated to `field`.
+    ///
+    /// If `field` is not a i64 fast field, this method returns `None`.
+    pub fn date(&self, field: Field) -> Option<FastFieldReader<crate::DateTime>> {
+        self.fast_field_date.get(&field).cloned()
+    }
+
     /// Returns the `f64` fast field reader reader associated to `field`.
     ///
     /// If `field` is not a f64 fast field, this method returns `None`.
@@ -223,6 +255,13 @@ impl FastFieldReaders {
         self.fast_field_f64s.get(&field).cloned()
     }
 
+    /// Returns a `crate::DateTime` multi-valued fast field reader reader associated to `field`.
+    ///
+    /// If `field` is not a `crate::DateTime` multi-valued fast field, this method returns `None`.
+    pub fn dates(&self, field: Field) -> Option<MultiValueIntFastFieldReader<crate::DateTime>> {
+        self.fast_field_dates.get(&field).cloned()
+    }
+
     /// Returns the `bytes` fast field reader associated to `field`.
     ///
     /// If `field` is not a bytes fast field, returns `None`.

src/fastfield/writer.rs

@@ -4,7 +4,7 @@ use crate::common::BinarySerializable;
 use crate::common::VInt;
 use crate::fastfield::{BytesFastFieldWriter, FastFieldSerializer};
 use crate::postings::UnorderedTermId;
-use crate::schema::{Cardinality, Document, Field, FieldType, Schema};
+use crate::schema::{Cardinality, Document, Field, FieldEntry, FieldType, Schema};
 use crate::termdict::TermOrdinal;
 use fnv::FnvHashMap;
 use std::collections::HashMap;
@@ -17,6 +17,14 @@ pub struct FastFieldsWriter {
     bytes_value_writers: Vec<BytesFastFieldWriter>,
 }
 
+fn fast_field_default_value(field_entry: &FieldEntry) -> u64 {
+    match *field_entry.field_type() {
+        FieldType::I64(_) | FieldType::Date(_) => common::i64_to_u64(0i64),
+        FieldType::F64(_) => common::f64_to_u64(0.0f64),
+        _ => 0u64,
+    }
+}
+
 impl FastFieldsWriter {
     /// Create all `FastFieldWriter` required by the schema.
     pub fn from_schema(schema: &Schema) -> FastFieldsWriter {
@@ -25,18 +33,15 @@ impl FastFieldsWriter {
         let mut bytes_value_writers = Vec::new();
 
         for (field, field_entry) in schema.fields() {
-            let default_value = match *field_entry.field_type() {
-                FieldType::I64(_) => common::i64_to_u64(0i64),
-                FieldType::F64(_) => common::f64_to_u64(0.0f64),
-                _ => 0u64,
-            };
             match *field_entry.field_type() {
                 FieldType::I64(ref int_options)
                 | FieldType::U64(ref int_options)
-                | FieldType::F64(ref int_options) => {
+                | FieldType::F64(ref int_options)
+                | FieldType::Date(ref int_options) => {
                     match int_options.get_fastfield_cardinality() {
                         Some(Cardinality::SingleValue) => {
                             let mut fast_field_writer = IntFastFieldWriter::new(field);
+                            let default_value = fast_field_default_value(field_entry);
                             fast_field_writer.set_val_if_missing(default_value);
                             single_value_writers.push(fast_field_writer);
                         }

src/indexer/index_writer.rs

@@ -142,8 +142,7 @@ pub(crate) fn advance_deletes(
         return Ok(());
     }
 
-    let mut delete_cursor = segment_entry.delete_cursor().clone();
-    if segment_entry.delete_bitset().is_none() && delete_cursor.get().is_none() {
+    if segment_entry.delete_bitset().is_none() && segment_entry.delete_cursor().get().is_none() {
         // There has been no `DeleteOperation` between the segment status and `target_opstamp`.
         return Ok(());
     }
@@ -159,7 +158,7 @@ pub(crate) fn advance_deletes(
     compute_deleted_bitset(
         &mut delete_bitset,
         &segment_reader,
-        &mut delete_cursor,
+        segment_entry.delete_cursor(),
         &DocToOpstampMapping::None,
         target_opstamp,
     )?;
@@ -515,9 +514,13 @@ impl IndexWriter {
     /// Merges a given list of segments
     ///
     /// `segment_ids` is required to be non-empty.
-    pub async fn merge(&mut self, segment_ids: &[SegmentId]) -> crate::Result<SegmentMeta> {
+    pub fn merge(
+        &mut self,
+        segment_ids: &[SegmentId],
+    ) -> impl Future<Output = crate::Result<SegmentMeta>> {
         let merge_operation = self.segment_updater.make_merge_operation(segment_ids);
-        self.segment_updater.start_merge(merge_operation)?.await
+        let segment_updater = self.segment_updater.clone();
+        async move { segment_updater.start_merge(merge_operation)?.await }
     }
 
     /// Closes the current document channel send.
@@ -894,7 +897,7 @@ mod tests {
         let index_writer = index.writer(3_000_000).unwrap();
         assert_eq!(
             format!("{:?}", index_writer.get_merge_policy()),
-            "LogMergePolicy { min_merge_size: 8, min_layer_size: 10000, \
+            "LogMergePolicy { min_merge_size: 8, max_merge_size: 10000000, min_layer_size: 10000, \
              level_log_size: 0.75 }"
         );
         let merge_policy = Box::new(NoMergePolicy::default());

src/indexer/log_merge_policy.rs

@@ -6,12 +6,14 @@ use std::f64;
 const DEFAULT_LEVEL_LOG_SIZE: f64 = 0.75;
 const DEFAULT_MIN_LAYER_SIZE: u32 = 10_000;
 const DEFAULT_MIN_MERGE_SIZE: usize = 8;
+const DEFAULT_MAX_MERGE_SIZE: usize = 10_000_000;
 
 /// `LogMergePolicy` tries tries to merge segments that have a similar number of
 /// documents.
 #[derive(Debug, Clone)]
 pub struct LogMergePolicy {
     min_merge_size: usize,
+    max_merge_size: usize,
     min_layer_size: u32,
     level_log_size: f64,
 }
@@ -26,6 +28,12 @@ impl LogMergePolicy {
         self.min_merge_size = min_merge_size;
     }
 
+    /// Set the maximum number docs in a segment for it to be considered for
+    /// merging.
+    pub fn set_max_merge_size(&mut self, max_merge_size: usize) {
+        self.max_merge_size = max_merge_size;
+    }
+
     /// Set the minimum segment size under which all segment belong
     /// to the same level.
     pub fn set_min_layer_size(&mut self, min_layer_size: u32) {
@@ -53,6 +61,7 @@ impl MergePolicy for LogMergePolicy {
         let mut size_sorted_tuples = segments
             .iter()
             .map(SegmentMeta::num_docs)
+            .filter(|s| s <= &(self.max_merge_size as u32))
             .enumerate()
             .collect::<Vec<(usize, u32)>>();
 
@@ -86,6 +95,7 @@ impl Default for LogMergePolicy {
     fn default() -> LogMergePolicy {
         LogMergePolicy {
             min_merge_size: DEFAULT_MIN_MERGE_SIZE,
+            max_merge_size: DEFAULT_MAX_MERGE_SIZE,
             min_layer_size: DEFAULT_MIN_LAYER_SIZE,
             level_log_size: DEFAULT_LEVEL_LOG_SIZE,
         }
@@ -104,6 +114,7 @@ mod tests {
     fn test_merge_policy() -> LogMergePolicy {
         let mut log_merge_policy = LogMergePolicy::default();
         log_merge_policy.set_min_merge_size(3);
+        log_merge_policy.set_max_merge_size(100_000);
         log_merge_policy.set_min_layer_size(2);
         log_merge_policy
     }
@@ -141,11 +152,11 @@ mod tests {
             create_random_segment_meta(10),
             create_random_segment_meta(10),
             create_random_segment_meta(10),
-            create_random_segment_meta(1000),
-            create_random_segment_meta(1000),
-            create_random_segment_meta(1000),
-            create_random_segment_meta(10000),
-            create_random_segment_meta(10000),
+            create_random_segment_meta(1_000),
+            create_random_segment_meta(1_000),
+            create_random_segment_meta(1_000),
+            create_random_segment_meta(10_000),
+            create_random_segment_meta(10_000),
             create_random_segment_meta(10),
             create_random_segment_meta(10),
             create_random_segment_meta(10),
@@ -182,4 +193,19 @@ mod tests {
         let result_list = test_merge_policy().compute_merge_candidates(&test_input);
         assert_eq!(result_list.len(), 1);
     }
+
+    #[test]
+    fn test_large_merge_segments() {
+        let test_input = vec![
+            create_random_segment_meta(1_000_000),
+            create_random_segment_meta(100_001),
+            create_random_segment_meta(100_000),
+            create_random_segment_meta(100_000),
+            create_random_segment_meta(100_000),
+        ];
+        let result_list = test_merge_policy().compute_merge_candidates(&test_input);
+        // Do not include large segments
+        assert_eq!(result_list.len(), 1);
+        assert_eq!(result_list[0].0.len(), 3)
+    }
 }

src/indexer/mod.rs

@@ -33,6 +33,7 @@ pub type DefaultMergePolicy = LogMergePolicy;
 mod tests {
     use crate::schema::{self, Schema};
     use crate::{Index, Term};
+
     #[test]
     fn test_advance_delete_bug() {
         let mut schema_builder = Schema::builder();

src/indexer/operation.rs

@@ -19,6 +19,8 @@ pub struct AddOperation {
 /// UserOperation is an enum type that encapsulates other operation types.
 #[derive(Eq, PartialEq, Debug)]
 pub enum UserOperation {
+    /// Add operation
     Add(Document),
+    /// Delete operation
     Delete(Term),
 }

src/indexer/segment_writer.rs

@@ -11,10 +11,9 @@ use crate::schema::Schema;
 use crate::schema::Term;
 use crate::schema::Value;
 use crate::schema::{Field, FieldEntry};
-use crate::tokenizer::BoxedTokenizer;
-use crate::tokenizer::FacetTokenizer;
-use crate::tokenizer::PreTokenizedStream;
-use crate::tokenizer::{TokenStream, TokenStreamChain, Tokenizer};
+use crate::tokenizer::{BoxTokenStream, PreTokenizedStream};
+use crate::tokenizer::{FacetTokenizer, TextAnalyzer};
+use crate::tokenizer::{TokenStreamChain, Tokenizer};
 use crate::DocId;
 use crate::Opstamp;
 use crate::Result;
@@ -50,7 +49,7 @@ pub struct SegmentWriter {
     fast_field_writers: FastFieldsWriter,
     fieldnorms_writer: FieldNormsWriter,
     doc_opstamps: Vec<Opstamp>,
-    tokenizers: Vec<Option<BoxedTokenizer>>,
+    tokenizers: Vec<Option<TextAnalyzer>>,
 }
 
 impl SegmentWriter {
@@ -159,7 +158,7 @@ impl SegmentWriter {
                     }
                 }
                 FieldType::Str(_) => {
-                    let mut token_streams: Vec<Box<dyn TokenStream>> = vec![];
+                    let mut token_streams: Vec<BoxTokenStream> = vec![];
                     let mut offsets = vec![];
                     let mut total_offset = 0;
 
@@ -172,7 +171,7 @@ impl SegmentWriter {
                                 }
 
                                 token_streams
-                                    .push(Box::new(PreTokenizedStream::from(tok_str.clone())));
+                                    .push(PreTokenizedStream::from(tok_str.clone()).into());
                             }
                             Value::Str(ref text) => {
                                 if let Some(ref mut tokenizer) =
@@ -191,8 +190,7 @@ impl SegmentWriter {
                     let num_tokens = if token_streams.is_empty() {
                         0
                     } else {
-                        let mut token_stream: Box<dyn TokenStream> =
-                            Box::new(TokenStreamChain::new(offsets, token_streams));
+                        let mut token_stream = TokenStreamChain::new(offsets, token_streams);
                         self.multifield_postings
                             .index_text(doc_id, field, &mut token_stream)
                     };

src/indexer/stamper.rs

@@ -1,18 +1,76 @@
 use crate::Opstamp;
 use std::ops::Range;
-use std::sync::atomic::{AtomicU64, Ordering};
+use std::sync::atomic::Ordering;
 use std::sync::Arc;
 
+#[cfg(not(target_arch = "arm"))]
+mod atomic_impl {
+
+    use crate::Opstamp;
+    use std::sync::atomic::{AtomicU64, Ordering};
+
+    #[derive(Default)]
+    pub struct AtomicU64Wrapper(AtomicU64);
+
+    impl AtomicU64Wrapper {
+        pub fn new(first_opstamp: Opstamp) -> AtomicU64Wrapper {
+            AtomicU64Wrapper(AtomicU64::new(first_opstamp as u64))
+        }
+
+        pub fn fetch_add(&self, val: u64, order: Ordering) -> u64 {
+            self.0.fetch_add(val as u64, order) as u64
+        }
+
+        pub fn revert(&self, val: u64, order: Ordering) -> u64 {
+            self.0.store(val, order);
+            val
+        }
+    }
+}
+
+#[cfg(target_arch = "arm")]
+mod atomic_impl {
+
+    use crate::Opstamp;
+    /// Under other architecture, we rely on a mutex.
+    use std::sync::atomic::Ordering;
+    use std::sync::RwLock;
+
+    #[derive(Default)]
+    pub struct AtomicU64Wrapper(RwLock<u64>);
+
+    impl AtomicU64Wrapper {
+        pub fn new(first_opstamp: Opstamp) -> AtomicU64Wrapper {
+            AtomicU64Wrapper(RwLock::new(first_opstamp))
+        }
+
+        pub fn fetch_add(&self, incr: u64, _order: Ordering) -> u64 {
+            let mut lock = self.0.write().unwrap();
+            let previous_val = *lock;
+            *lock = previous_val + incr;
+            previous_val
+        }
+
+        pub fn revert(&self, val: u64, _order: Ordering) -> u64 {
+            let mut lock = self.0.write().unwrap();
+            *lock = val;
+            val
+        }
+    }
+}
+
+use self::atomic_impl::AtomicU64Wrapper;
+
 /// Stamper provides Opstamps, which is just an auto-increment id to label
 /// an operation.
 ///
 /// Cloning does not "fork" the stamp generation. The stamper actually wraps an `Arc`.
 #[derive(Clone, Default)]
-pub struct Stamper(Arc<AtomicU64>);
+pub struct Stamper(Arc<AtomicU64Wrapper>);
 
 impl Stamper {
     pub fn new(first_opstamp: Opstamp) -> Stamper {
-        Stamper(Arc::new(AtomicU64::new(first_opstamp)))
+        Stamper(Arc::new(AtomicU64Wrapper::new(first_opstamp)))
     }
 
     pub fn stamp(&self) -> Opstamp {
@@ -31,8 +89,7 @@ impl Stamper {
 
     /// Reverts the stamper to a given `Opstamp` value and returns it
     pub fn revert(&self, to_opstamp: Opstamp) -> Opstamp {
-        self.0.store(to_opstamp, Ordering::SeqCst);
-        to_opstamp
+        self.0.revert(to_opstamp, Ordering::SeqCst)
     }
 }
 

src/lib.rs

@@ -161,10 +161,11 @@ pub use self::snippet::{Snippet, SnippetGenerator};
 mod docset;
 pub use self::docset::{DocSet, SkipResult};
 pub use crate::common::{f64_to_u64, i64_to_u64, u64_to_f64, u64_to_i64};
-pub use crate::core::SegmentComponent;
+pub use crate::core::{Executor, SegmentComponent};
 pub use crate::core::{Index, IndexMeta, Searcher, Segment, SegmentId, SegmentMeta};
 pub use crate::core::{InvertedIndexReader, SegmentReader};
 pub use crate::directory::Directory;
+pub use crate::indexer::operation::UserOperation;
 pub use crate::indexer::IndexWriter;
 pub use crate::postings::Postings;
 pub use crate::reader::LeasedItem;
@@ -940,4 +941,73 @@ mod tests {
             assert_eq!(fast_field_reader.get(0), 4f64)
         }
     }
+
+    // motivated by #729
+    #[test]
+    fn test_update_via_delete_insert() {
+        use crate::collector::Count;
+        use crate::indexer::NoMergePolicy;
+        use crate::query::AllQuery;
+        use crate::SegmentId;
+        use futures::executor::block_on;
+
+        const DOC_COUNT: u64 = 2u64;
+
+        let mut schema_builder = SchemaBuilder::default();
+        let id = schema_builder.add_u64_field("id", INDEXED);
+        let schema = schema_builder.build();
+
+        let index = Index::create_in_ram(schema.clone());
+        let index_reader = index.reader().unwrap();
+
+        let mut index_writer = index.writer(3_000_000).unwrap();
+        index_writer.set_merge_policy(Box::new(NoMergePolicy));
+
+        for doc_id in 0u64..DOC_COUNT {
+            index_writer.add_document(doc!(id => doc_id));
+        }
+        index_writer.commit().unwrap();
+
+        index_reader.reload().unwrap();
+        let searcher = index_reader.searcher();
+
+        assert_eq!(
+            searcher.search(&AllQuery, &Count).unwrap(),
+            DOC_COUNT as usize
+        );
+
+        // update the 10 elements by deleting and re-adding
+        for doc_id in 0u64..DOC_COUNT {
+            index_writer.delete_term(Term::from_field_u64(id, doc_id));
+            index_writer.commit().unwrap();
+            index_reader.reload().unwrap();
+            let doc = doc!(id =>  doc_id);
+            index_writer.add_document(doc);
+            index_writer.commit().unwrap();
+            index_reader.reload().unwrap();
+            let searcher = index_reader.searcher();
+            // The number of document should be stable.
+            assert_eq!(
+                searcher.search(&AllQuery, &Count).unwrap(),
+                DOC_COUNT as usize
+            );
+        }
+
+        index_reader.reload().unwrap();
+        let searcher = index_reader.searcher();
+        let segment_ids: Vec<SegmentId> = searcher
+            .segment_readers()
+            .into_iter()
+            .map(|reader| reader.segment_id())
+            .collect();
+        block_on(index_writer.merge(&segment_ids)).unwrap();
+
+        index_reader.reload().unwrap();
+        let searcher = index_reader.searcher();
+
+        assert_eq!(
+            searcher.search(&AllQuery, &Count).unwrap(),
+            DOC_COUNT as usize
+        );
+    }
 }

src/postings/serializer.rs

@@ -148,8 +148,7 @@ impl<'a> FieldSerializer<'a> {
             }
             _ => (false, false),
         };
-        let term_dictionary_builder =
-            TermDictionaryBuilder::create(term_dictionary_write, &field_type)?;
+        let term_dictionary_builder = TermDictionaryBuilder::create(term_dictionary_write)?;
         let postings_serializer =
             PostingsSerializer::new(postings_write, term_freq_enabled, position_enabled);
         let positions_serializer_opt = if position_enabled {

src/query/automaton_weight.rs

@@ -15,6 +15,7 @@ use tantivy_fst::Automaton;
 pub struct AutomatonWeight<A> {
     field: Field,
     automaton: Arc<A>,
+    boost: f32,
 }
 
 impl<A> AutomatonWeight<A>
@@ -26,9 +27,15 @@ where
         AutomatonWeight {
             field,
             automaton: automaton.into(),
+            boost: 1.0,
         }
     }
 
+    /// Boost the scorer by the given factor.
+    pub fn boost_by(self, boost: f32) -> Self {
+        Self { boost, ..self }
+    }
+
     fn automaton_stream<'a>(&'a self, term_dict: &'a TermDictionary) -> TermStreamer<'a, &'a A> {
         let automaton: &A = &*self.automaton;
         let term_stream_builder = term_dict.search(automaton);
@@ -58,7 +65,7 @@ where
             }
         }
         let doc_bitset = BitSetDocSet::from(doc_bitset);
-        Ok(Box::new(ConstScorer::new(doc_bitset)))
+        Ok(Box::new(ConstScorer::with_score(doc_bitset, self.boost)))
     }
 
     fn explain(&self, reader: &SegmentReader, doc: DocId) -> Result<Explanation> {

src/query/bm25.rs

@@ -34,7 +34,7 @@ pub struct BM25Weight {
 }
 
 impl BM25Weight {
-    pub fn for_terms(searcher: &Searcher, terms: &[Term]) -> BM25Weight {
+    pub fn for_terms(searcher: &Searcher, terms: &[Term], boost: f32) -> BM25Weight {
         assert!(!terms.is_empty(), "BM25 requires at least one term");
         let field = terms[0].field();
         for term in &terms[1..] {
@@ -75,11 +75,11 @@ impl BM25Weight {
                 .sum::<f32>();
             idf_explain = Explanation::new("idf", idf);
         }
-        BM25Weight::new(idf_explain, average_fieldnorm)
+        BM25Weight::new(idf_explain, average_fieldnorm, boost)
     }
 
-    fn new(idf_explain: Explanation, average_fieldnorm: f32) -> BM25Weight {
-        let weight = idf_explain.value() * (1f32 + K1);
+    fn new(idf_explain: Explanation, average_fieldnorm: f32, boost: f32) -> BM25Weight {
+        let weight = idf_explain.value() * (1f32 + K1) * boost;
         BM25Weight {
             idf_explain,
             weight,

src/query/fuzzy_query.rs

@@ -79,6 +79,7 @@ pub struct FuzzyTermQuery {
     transposition_cost_one: bool,
     ///
     prefix: bool,
+    boost: f32,
 }
 
 impl FuzzyTermQuery {
@@ -89,6 +90,7 @@ impl FuzzyTermQuery {
             distance,
             transposition_cost_one,
             prefix: false,
+            boost: 1.0,
         }
     }
 
@@ -99,16 +101,22 @@ impl FuzzyTermQuery {
             distance,
             transposition_cost_one,
             prefix: true,
+            boost: 1.0,
         }
     }
 
+    /// Boost the query score by the given factor.
+    pub fn boost_by(self, boost: f32) -> Self {
+        Self { boost, ..self }
+    }
+
     fn specialized_weight(&self) -> Result<AutomatonWeight<DFA>> {
         // LEV_BUILDER is a HashMap, whose `get` method returns an Option
         match LEV_BUILDER.get(&(self.distance, false)) {
             // Unwrap the option and build the Ok(AutomatonWeight)
             Some(automaton_builder) => {
                 let automaton = automaton_builder.build_dfa(self.term.text());
-                Ok(AutomatonWeight::new(self.term.field(), automaton))
+                Ok(AutomatonWeight::new(self.term.field(), automaton).boost_by(self.boost))
             }
             None => Err(InvalidArgument(format!(
                 "Levenshtein distance of {} is not allowed. Choose a value in the {:?} range",

src/query/phrase_query/phrase_query.rs

@@ -27,6 +27,7 @@ use std::collections::BTreeSet;
 pub struct PhraseQuery {
     field: Field,
     phrase_terms: Vec<(usize, Term)>,
+    boost: f32,
 }
 
 impl PhraseQuery {
@@ -57,9 +58,15 @@ impl PhraseQuery {
         PhraseQuery {
             field,
             phrase_terms: terms,
+            boost: 1.0,
         }
     }
 
+    /// Boost the query score by the given factor.
+    pub fn boost_by(self, boost: f32) -> Self {
+        Self { boost, ..self }
+    }
+
     /// The `Field` this `PhraseQuery` is targeting.
     pub fn field(&self) -> Field {
         self.field
@@ -97,7 +104,7 @@ impl PhraseQuery {
             )));
         }
         let terms = self.phrase_terms();
-        let bm25_weight = BM25Weight::for_terms(searcher, &terms);
+        let bm25_weight = BM25Weight::for_terms(searcher, &terms, self.boost);
         Ok(PhraseWeight::new(
             self.phrase_terms.clone(),
             bm25_weight,

src/query/query_parser/query_parser.rs

@@ -8,7 +8,7 @@ use crate::query::PhraseQuery;
 use crate::query::Query;
 use crate::query::RangeQuery;
 use crate::query::TermQuery;
-use crate::schema::IndexRecordOption;
+use crate::schema::{Facet, IndexRecordOption};
 use crate::schema::{Field, Schema};
 use crate::schema::{FieldType, Term};
 use crate::tokenizer::TokenizerManager;
@@ -319,7 +319,10 @@ impl QueryParser {
                     ))
                 }
             }
-            FieldType::HierarchicalFacet => Ok(vec![(0, Term::from_field_text(field, phrase))]),
+            FieldType::HierarchicalFacet => {
+                let facet = Facet::from_text(phrase);
+                Ok(vec![(0, Term::from_field_text(field, facet.encoded_str()))])
+            }
             FieldType::Bytes => {
                 let field_name = self.schema.get_field_name(field).to_string();
                 Err(QueryParserError::FieldNotIndexed(field_name))
@@ -530,7 +533,7 @@ mod test {
     use crate::schema::{IndexRecordOption, TextFieldIndexing, TextOptions};
     use crate::schema::{Schema, Term, INDEXED, STORED, STRING, TEXT};
     use crate::tokenizer::{
-        LowerCaser, SimpleTokenizer, StopWordFilter, Tokenizer, TokenizerManager,
+        LowerCaser, SimpleTokenizer, StopWordFilter, TextAnalyzer, TokenizerManager,
     };
     use crate::Index;
     use matches::assert_matches;
@@ -554,12 +557,13 @@ mod test {
         schema_builder.add_text_field("with_stop_words", text_options);
         schema_builder.add_date_field("date", INDEXED);
         schema_builder.add_f64_field("float", INDEXED);
+        schema_builder.add_facet_field("facet");
         let schema = schema_builder.build();
         let default_fields = vec![title, text];
         let tokenizer_manager = TokenizerManager::default();
         tokenizer_manager.register(
             "en_with_stop_words",
-            SimpleTokenizer
+            TextAnalyzer::from(SimpleTokenizer)
                 .filter(LowerCaser)
                 .filter(StopWordFilter::remove(vec!["the".to_string()])),
         );
@@ -588,9 +592,13 @@ mod test {
     }
 
     #[test]
-    pub fn test_parse_query_simple() {
+    pub fn test_parse_query_facet() {
         let query_parser = make_query_parser();
-        assert!(query_parser.parse_query("toto").is_ok());
+        let query = query_parser.parse_query("facet:/root/branch/leaf").unwrap();
+        assert_eq!(
+            format!("{:?}", query),
+            "TermQuery(Term(field=11,bytes=[114, 111, 111, 116, 0, 98, 114, 97, 110, 99, 104, 0, 108, 101, 97, 102]))"
+        );
     }
 
     #[test]

src/query/regex_query.rs

@@ -54,6 +54,7 @@ use tantivy_fst::Regex;
 pub struct RegexQuery {
     regex: Arc<Regex>,
     field: Field,
+    boost: f32,
 }
 
 impl RegexQuery {
@@ -69,11 +70,17 @@ impl RegexQuery {
         RegexQuery {
             regex: regex.into(),
             field,
+            boost: 1.0,
         }
     }
 
+    /// Boost the query score by the given factor.
+    pub fn boost_by(self, boost: f32) -> Self {
+        Self { boost, ..self }
+    }
+
     fn specialized_weight(&self) -> AutomatonWeight<Regex> {
-        AutomatonWeight::new(self.field, self.regex.clone())
+        AutomatonWeight::new(self.field, self.regex.clone()).boost_by(self.boost)
     }
 }
 

src/query/scorer.rs

@@ -56,6 +56,11 @@ impl<TDocSet: DocSet> ConstScorer<TDocSet> {
         }
     }
 
+    /// Creates a new `ConstScorer` with a custom score value
+    pub fn with_score(docset: TDocSet, score: f32) -> ConstScorer<TDocSet> {
+        ConstScorer { docset, score }
+    }
+
     /// Sets the constant score to a different value.
     pub fn set_score(&mut self, score: Score) {
         self.score = score;

src/query/term_query/mod.rs

@@ -45,6 +45,35 @@ mod tests {
         assert_eq!(term_scorer.score(), 0.28768212);
     }
 
+    #[test]
+    pub fn test_term_query_boost_by() {
+        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);
+        {
+            // writing the segment
+            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            {
+                let doc = doc!(text_field => "a");
+                index_writer.add_document(doc);
+            }
+            assert!(index_writer.commit().is_ok());
+        }
+        let searcher = index.reader().unwrap().searcher();
+        let term_query = TermQuery::new(
+            Term::from_field_text(text_field, "a"),
+            IndexRecordOption::Basic,
+        )
+        .boost_by(42.0);
+        let term_weight = term_query.weight(&searcher, true).unwrap();
+        let segment_reader = searcher.segment_reader(0);
+        let mut term_scorer = term_weight.scorer(segment_reader).unwrap();
+        assert!(term_scorer.advance());
+        assert_eq!(term_scorer.doc(), 0);
+        assert_nearly_equals(0.28768212 * 42.0, term_scorer.score());
+    }
+
     #[test]
     pub fn test_term_weight() {
         let mut schema_builder = Schema::builder();

src/query/term_query/term_query.rs

@@ -61,6 +61,7 @@ use std::fmt;
 pub struct TermQuery {
     term: Term,
     index_record_option: IndexRecordOption,
+    boost: f32,
 }
 
 impl fmt::Debug for TermQuery {
@@ -75,9 +76,15 @@ impl TermQuery {
         TermQuery {
             term,
             index_record_option: segment_postings_options,
+            boost: 1.0,
         }
     }
 
+    /// Boost the query score by the given factor.
+    pub fn boost_by(self, boost: f32) -> Self {
+        Self { boost, ..self }
+    }
+
     /// The `Term` this query is built out of.
     pub fn term(&self) -> &Term {
         &self.term
@@ -90,7 +97,7 @@ impl TermQuery {
     /// This is useful for optimization purpose.
     pub fn specialized_weight(&self, searcher: &Searcher, scoring_enabled: bool) -> TermWeight {
         let term = self.term.clone();
-        let bm25_weight = BM25Weight::for_terms(searcher, &[term]);
+        let bm25_weight = BM25Weight::for_terms(searcher, &[term], self.boost);
         let index_record_option = if scoring_enabled {
             self.index_record_option
         } else {

src/schema/facet.rs

@@ -122,6 +122,11 @@ impl Facet {
     pub fn to_path(&self) -> Vec<&str> {
         self.encoded_str().split(|c| c == FACET_SEP_CHAR).collect()
     }
+
+    /// This function is the inverse of Facet::from(&str).
+    pub fn to_path_string(&self) -> String {
+        format!("{}", self.to_string())
+    }
 }
 
 impl Borrow<str> for Facet {
@@ -265,4 +270,21 @@ mod tests {
         let facet = Facet::from_path(v.iter());
         assert_eq!(facet.to_path(), v);
     }
+
+    #[test]
+    fn test_to_path_string() {
+        let v = ["first", "second", "third/not_fourth"];
+        let facet = Facet::from_path(v.iter());
+        assert_eq!(
+            facet.to_path_string(),
+            String::from("/first/second/third\\/not_fourth")
+        );
+    }
+
+    #[test]
+    fn test_to_path_string_empty() {
+        let v: Vec<&str> = vec![];
+        let facet = Facet::from_path(v.iter());
+        assert_eq!(facet.to_path_string(), "/");
+    }
 }

src/schema/schema.rs

@@ -166,7 +166,7 @@ impl SchemaBuilder {
     }
 
     /// Adds a field entry to the schema in build.
-    fn add_field(&mut self, field_entry: FieldEntry) -> Field {
+    pub fn add_field(&mut self, field_entry: FieldEntry) -> Field {
         let field = Field::from_field_id(self.fields.len() as u32);
         let field_name = field_entry.name().to_string();
         self.fields.push(field_entry);
@@ -401,6 +401,7 @@ pub enum DocParsingError {
 mod tests {
 
     use crate::schema::field_type::ValueParsingError;
+    use crate::schema::int_options::Cardinality::SingleValue;
     use crate::schema::schema::DocParsingError::NotJSON;
     use crate::schema::*;
     use matches::{assert_matches, matches};
@@ -715,4 +716,94 @@ mod tests {
             assert_matches!(json_err, Err(NotJSON(_)));
         }
     }
+
+    #[test]
+    pub fn test_schema_add_field() {
+        let mut schema_builder = SchemaBuilder::default();
+        let id_options = TextOptions::default().set_stored().set_indexing_options(
+            TextFieldIndexing::default()
+                .set_tokenizer("raw")
+                .set_index_option(IndexRecordOption::Basic),
+        );
+        let timestamp_options = IntOptions::default()
+            .set_stored()
+            .set_indexed()
+            .set_fast(SingleValue);
+        schema_builder.add_text_field("_id", id_options);
+        schema_builder.add_date_field("_timestamp", timestamp_options);
+
+        let schema_content = r#"[
+  {
+    "name": "text",
+    "type": "text",
+    "options": {
+      "indexing": {
+        "record": "position",
+        "tokenizer": "default"
+      },
+      "stored": false
+    }
+  },
+  {
+    "name": "popularity",
+    "type": "i64",
+    "options": {
+      "indexed": false,
+      "fast": "single",
+      "stored": true
+    }
+  }
+]"#;
+        let tmp_schema: Schema =
+            serde_json::from_str(&schema_content).expect("error while reading json");
+        for (_field, field_entry) in tmp_schema.fields() {
+            schema_builder.add_field(field_entry.clone());
+        }
+
+        let schema = schema_builder.build();
+        let schema_json = serde_json::to_string_pretty(&schema).unwrap();
+        let expected = r#"[
+  {
+    "name": "_id",
+    "type": "text",
+    "options": {
+      "indexing": {
+        "record": "basic",
+        "tokenizer": "raw"
+      },
+      "stored": true
+    }
+  },
+  {
+    "name": "_timestamp",
+    "type": "date",
+    "options": {
+      "indexed": true,
+      "fast": "single",
+      "stored": true
+    }
+  },
+  {
+    "name": "text",
+    "type": "text",
+    "options": {
+      "indexing": {
+        "record": "position",
+        "tokenizer": "default"
+      },
+      "stored": false
+    }
+  },
+  {
+    "name": "popularity",
+    "type": "i64",
+    "options": {
+      "indexed": false,
+      "fast": "single",
+      "stored": true
+    }
+  }
+]"#;
+        assert_eq!(schema_json, expected);
+    }
 }

src/schema/value.rs

@@ -209,8 +209,8 @@ impl From<f64> for Value {
     }
 }
 
-impl From<DateTime> for Value {
-    fn from(date_time: DateTime) -> Value {
+impl From<crate::DateTime> for Value {
+    fn from(date_time: crate::DateTime) -> Value {
         Value::Date(date_time)
     }
 }

src/snippet/mod.rs

@@ -1,8 +1,7 @@
 use crate::query::Query;
 use crate::schema::Field;
 use crate::schema::Value;
-use crate::tokenizer::BoxedTokenizer;
-use crate::tokenizer::{Token, TokenStream};
+use crate::tokenizer::{TextAnalyzer, Token};
 use crate::Document;
 use crate::Result;
 use crate::Searcher;
@@ -142,7 +141,7 @@ impl Snippet {
 /// Fragments must be valid in the sense that `&text[fragment.start..fragment.stop]`\
 /// has to be a valid string.
 fn search_fragments<'a>(
-    tokenizer: &BoxedTokenizer,
+    tokenizer: &TextAnalyzer,
     text: &'a str,
     terms: &BTreeMap<String, f32>,
     max_num_chars: usize,
@@ -251,7 +250,7 @@ fn select_best_fragment_combination(fragments: &[FragmentCandidate], text: &str)
 /// ```
 pub struct SnippetGenerator {
     terms_text: BTreeMap<String, f32>,
-    tokenizer: BoxedTokenizer,
+    tokenizer: TextAnalyzer,
     field: Field,
     max_num_chars: usize,
 }
@@ -347,12 +346,11 @@ Survey in 2016, 2017, and 2018."#;
 
     #[test]
     fn test_snippet() {
-        let boxed_tokenizer = SimpleTokenizer.into();
         let terms = btreemap! {
             String::from("rust") => 1.0,
             String::from("language") => 0.9
         };
-        let fragments = search_fragments(&boxed_tokenizer, TEST_TEXT, &terms, 100);
+        let fragments = search_fragments(&From::from(SimpleTokenizer), TEST_TEXT, &terms, 100);
         assert_eq!(fragments.len(), 7);
         {
             let first = &fragments[0];
@@ -374,13 +372,12 @@ Survey in 2016, 2017, and 2018."#;
 
     #[test]
     fn test_snippet_scored_fragment() {
-        let boxed_tokenizer = SimpleTokenizer.into();
         {
             let terms = btreemap! {
                 String::from("rust") =>1.0f32,
                 String::from("language") => 0.9f32
             };
-            let fragments = search_fragments(&boxed_tokenizer, TEST_TEXT, &terms, 20);
+            let fragments = search_fragments(&From::from(SimpleTokenizer), TEST_TEXT, &terms, 20);
             {
                 let first = &fragments[0];
                 assert_eq!(first.score, 1.0);
@@ -389,13 +386,12 @@ Survey in 2016, 2017, and 2018."#;
             let snippet = select_best_fragment_combination(&fragments[..], &TEST_TEXT);
             assert_eq!(snippet.to_html(), "<b>Rust</b> is a systems")
         }
-        let boxed_tokenizer = SimpleTokenizer.into();
         {
             let terms = btreemap! {
                 String::from("rust") =>0.9f32,
                 String::from("language") => 1.0f32
             };
-            let fragments = search_fragments(&boxed_tokenizer, TEST_TEXT, &terms, 20);
+            let fragments = search_fragments(&From::from(SimpleTokenizer), TEST_TEXT, &terms, 20);
             //assert_eq!(fragments.len(), 7);
             {
                 let first = &fragments[0];
@@ -409,14 +405,12 @@ Survey in 2016, 2017, and 2018."#;
 
     #[test]
     fn test_snippet_in_second_fragment() {
-        let boxed_tokenizer = SimpleTokenizer.into();
-
         let text = "a b c d e f g";
 
         let mut terms = BTreeMap::new();
         terms.insert(String::from("c"), 1.0);
 
-        let fragments = search_fragments(&boxed_tokenizer, &text, &terms, 3);
+        let fragments = search_fragments(&From::from(SimpleTokenizer), &text, &terms, 3);
 
         assert_eq!(fragments.len(), 1);
         {
@@ -433,14 +427,12 @@ Survey in 2016, 2017, and 2018."#;
 
     #[test]
     fn test_snippet_with_term_at_the_end_of_fragment() {
-        let boxed_tokenizer = SimpleTokenizer.into();
-
         let text = "a b c d e f f g";
 
         let mut terms = BTreeMap::new();
         terms.insert(String::from("f"), 1.0);
 
-        let fragments = search_fragments(&boxed_tokenizer, &text, &terms, 3);
+        let fragments = search_fragments(&From::from(SimpleTokenizer), &text, &terms, 3);
 
         assert_eq!(fragments.len(), 2);
         {
@@ -457,15 +449,13 @@ Survey in 2016, 2017, and 2018."#;
 
     #[test]
     fn test_snippet_with_second_fragment_has_the_highest_score() {
-        let boxed_tokenizer = SimpleTokenizer.into();
-
         let text = "a b c d e f g";
 
         let mut terms = BTreeMap::new();
         terms.insert(String::from("f"), 1.0);
         terms.insert(String::from("a"), 0.9);
 
-        let fragments = search_fragments(&boxed_tokenizer, &text, &terms, 7);
+        let fragments = search_fragments(&From::from(SimpleTokenizer), &text, &terms, 7);
 
         assert_eq!(fragments.len(), 2);
         {
@@ -482,14 +472,12 @@ Survey in 2016, 2017, and 2018."#;
 
     #[test]
     fn test_snippet_with_term_not_in_text() {
-        let boxed_tokenizer = SimpleTokenizer.into();
-
         let text = "a b c d";
 
         let mut terms = BTreeMap::new();
         terms.insert(String::from("z"), 1.0);
 
-        let fragments = search_fragments(&boxed_tokenizer, &text, &terms, 3);
+        let fragments = search_fragments(&From::from(SimpleTokenizer), &text, &terms, 3);
 
         assert_eq!(fragments.len(), 0);
 
@@ -500,12 +488,10 @@ Survey in 2016, 2017, and 2018."#;
 
     #[test]
     fn test_snippet_with_no_terms() {
-        let boxed_tokenizer = SimpleTokenizer.into();
-
         let text = "a b c d";
 
         let terms = BTreeMap::new();
-        let fragments = search_fragments(&boxed_tokenizer, &text, &terms, 3);
+        let fragments = search_fragments(&From::from(SimpleTokenizer), &text, &terms, 3);
         assert_eq!(fragments.len(), 0);
 
         let snippet = select_best_fragment_combination(&fragments[..], &text);

src/termdict/mod.rs

@@ -38,7 +38,7 @@ mod tests {
     use crate::core::Index;
     use crate::directory::{Directory, RAMDirectory, ReadOnlySource};
     use crate::postings::TermInfo;
-    use crate::schema::{Document, FieldType, Schema, TEXT};
+    use crate::schema::{Document, Schema, TEXT};
     use std::path::PathBuf;
     use std::str;
 
@@ -52,6 +52,12 @@ mod tests {
         }
     }
 
+    #[test]
+    fn test_empty_term_dictionary() {
+        let empty = TermDictionary::empty();
+        assert!(empty.stream().next().is_none());
+    }
+
     #[test]
     fn test_term_ordinals() {
         const COUNTRIES: [&'static str; 7] = [
@@ -67,9 +73,7 @@ mod tests {
         let path = PathBuf::from("TermDictionary");
         {
             let write = directory.open_write(&path).unwrap();
-            let field_type = FieldType::Str(TEXT);
-            let mut term_dictionary_builder =
-                TermDictionaryBuilder::create(write, &field_type).unwrap();
+            let mut term_dictionary_builder = TermDictionaryBuilder::create(write).unwrap();
             for term in COUNTRIES.iter() {
                 term_dictionary_builder
                     .insert(term.as_bytes(), &make_term_info(0u64))
@@ -93,9 +97,7 @@ mod tests {
         let path = PathBuf::from("TermDictionary");
         {
             let write = directory.open_write(&path).unwrap();
-            let field_type = FieldType::Str(TEXT);
-            let mut term_dictionary_builder =
-                TermDictionaryBuilder::create(write, &field_type).unwrap();
+            let mut term_dictionary_builder = TermDictionaryBuilder::create(write).unwrap();
             term_dictionary_builder
                 .insert("abc".as_bytes(), &make_term_info(34u64))
                 .unwrap();
@@ -179,10 +181,8 @@ mod tests {
         let ids: Vec<_> = (0u32..10_000u32)
             .map(|i| (format!("doc{:0>6}", i), i))
             .collect();
-        let field_type = FieldType::Str(TEXT);
         let buffer: Vec<u8> = {
-            let mut term_dictionary_builder =
-                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
+            let mut term_dictionary_builder = TermDictionaryBuilder::create(vec![]).unwrap();
             for &(ref id, ref i) in &ids {
                 term_dictionary_builder
                     .insert(id.as_bytes(), &make_term_info(*i as u64))
@@ -209,10 +209,8 @@ mod tests {
 
     #[test]
     fn test_stream_high_range_prefix_suffix() {
-        let field_type = FieldType::Str(TEXT);
         let buffer: Vec<u8> = {
-            let mut term_dictionary_builder =
-                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
+            let mut term_dictionary_builder = TermDictionaryBuilder::create(vec![]).unwrap();
             // term requires more than 16bits
             term_dictionary_builder
                 .insert("abcdefghijklmnopqrstuvwxy", &make_term_info(1))
@@ -244,10 +242,8 @@ mod tests {
         let ids: Vec<_> = (0u32..10_000u32)
             .map(|i| (format!("doc{:0>6}", i), i))
             .collect();
-        let field_type = FieldType::Str(TEXT);
         let buffer: Vec<u8> = {
-            let mut term_dictionary_builder =
-                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
+            let mut term_dictionary_builder = TermDictionaryBuilder::create(vec![]).unwrap();
             for &(ref id, ref i) in &ids {
                 term_dictionary_builder
                     .insert(id.as_bytes(), &make_term_info(*i as u64))
@@ -313,10 +309,8 @@ mod tests {
 
     #[test]
     fn test_empty_string() {
-        let field_type = FieldType::Str(TEXT);
         let buffer: Vec<u8> = {
-            let mut term_dictionary_builder =
-                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
+            let mut term_dictionary_builder = TermDictionaryBuilder::create(vec![]).unwrap();
             term_dictionary_builder
                 .insert(&[], &make_term_info(1 as u64))
                 .unwrap();
@@ -337,10 +331,8 @@ mod tests {
 
     #[test]
     fn test_stream_range_boundaries() {
-        let field_type = FieldType::Str(TEXT);
         let buffer: Vec<u8> = {
-            let mut term_dictionary_builder =
-                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
+            let mut term_dictionary_builder = TermDictionaryBuilder::create(vec![]).unwrap();
             for i in 0u8..10u8 {
                 let number_arr = [i; 1];
                 term_dictionary_builder
@@ -352,41 +344,91 @@ mod tests {
         let source = ReadOnlySource::from(buffer);
         let term_dictionary: TermDictionary = TermDictionary::from_source(&source);
 
-        let value_list = |mut streamer: TermStreamer<'_>| {
+        let value_list = |mut streamer: TermStreamer<'_>, backwards: bool| {
             let mut res: Vec<u32> = vec![];
             while let Some((_, ref v)) = streamer.next() {
                 res.push(v.doc_freq);
             }
+            if backwards {
+                res.reverse();
+            }
             res
         };
+        {
+            let range = term_dictionary.range().backward().into_stream();
+            assert_eq!(
+                value_list(range, true),
+                vec![0u32, 1u32, 2u32, 3u32, 4u32, 5u32, 6u32, 7u32, 8u32, 9u32]
+            );
+        }
         {
             let range = term_dictionary.range().ge([2u8]).into_stream();
             assert_eq!(
-                value_list(range),
+                value_list(range, false),
+                vec![2u32, 3u32, 4u32, 5u32, 6u32, 7u32, 8u32, 9u32]
+            );
+        }
+        {
+            let range = term_dictionary.range().ge([2u8]).backward().into_stream();
+            assert_eq!(
+                value_list(range, true),
                 vec![2u32, 3u32, 4u32, 5u32, 6u32, 7u32, 8u32, 9u32]
             );
         }
         {
             let range = term_dictionary.range().gt([2u8]).into_stream();
             assert_eq!(
-                value_list(range),
+                value_list(range, false),
+                vec![3u32, 4u32, 5u32, 6u32, 7u32, 8u32, 9u32]
+            );
+        }
+        {
+            let range = term_dictionary.range().gt([2u8]).backward().into_stream();
+            assert_eq!(
+                value_list(range, true),
                 vec![3u32, 4u32, 5u32, 6u32, 7u32, 8u32, 9u32]
             );
         }
         {
             let range = term_dictionary.range().lt([6u8]).into_stream();
-            assert_eq!(value_list(range), vec![0u32, 1u32, 2u32, 3u32, 4u32, 5u32]);
+            assert_eq!(
+                value_list(range, false),
+                vec![0u32, 1u32, 2u32, 3u32, 4u32, 5u32]
+            );
+        }
+        {
+            let range = term_dictionary.range().lt([6u8]).backward().into_stream();
+            assert_eq!(
+                value_list(range, true),
+                vec![0u32, 1u32, 2u32, 3u32, 4u32, 5u32]
+            );
         }
         {
             let range = term_dictionary.range().le([6u8]).into_stream();
             assert_eq!(
-                value_list(range),
+                value_list(range, false),
+                vec![0u32, 1u32, 2u32, 3u32, 4u32, 5u32, 6u32]
+            );
+        }
+        {
+            let range = term_dictionary.range().le([6u8]).backward().into_stream();
+            assert_eq!(
+                value_list(range, true),
                 vec![0u32, 1u32, 2u32, 3u32, 4u32, 5u32, 6u32]
             );
         }
         {
             let range = term_dictionary.range().ge([0u8]).lt([5u8]).into_stream();
-            assert_eq!(value_list(range), vec![0u32, 1u32, 2u32, 3u32, 4u32]);
+            assert_eq!(value_list(range, false), vec![0u32, 1u32, 2u32, 3u32, 4u32]);
+        }
+        {
+            let range = term_dictionary
+                .range()
+                .ge([0u8])
+                .lt([5u8])
+                .backward()
+                .into_stream();
+            assert_eq!(value_list(range, true), vec![0u32, 1u32, 2u32, 3u32, 4u32]);
         }
     }
 
@@ -408,9 +450,7 @@ mod tests {
         let path = PathBuf::from("TermDictionary");
         {
             let write = directory.open_write(&path).unwrap();
-            let field_type = FieldType::Str(TEXT);
-            let mut term_dictionary_builder =
-                TermDictionaryBuilder::create(write, &field_type).unwrap();
+            let mut term_dictionary_builder = TermDictionaryBuilder::create(write).unwrap();
             for term in COUNTRIES.iter() {
                 term_dictionary_builder
                     .insert(term.as_bytes(), &make_term_info(0u64))

src/termdict/streamer.rs

@@ -51,6 +51,12 @@ where
         self
     }
 
+    /// Iterate over the range backwards.
+    pub fn backward(mut self) -> Self {
+        self.stream_builder = self.stream_builder.backward();
+        self
+    }
+
     /// Creates the stream corresponding to the range
     /// of terms defined using the `TermStreamerBuilder`.
     pub fn into_stream(self) -> TermStreamer<'a, A> {

src/termdict/termdict.rs

@@ -4,8 +4,8 @@ use crate::common::BinarySerializable;
 use crate::common::CountingWriter;
 use crate::directory::ReadOnlySource;
 use crate::postings::TermInfo;
-use crate::schema::FieldType;
 use crate::termdict::TermOrdinal;
+use once_cell::sync::Lazy;
 use std::io::{self, Write};
 use tantivy_fst;
 use tantivy_fst::raw::Fst;
@@ -29,7 +29,7 @@ where
     W: Write,
 {
     /// Creates a new `TermDictionaryBuilder`
-    pub fn create(w: W, _field_type: &FieldType) -> io::Result<Self> {
+    pub fn create(w: W) -> io::Result<Self> {
         let fst_builder = tantivy_fst::MapBuilder::new(w).map_err(convert_fst_error)?;
         Ok(TermDictionaryBuilder {
             fst_builder,
@@ -92,6 +92,14 @@ fn open_fst_index(source: ReadOnlySource) -> tantivy_fst::Map<ReadOnlySource> {
     tantivy_fst::Map::from(fst)
 }
 
+static EMPTY_DATA_SOURCE: Lazy<ReadOnlySource> = Lazy::new(|| {
+    let term_dictionary_data: Vec<u8> = TermDictionaryBuilder::create(Vec::<u8>::new())
+        .expect("Creating a TermDictionaryBuilder in a Vec<u8> should never fail")
+        .finish()
+        .expect("Writing in a Vec<u8> should never fail");
+    ReadOnlySource::from(term_dictionary_data)
+});
+
 /// The term dictionary contains all of the terms in
 /// `tantivy index` in a sorted manner.
 ///
@@ -122,14 +130,8 @@ 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::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");
-        let source = ReadOnlySource::from(term_dictionary_data);
-        Self::from_source(&source)
+    pub fn empty() -> Self {
+        TermDictionary::from_source(&*EMPTY_DATA_SOURCE)
     }
 
     /// Returns the number of terms in the dictionary.

src/tokenizer/alphanum_only.rs

@@ -2,7 +2,7 @@
 //! ```rust
 //! use tantivy::tokenizer::*;
 //!
-//! let tokenizer = RawTokenizer
+//! let tokenizer = TextAnalyzer::from(RawTokenizer)
 //!   .filter(AlphaNumOnlyFilter);
 //!
 //! let mut stream = tokenizer.token_stream("hello there");
@@ -10,7 +10,7 @@
 //! // contains a space
 //! assert!(stream.next().is_none());
 //!
-//! let tokenizer = SimpleTokenizer
+//! let tokenizer = TextAnalyzer::from(SimpleTokenizer)
 //!   .filter(AlphaNumOnlyFilter);
 //!
 //! let mut stream = tokenizer.token_stream("hello there 💣");
@@ -19,56 +19,30 @@
 //! // the "emoji" is dropped because its not an alphanum
 //! assert!(stream.next().is_none());
 //! ```
-use super::{Token, TokenFilter, TokenStream};
+use super::{BoxTokenStream, Token, TokenFilter, TokenStream};
 
 /// `TokenFilter` that removes all tokens that contain non
 /// ascii alphanumeric characters.
 #[derive(Clone)]
 pub struct AlphaNumOnlyFilter;
 
-pub struct AlphaNumOnlyFilterStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
-    tail: TailTokenStream,
+pub struct AlphaNumOnlyFilterStream<'a> {
+    tail: BoxTokenStream<'a>,
 }
 
-impl<TailTokenStream> AlphaNumOnlyFilterStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
+impl<'a> AlphaNumOnlyFilterStream<'a> {
     fn predicate(&self, token: &Token) -> bool {
         token.text.chars().all(|c| c.is_ascii_alphanumeric())
     }
+}
 
-    fn wrap(tail: TailTokenStream) -> AlphaNumOnlyFilterStream<TailTokenStream> {
-        AlphaNumOnlyFilterStream { tail }
+impl TokenFilter for AlphaNumOnlyFilter {
+    fn transform<'a>(&self, token_stream: BoxTokenStream<'a>) -> BoxTokenStream<'a> {
+        BoxTokenStream::from(AlphaNumOnlyFilterStream { tail: token_stream })
     }
 }
 
-impl<TailTokenStream> TokenFilter<TailTokenStream> for AlphaNumOnlyFilter
-where
-    TailTokenStream: TokenStream,
-{
-    type ResultTokenStream = AlphaNumOnlyFilterStream<TailTokenStream>;
-
-    fn transform(&self, token_stream: TailTokenStream) -> Self::ResultTokenStream {
-        AlphaNumOnlyFilterStream::wrap(token_stream)
-    }
-}
-
-impl<TailTokenStream> TokenStream for AlphaNumOnlyFilterStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
-    fn token(&self) -> &Token {
-        self.tail.token()
-    }
-
-    fn token_mut(&mut self) -> &mut Token {
-        self.tail.token_mut()
-    }
-
+impl<'a> TokenStream for AlphaNumOnlyFilterStream<'a> {
     fn advance(&mut self) -> bool {
         while self.tail.advance() {
             if self.predicate(self.tail.token()) {
@@ -78,4 +52,12 @@ where
 
         false
     }
+
+    fn token(&self) -> &Token {
+        self.tail.token()
+    }
+
+    fn token_mut(&mut self) -> &mut Token {
+        self.tail.token_mut()
+    }
 }

src/tokenizer/ascii_folding_filter.rs

@@ -1,4 +1,4 @@
-use super::{Token, TokenFilter, TokenStream};
+use super::{BoxTokenStream, Token, TokenFilter, TokenStream};
 use std::mem;
 
 /// This class converts alphabetic, numeric, and symbolic Unicode characters
@@ -7,26 +7,21 @@ use std::mem;
 #[derive(Clone)]
 pub struct AsciiFoldingFilter;
 
-impl<TailTokenStream> TokenFilter<TailTokenStream> for AsciiFoldingFilter
-where
-    TailTokenStream: TokenStream,
-{
-    type ResultTokenStream = AsciiFoldingFilterTokenStream<TailTokenStream>;
-
-    fn transform(&self, token_stream: TailTokenStream) -> Self::ResultTokenStream {
-        AsciiFoldingFilterTokenStream::wrap(token_stream)
+impl TokenFilter for AsciiFoldingFilter {
+    fn transform<'a>(&self, token_stream: BoxTokenStream<'a>) -> BoxTokenStream<'a> {
+        From::from(AsciiFoldingFilterTokenStream {
+            tail: token_stream,
+            buffer: String::with_capacity(100),
+        })
     }
 }
 
-pub struct AsciiFoldingFilterTokenStream<TailTokenStream> {
+pub struct AsciiFoldingFilterTokenStream<'a> {
     buffer: String,
-    tail: TailTokenStream,
+    tail: BoxTokenStream<'a>,
 }
 
-impl<TailTokenStream> TokenStream for AsciiFoldingFilterTokenStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
+impl<'a> TokenStream for AsciiFoldingFilterTokenStream<'a> {
     fn advance(&mut self) -> bool {
         if !self.tail.advance() {
             return false;
@@ -48,18 +43,6 @@ where
     }
 }
 
-impl<TailTokenStream> AsciiFoldingFilterTokenStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
-    fn wrap(tail: TailTokenStream) -> AsciiFoldingFilterTokenStream<TailTokenStream> {
-        AsciiFoldingFilterTokenStream {
-            tail,
-            buffer: String::with_capacity(100),
-        }
-    }
-}
-
 // Returns a string that represents the ascii folded version of
 // the character. If the `char` does not require ascii folding
 // (e.g. simple ASCII chars like `A`) or if the `char`
@@ -1561,8 +1544,7 @@ mod tests {
     use crate::tokenizer::AsciiFoldingFilter;
     use crate::tokenizer::RawTokenizer;
     use crate::tokenizer::SimpleTokenizer;
-    use crate::tokenizer::TokenStream;
-    use crate::tokenizer::Tokenizer;
+    use crate::tokenizer::TextAnalyzer;
     use std::iter;
 
     #[test]
@@ -1579,7 +1561,7 @@ mod tests {
 
     fn folding_helper(text: &str) -> Vec<String> {
         let mut tokens = Vec::new();
-        SimpleTokenizer
+        TextAnalyzer::from(SimpleTokenizer)
             .filter(AsciiFoldingFilter)
             .token_stream(text)
             .process(&mut |token| {
@@ -1589,7 +1571,9 @@ mod tests {
     }
 
     fn folding_using_raw_tokenizer_helper(text: &str) -> String {
-        let mut token_stream = RawTokenizer.filter(AsciiFoldingFilter).token_stream(text);
+        let mut token_stream = TextAnalyzer::from(RawTokenizer)
+            .filter(AsciiFoldingFilter)
+            .token_stream(text);
         token_stream.advance();
         token_stream.token().text.clone()
     }

src/tokenizer/facet_tokenizer.rs

@@ -1,4 +1,4 @@
-use super::{Token, TokenStream, Tokenizer};
+use super::{BoxTokenStream, Token, TokenStream, Tokenizer};
 use crate::schema::FACET_SEP_BYTE;
 
 /// The `FacetTokenizer` process a `Facet` binary representation
@@ -25,15 +25,14 @@ pub struct FacetTokenStream<'a> {
     token: Token,
 }
 
-impl<'a> Tokenizer<'a> for FacetTokenizer {
-    type TokenStreamImpl = FacetTokenStream<'a>;
-
-    fn token_stream(&self, text: &'a str) -> Self::TokenStreamImpl {
+impl Tokenizer for FacetTokenizer {
+    fn token_stream<'a>(&self, text: &'a str) -> BoxTokenStream<'a> {
         FacetTokenStream {
             text,
             state: State::RootFacetNotEmitted, //< pos is the first char that has not been processed yet.
             token: Token::default(),
         }
+        .into()
     }
 }
 
@@ -84,7 +83,7 @@ mod tests {
 
     use super::FacetTokenizer;
     use crate::schema::Facet;
-    use crate::tokenizer::{Token, TokenStream, Tokenizer};
+    use crate::tokenizer::{Token, Tokenizer};
 
     #[test]
     fn test_facet_tokenizer() {

src/tokenizer/lower_caser.rs

@@ -1,24 +1,23 @@
 use super::{Token, TokenFilter, TokenStream};
+use crate::tokenizer::BoxTokenStream;
 use std::mem;
 
+impl TokenFilter for LowerCaser {
+    fn transform<'a>(&self, token_stream: BoxTokenStream<'a>) -> BoxTokenStream<'a> {
+        BoxTokenStream::from(LowerCaserTokenStream {
+            tail: token_stream,
+            buffer: String::with_capacity(100),
+        })
+    }
+}
+
 /// Token filter that lowercase terms.
 #[derive(Clone)]
 pub struct LowerCaser;
 
-impl<TailTokenStream> TokenFilter<TailTokenStream> for LowerCaser
-where
-    TailTokenStream: TokenStream,
-{
-    type ResultTokenStream = LowerCaserTokenStream<TailTokenStream>;
-
-    fn transform(&self, token_stream: TailTokenStream) -> Self::ResultTokenStream {
-        LowerCaserTokenStream::wrap(token_stream)
-    }
-}
-
-pub struct LowerCaserTokenStream<TailTokenStream> {
+pub struct LowerCaserTokenStream<'a> {
     buffer: String,
-    tail: TailTokenStream,
+    tail: BoxTokenStream<'a>,
 }
 
 // writes a lowercased version of text into output.
@@ -31,18 +30,7 @@ fn to_lowercase_unicode(text: &mut String, output: &mut String) {
     }
 }
 
-impl<TailTokenStream> TokenStream for LowerCaserTokenStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
-    fn token(&self) -> &Token {
-        self.tail.token()
-    }
-
-    fn token_mut(&mut self) -> &mut Token {
-        self.tail.token_mut()
-    }
-
+impl<'a> TokenStream for LowerCaserTokenStream<'a> {
     fn advance(&mut self) -> bool {
         if !self.tail.advance() {
             return false;
@@ -56,26 +44,19 @@ where
         }
         true
     }
-}
 
-impl<TailTokenStream> LowerCaserTokenStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
-    fn wrap(tail: TailTokenStream) -> LowerCaserTokenStream<TailTokenStream> {
-        LowerCaserTokenStream {
-            tail,
-            buffer: String::with_capacity(100),
-        }
+    fn token(&self) -> &Token {
+        self.tail.token()
+    }
+
+    fn token_mut(&mut self) -> &mut Token {
+        self.tail.token_mut()
     }
 }
 
 #[cfg(test)]
 mod tests {
-    use crate::tokenizer::LowerCaser;
-    use crate::tokenizer::SimpleTokenizer;
-    use crate::tokenizer::TokenStream;
-    use crate::tokenizer::Tokenizer;
+    use crate::tokenizer::{LowerCaser, SimpleTokenizer, TextAnalyzer};
 
     #[test]
     fn test_to_lower_case() {
@@ -87,7 +68,9 @@ mod tests {
 
     fn lowercase_helper(text: &str) -> Vec<String> {
         let mut tokens = vec![];
-        let mut token_stream = SimpleTokenizer.filter(LowerCaser).token_stream(text);
+        let mut token_stream = TextAnalyzer::from(SimpleTokenizer)
+            .filter(LowerCaser)
+            .token_stream(text);
         while token_stream.advance() {
             let token_text = token_stream.token().text.clone();
             tokens.push(token_text);

src/tokenizer/mod.rs

@@ -64,7 +64,7 @@
 //! ```rust
 //! use tantivy::tokenizer::*;
 //!
-//! let en_stem = SimpleTokenizer
+//! let en_stem = TextAnalyzer::from(SimpleTokenizer)
 //!     .filter(RemoveLongFilter::limit(40))
 //!     .filter(LowerCaser)
 //!     .filter(Stemmer::new(Language::English));
@@ -109,7 +109,7 @@
 //! let index = Index::create_in_ram(schema);
 //!
 //! // We need to register our tokenizer :
-//! let custom_en_tokenizer = SimpleTokenizer
+//! let custom_en_tokenizer = TextAnalyzer::from(SimpleTokenizer)
 //!     .filter(RemoveLongFilter::limit(40))
 //!     .filter(LowerCaser);
 //! index
@@ -143,10 +143,11 @@ pub use self::simple_tokenizer::SimpleTokenizer;
 pub use self::stemmer::{Language, Stemmer};
 pub use self::stop_word_filter::StopWordFilter;
 pub(crate) use self::token_stream_chain::TokenStreamChain;
-pub use self::tokenizer::BoxedTokenizer;
 
 pub use self::tokenized_string::{PreTokenizedStream, PreTokenizedString};
-pub use self::tokenizer::{Token, TokenFilter, TokenStream, Tokenizer};
+pub use self::tokenizer::{
+    BoxTokenFilter, BoxTokenStream, TextAnalyzer, Token, TokenFilter, TokenStream, Tokenizer,
+};
 
 pub use self::tokenizer_manager::TokenizerManager;
 
@@ -160,9 +161,9 @@ pub const MAX_TOKEN_LEN: usize = u16::max_value() as usize - 4;
 #[cfg(test)]
 pub mod tests {
     use super::{
-        Language, LowerCaser, RemoveLongFilter, SimpleTokenizer, Stemmer, Token, Tokenizer,
-        TokenizerManager,
+        Language, LowerCaser, RemoveLongFilter, SimpleTokenizer, Stemmer, Token, TokenizerManager,
     };
+    use crate::tokenizer::TextAnalyzer;
 
     /// This is a function that can be used in tests and doc tests
     /// to assert a token's correctness.
@@ -229,7 +230,7 @@ pub mod tests {
         let tokenizer_manager = TokenizerManager::default();
         tokenizer_manager.register(
             "el_stem",
-            SimpleTokenizer
+            TextAnalyzer::from(SimpleTokenizer)
                 .filter(RemoveLongFilter::limit(40))
                 .filter(LowerCaser)
                 .filter(Stemmer::new(Language::Greek)),

src/tokenizer/ngram_tokenizer.rs

@@ -1,4 +1,5 @@
 use super::{Token, TokenStream, Tokenizer};
+use crate::tokenizer::BoxTokenStream;
 
 /// Tokenize the text by splitting words into n-grams of the given size(s)
 ///
@@ -129,11 +130,9 @@ pub struct NgramTokenStream<'a> {
     token: Token,
 }
 
-impl<'a> Tokenizer<'a> for NgramTokenizer {
-    type TokenStreamImpl = NgramTokenStream<'a>;
-
-    fn token_stream(&self, text: &'a str) -> Self::TokenStreamImpl {
-        NgramTokenStream {
+impl Tokenizer for NgramTokenizer {
+    fn token_stream<'a>(&self, text: &'a str) -> BoxTokenStream<'a> {
+        From::from(NgramTokenStream {
             ngram_charidx_iterator: StutteringIterator::new(
                 CodepointFrontiers::for_str(text),
                 self.min_gram,
@@ -142,7 +141,7 @@ impl<'a> Tokenizer<'a> for NgramTokenizer {
             prefix_only: self.prefix_only,
             text,
             token: Token::default(),
-        }
+        })
     }
 }
 
@@ -308,10 +307,10 @@ mod tests {
     use super::NgramTokenizer;
     use super::StutteringIterator;
     use crate::tokenizer::tests::assert_token;
-    use crate::tokenizer::tokenizer::{TokenStream, Tokenizer};
-    use crate::tokenizer::Token;
+    use crate::tokenizer::tokenizer::Tokenizer;
+    use crate::tokenizer::{BoxTokenStream, Token};
 
-    fn test_helper<T: TokenStream>(mut tokenizer: T) -> Vec<Token> {
+    fn test_helper(mut tokenizer: BoxTokenStream) -> Vec<Token> {
         let mut tokens: Vec<Token> = vec![];
         tokenizer.process(&mut |token: &Token| tokens.push(token.clone()));
         tokens

src/tokenizer/raw_tokenizer.rs

@@ -1,4 +1,5 @@
 use super::{Token, TokenStream, Tokenizer};
+use crate::tokenizer::BoxTokenStream;
 
 /// For each value of the field, emit a single unprocessed token.
 #[derive(Clone)]
@@ -9,10 +10,8 @@ pub struct RawTokenStream {
     has_token: bool,
 }
 
-impl<'a> Tokenizer<'a> for RawTokenizer {
-    type TokenStreamImpl = RawTokenStream;
-
-    fn token_stream(&self, text: &'a str) -> Self::TokenStreamImpl {
+impl Tokenizer for RawTokenizer {
+    fn token_stream<'a>(&self, text: &'a str) -> BoxTokenStream<'a> {
         let token = Token {
             offset_from: 0,
             offset_to: text.len(),
@@ -24,6 +23,7 @@ impl<'a> Tokenizer<'a> for RawTokenizer {
             token,
             has_token: true,
         }
+        .into()
     }
 }
 

src/tokenizer/remove_long.rs

@@ -2,7 +2,7 @@
 //! ```rust
 //! use tantivy::tokenizer::*;
 //!
-//! let tokenizer = SimpleTokenizer
+//! let tokenizer = TextAnalyzer::from(SimpleTokenizer)
 //!   .filter(RemoveLongFilter::limit(5));
 //!
 //! let mut stream = tokenizer.token_stream("toolong nice");
@@ -13,6 +13,7 @@
 //! ```
 //!
 use super::{Token, TokenFilter, TokenStream};
+use crate::tokenizer::BoxTokenStream;
 
 /// `RemoveLongFilter` removes tokens that are longer
 /// than a given number of bytes (in UTF-8 representation).
@@ -31,56 +32,27 @@ impl RemoveLongFilter {
     }
 }
 
-impl<TailTokenStream> RemoveLongFilterStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
+impl<'a> RemoveLongFilterStream<'a> {
     fn predicate(&self, token: &Token) -> bool {
         token.text.len() < self.token_length_limit
     }
+}
 
-    fn wrap(
-        token_length_limit: usize,
-        tail: TailTokenStream,
-    ) -> RemoveLongFilterStream<TailTokenStream> {
-        RemoveLongFilterStream {
-            token_length_limit,
-            tail,
-        }
+impl TokenFilter for RemoveLongFilter {
+    fn transform<'a>(&self, token_stream: BoxTokenStream<'a>) -> BoxTokenStream<'a> {
+        BoxTokenStream::from(RemoveLongFilterStream {
+            token_length_limit: self.length_limit,
+            tail: token_stream,
+        })
     }
 }
 
-impl<TailTokenStream> TokenFilter<TailTokenStream> for RemoveLongFilter
-where
-    TailTokenStream: TokenStream,
-{
-    type ResultTokenStream = RemoveLongFilterStream<TailTokenStream>;
-
-    fn transform(&self, token_stream: TailTokenStream) -> Self::ResultTokenStream {
-        RemoveLongFilterStream::wrap(self.length_limit, token_stream)
-    }
-}
-
-pub struct RemoveLongFilterStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
+pub struct RemoveLongFilterStream<'a> {
     token_length_limit: usize,
-    tail: TailTokenStream,
+    tail: BoxTokenStream<'a>,
 }
 
-impl<TailTokenStream> TokenStream for RemoveLongFilterStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
-    fn token(&self) -> &Token {
-        self.tail.token()
-    }
-
-    fn token_mut(&mut self) -> &mut Token {
-        self.tail.token_mut()
-    }
-
+impl<'a> TokenStream for RemoveLongFilterStream<'a> {
     fn advance(&mut self) -> bool {
         while self.tail.advance() {
             if self.predicate(self.tail.token()) {
@@ -89,4 +61,12 @@ where
         }
         false
     }
+
+    fn token(&self) -> &Token {
+        self.tail.token()
+    }
+
+    fn token_mut(&mut self) -> &mut Token {
+        self.tail.token_mut()
+    }
 }

src/tokenizer/simple_tokenizer.rs

@@ -1,3 +1,4 @@
+use super::BoxTokenStream;
 use super::{Token, TokenStream, Tokenizer};
 use std::str::CharIndices;
 
@@ -11,15 +12,13 @@ pub struct SimpleTokenStream<'a> {
     token: Token,
 }
 
-impl<'a> Tokenizer<'a> for SimpleTokenizer {
-    type TokenStreamImpl = SimpleTokenStream<'a>;
-
-    fn token_stream(&self, text: &'a str) -> Self::TokenStreamImpl {
-        SimpleTokenStream {
+impl Tokenizer for SimpleTokenizer {
+    fn token_stream<'a>(&self, text: &'a str) -> BoxTokenStream<'a> {
+        BoxTokenStream::from(SimpleTokenStream {
             text,
             chars: text.char_indices(),
             token: Token::default(),
-        }
+        })
     }
 }
 

src/tokenizer/stemmer.rs

@@ -1,4 +1,5 @@
 use super::{Token, TokenFilter, TokenStream};
+use crate::tokenizer::BoxTokenStream;
 use rust_stemmers::{self, Algorithm};
 
 /// Available stemmer languages.
@@ -75,38 +76,22 @@ impl Default for Stemmer {
     }
 }
 
-impl<TailTokenStream> TokenFilter<TailTokenStream> for Stemmer
-where
-    TailTokenStream: TokenStream,
-{
-    type ResultTokenStream = StemmerTokenStream<TailTokenStream>;
-
-    fn transform(&self, token_stream: TailTokenStream) -> Self::ResultTokenStream {
+impl TokenFilter for Stemmer {
+    fn transform<'a>(&self, token_stream: BoxTokenStream<'a>) -> BoxTokenStream<'a> {
         let inner_stemmer = rust_stemmers::Stemmer::create(self.stemmer_algorithm);
-        StemmerTokenStream::wrap(inner_stemmer, token_stream)
+        BoxTokenStream::from(StemmerTokenStream {
+            tail: token_stream,
+            stemmer: inner_stemmer,
+        })
     }
 }
 
-pub struct StemmerTokenStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
-    tail: TailTokenStream,
+pub struct StemmerTokenStream<'a> {
+    tail: BoxTokenStream<'a>,
     stemmer: rust_stemmers::Stemmer,
 }
 
-impl<TailTokenStream> TokenStream for StemmerTokenStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
-    fn token(&self) -> &Token {
-        self.tail.token()
-    }
-
-    fn token_mut(&mut self) -> &mut Token {
-        self.tail.token_mut()
-    }
-
+impl<'a> TokenStream for StemmerTokenStream<'a> {
     fn advance(&mut self) -> bool {
         if !self.tail.advance() {
             return false;
@@ -117,16 +102,12 @@ where
         self.token_mut().text.push_str(&stemmed_str);
         true
     }
-}
 
-impl<TailTokenStream> StemmerTokenStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
-    fn wrap(
-        stemmer: rust_stemmers::Stemmer,
-        tail: TailTokenStream,
-    ) -> StemmerTokenStream<TailTokenStream> {
-        StemmerTokenStream { tail, stemmer }
+    fn token(&self) -> &Token {
+        self.tail.token()
+    }
+
+    fn token_mut(&mut self) -> &mut Token {
+        self.tail.token_mut()
     }
 }

src/tokenizer/stop_word_filter.rs

@@ -2,7 +2,7 @@
 //! ```rust
 //! use tantivy::tokenizer::*;
 //!
-//! let tokenizer = SimpleTokenizer
+//! let tokenizer = TextAnalyzer::from(SimpleTokenizer)
 //!   .filter(StopWordFilter::remove(vec!["the".to_string(), "is".to_string()]));
 //!
 //! let mut stream = tokenizer.token_stream("the fox is crafty");
@@ -11,6 +11,7 @@
 //! assert!(stream.next().is_none());
 //! ```
 use super::{Token, TokenFilter, TokenStream};
+use crate::tokenizer::BoxTokenStream;
 use fnv::FnvHasher;
 use std::collections::HashSet;
 use std::hash::BuildHasherDefault;
@@ -48,53 +49,27 @@ impl StopWordFilter {
     }
 }
 
-pub struct StopWordFilterStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
+pub struct StopWordFilterStream<'a> {
     words: StopWordHashSet,
-    tail: TailTokenStream,
+    tail: BoxTokenStream<'a>,
 }
 
-impl<TailTokenStream> TokenFilter<TailTokenStream> for StopWordFilter
-where
-    TailTokenStream: TokenStream,
-{
-    type ResultTokenStream = StopWordFilterStream<TailTokenStream>;
-
-    fn transform(&self, token_stream: TailTokenStream) -> Self::ResultTokenStream {
-        StopWordFilterStream::wrap(self.words.clone(), token_stream)
+impl TokenFilter for StopWordFilter {
+    fn transform<'a>(&self, token_stream: BoxTokenStream<'a>) -> BoxTokenStream<'a> {
+        BoxTokenStream::from(StopWordFilterStream {
+            words: self.words.clone(),
+            tail: token_stream,
+        })
     }
 }
 
-impl<TailTokenStream> StopWordFilterStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
+impl<'a> StopWordFilterStream<'a> {
     fn predicate(&self, token: &Token) -> bool {
         !self.words.contains(&token.text)
     }
-
-    fn wrap(
-        words: StopWordHashSet,
-        tail: TailTokenStream,
-    ) -> StopWordFilterStream<TailTokenStream> {
-        StopWordFilterStream { words, tail }
-    }
 }
 
-impl<TailTokenStream> TokenStream for StopWordFilterStream<TailTokenStream>
-where
-    TailTokenStream: TokenStream,
-{
-    fn token(&self) -> &Token {
-        self.tail.token()
-    }
-
-    fn token_mut(&mut self) -> &mut Token {
-        self.tail.token_mut()
-    }
-
+impl<'a> TokenStream for StopWordFilterStream<'a> {
     fn advance(&mut self) -> bool {
         while self.tail.advance() {
             if self.predicate(self.tail.token()) {
@@ -103,6 +78,14 @@ where
         }
         false
     }
+
+    fn token(&self) -> &Token {
+        self.tail.token()
+    }
+
+    fn token_mut(&mut self) -> &mut Token {
+        self.tail.token_mut()
+    }
 }
 
 impl Default for StopWordFilter {

src/tokenizer/token_stream_chain.rs

@@ -1,23 +1,21 @@
-use crate::tokenizer::{Token, TokenStream};
+use crate::tokenizer::{BoxTokenStream, Token, TokenStream};
+use std::ops::DerefMut;
 
 const POSITION_GAP: usize = 2;
 
-pub(crate) struct TokenStreamChain<TTokenStream: TokenStream> {
+pub(crate) struct TokenStreamChain<'a> {
     offsets: Vec<usize>,
-    token_streams: Vec<TTokenStream>,
+    token_streams: Vec<BoxTokenStream<'a>>,
     position_shift: usize,
     stream_idx: usize,
     token: Token,
 }
 
-impl<'a, TTokenStream> TokenStreamChain<TTokenStream>
-where
-    TTokenStream: TokenStream,
-{
+impl<'a> TokenStreamChain<'a> {
     pub fn new(
         offsets: Vec<usize>,
-        token_streams: Vec<TTokenStream>,
-    ) -> TokenStreamChain<TTokenStream> {
+        token_streams: Vec<BoxTokenStream<'a>>,
+    ) -> TokenStreamChain<'a> {
         TokenStreamChain {
             offsets,
             stream_idx: 0,
@@ -28,13 +26,10 @@ where
     }
 }
 
-impl<'a, TTokenStream> TokenStream for TokenStreamChain<TTokenStream>
-where
-    TTokenStream: TokenStream,
-{
+impl<'a> TokenStream for TokenStreamChain<'a> {
     fn advance(&mut self) -> bool {
         while self.stream_idx < self.token_streams.len() {
-            let token_stream = &mut self.token_streams[self.stream_idx];
+            let token_stream = self.token_streams[self.stream_idx].deref_mut();
             if token_stream.advance() {
                 let token = token_stream.token();
                 let offset_offset = self.offsets[self.stream_idx];

src/tokenizer/tokenized_string.rs

@@ -1,4 +1,4 @@
-use crate::tokenizer::{Token, TokenStream, TokenStreamChain};
+use crate::tokenizer::{BoxTokenStream, Token, TokenStream, TokenStreamChain};
 use std::cmp::Ordering;
 
 /// Struct representing pre-tokenized text
@@ -41,9 +41,9 @@ impl PreTokenizedStream {
     /// Creates a TokenStream from PreTokenizedString array
     pub fn chain_tokenized_strings<'a>(
         tok_strings: &'a [&'a PreTokenizedString],
-    ) -> Box<dyn TokenStream + 'a> {
+    ) -> BoxTokenStream {
         if tok_strings.len() == 1 {
-            Box::new(PreTokenizedStream::from((*tok_strings[0]).clone()))
+            PreTokenizedStream::from((*tok_strings[0]).clone()).into()
         } else {
             let mut offsets = vec![];
             let mut total_offset = 0;
@@ -53,11 +53,12 @@ impl PreTokenizedStream {
                     total_offset += last_token.offset_to;
                 }
             }
-            let token_streams: Vec<_> = tok_strings
+            // TODO remove the string cloning.
+            let token_streams: Vec<BoxTokenStream<'static>> = tok_strings
                 .iter()
-                .map(|tok_string| PreTokenizedStream::from((*tok_string).clone()))
+                .map(|&tok_string| PreTokenizedStream::from((*tok_string).clone()).into())
                 .collect();
-            Box::new(TokenStreamChain::new(offsets, token_streams))
+            TokenStreamChain::new(offsets, token_streams).into()
         }
     }
 }

src/tokenizer/tokenizer.rs

@@ -2,6 +2,7 @@ use crate::tokenizer::TokenStreamChain;
 /// The tokenizer module contains all of the tools used to process
 /// text in `tantivy`.
 use std::borrow::{Borrow, BorrowMut};
+use std::ops::{Deref, DerefMut};
 
 /// Token
 #[derive(Debug, Clone, Serialize, Deserialize, Eq, PartialEq)]
@@ -33,20 +34,31 @@ impl Default for Token {
     }
 }
 
-/// `Tokenizer` are in charge of splitting text into a stream of token
-/// before indexing.
+/// `TextAnalyzer` tokenizes an input text into tokens and modifies the resulting `TokenStream`.
 ///
-/// See the [module documentation](./index.html) for more detail.
-///
-/// # Warning
-///
-/// This API may change to use associated types.
-pub trait Tokenizer<'a>: Sized + Clone {
-    /// Type associated to the resulting tokenstream tokenstream.
-    type TokenStreamImpl: TokenStream;
+/// It simply wraps a `Tokenizer` and a list of `TokenFilter` that are applied sequentially.
+pub struct TextAnalyzer {
+    tokenizer: Box<dyn Tokenizer>,
+    token_filters: Vec<BoxTokenFilter>,
+}
 
-    /// Creates a token stream for a given `str`.
-    fn token_stream(&self, text: &'a str) -> Self::TokenStreamImpl;
+impl<T: Tokenizer> From<T> for TextAnalyzer {
+    fn from(tokenizer: T) -> Self {
+        TextAnalyzer::new(tokenizer, Vec::new())
+    }
+}
+
+impl TextAnalyzer {
+    /// Creates a new `TextAnalyzer` given a tokenizer and a vector of `BoxTokenFilter`.
+    ///
+    /// When creating a `TextAnalyzer` from a `Tokenizer` alone, prefer using
+    /// `TextAnalyzer::from(tokenizer)`.
+    pub fn new<T: Tokenizer>(tokenizer: T, token_filters: Vec<BoxTokenFilter>) -> TextAnalyzer {
+        TextAnalyzer {
+            tokenizer: Box::new(tokenizer),
+            token_filters,
+        }
+    }
 
     /// Appends a token filter to the current tokenizer.
     ///
@@ -58,90 +70,26 @@ pub trait Tokenizer<'a>: Sized + Clone {
     /// ```rust
     /// use tantivy::tokenizer::*;
     ///
-    /// let en_stem = SimpleTokenizer
+    /// let en_stem = TextAnalyzer::from(SimpleTokenizer)
     ///     .filter(RemoveLongFilter::limit(40))
     ///     .filter(LowerCaser)
     ///     .filter(Stemmer::default());
     /// ```
     ///
-    fn filter<NewFilter>(self, new_filter: NewFilter) -> ChainTokenizer<NewFilter, Self>
-    where
-        NewFilter: TokenFilter<<Self as Tokenizer<'a>>::TokenStreamImpl>,
-    {
-        ChainTokenizer {
-            head: new_filter,
-            tail: self,
-        }
-    }
-}
-
-/// A boxed tokenizer
-trait BoxedTokenizerTrait: Send + Sync {
-    /// Tokenize a `&str`
-    fn token_stream<'a>(&self, text: &'a str) -> Box<dyn TokenStream + 'a>;
-
-    /// Tokenize an array`&str`
-    ///
-    /// The resulting `TokenStream` is equivalent to what would be obtained if the &str were
-    /// one concatenated `&str`, with an artificial position gap of `2` between the different fields
-    /// to prevent accidental `PhraseQuery` to match accross two terms.
-    fn token_stream_texts<'b>(&self, texts: &'b [&'b str]) -> Box<dyn TokenStream + 'b>;
-
-    /// Return a boxed clone of the tokenizer
-    fn boxed_clone(&self) -> BoxedTokenizer;
-}
-
-/// A boxed tokenizer
-pub struct BoxedTokenizer(Box<dyn BoxedTokenizerTrait>);
-
-impl<T> From<T> for BoxedTokenizer
-where
-    T: 'static + Send + Sync + for<'a> Tokenizer<'a>,
-{
-    fn from(tokenizer: T) -> BoxedTokenizer {
-        BoxedTokenizer(Box::new(BoxableTokenizer(tokenizer)))
-    }
-}
-
-impl BoxedTokenizer {
-    /// Tokenize a `&str`
-    pub fn token_stream<'a>(&self, text: &'a str) -> Box<dyn TokenStream + 'a> {
-        self.0.token_stream(text)
+    pub fn filter<F: Into<BoxTokenFilter>>(mut self, token_filter: F) -> Self {
+        self.token_filters.push(token_filter.into());
+        self
     }
 
     /// Tokenize an array`&str`
     ///
-    /// The resulting `TokenStream` is equivalent to what would be obtained if the &str were
+    /// The resulting `BoxTokenStream` is equivalent to what would be obtained if the &str were
     /// one concatenated `&str`, with an artificial position gap of `2` between the different fields
     /// to prevent accidental `PhraseQuery` to match accross two terms.
-    pub fn token_stream_texts<'b>(&self, texts: &'b [&'b str]) -> Box<dyn TokenStream + 'b> {
-        self.0.token_stream_texts(texts)
-    }
-}
-
-impl Clone for BoxedTokenizer {
-    fn clone(&self) -> BoxedTokenizer {
-        self.0.boxed_clone()
-    }
-}
-
-#[derive(Clone)]
-struct BoxableTokenizer<A>(A)
-where
-    A: for<'a> Tokenizer<'a> + Send + Sync;
-
-impl<A> BoxedTokenizerTrait for BoxableTokenizer<A>
-where
-    A: 'static + Send + Sync + for<'a> Tokenizer<'a>,
-{
-    fn token_stream<'a>(&self, text: &'a str) -> Box<dyn TokenStream + 'a> {
-        Box::new(self.0.token_stream(text))
-    }
-
-    fn token_stream_texts<'b>(&self, texts: &'b [&'b str]) -> Box<dyn TokenStream + 'b> {
+    pub fn token_stream_texts<'a>(&self, texts: &'a [&'a str]) -> BoxTokenStream<'a> {
         assert!(!texts.is_empty());
         if texts.len() == 1 {
-            Box::new(self.0.token_stream(texts[0]))
+            self.token_stream(texts[0])
         } else {
             let mut offsets = vec![];
             let mut total_offset = 0;
@@ -149,34 +97,124 @@ where
                 offsets.push(total_offset);
                 total_offset += text.len();
             }
-            let token_streams: Vec<_> =
-                texts.iter().map(|text| self.0.token_stream(text)).collect();
-            Box::new(TokenStreamChain::new(offsets, token_streams))
+            let token_streams: Vec<BoxTokenStream<'a>> = texts
+                .iter()
+                .cloned()
+                .map(|text| self.token_stream(text))
+                .collect();
+            From::from(TokenStreamChain::new(offsets, token_streams))
         }
     }
 
-    fn boxed_clone(&self) -> BoxedTokenizer {
-        self.0.clone().into()
+    /// Creates a token stream for a given `str`.
+    pub fn token_stream<'a>(&self, text: &'a str) -> BoxTokenStream<'a> {
+        let mut token_stream = self.tokenizer.token_stream(text);
+        for token_filter in &self.token_filters {
+            token_stream = token_filter.transform(token_stream);
+        }
+        token_stream
     }
 }
 
-impl<'b> TokenStream for Box<dyn TokenStream + 'b> {
+impl Clone for TextAnalyzer {
+    fn clone(&self) -> Self {
+        TextAnalyzer {
+            tokenizer: self.tokenizer.box_clone(),
+            token_filters: self
+                .token_filters
+                .iter()
+                .map(|token_filter| token_filter.box_clone())
+                .collect(),
+        }
+    }
+}
+
+/// `Tokenizer` are in charge of splitting text into a stream of token
+/// before indexing.
+///
+/// See the [module documentation](./index.html) for more detail.
+///
+/// # Warning
+///
+/// This API may change to use associated types.
+pub trait Tokenizer: 'static + Send + Sync + TokenizerClone {
+    /// Creates a token stream for a given `str`.
+    fn token_stream<'a>(&self, text: &'a str) -> BoxTokenStream<'a>;
+}
+
+pub trait TokenizerClone {
+    fn box_clone(&self) -> Box<dyn Tokenizer>;
+}
+
+impl<T: Tokenizer + Clone> TokenizerClone for T {
+    fn box_clone(&self) -> Box<dyn Tokenizer> {
+        Box::new(self.clone())
+    }
+}
+
+impl<'a> TokenStream for Box<dyn TokenStream + 'a> {
     fn advance(&mut self) -> bool {
         let token_stream: &mut dyn TokenStream = self.borrow_mut();
         token_stream.advance()
     }
 
-    fn token(&self) -> &Token {
-        let token_stream: &dyn TokenStream = self.borrow();
+    fn token<'b>(&'b self) -> &'b Token {
+        let token_stream: &'b (dyn TokenStream + 'a) = self.borrow();
         token_stream.token()
     }
 
-    fn token_mut(&mut self) -> &mut Token {
-        let token_stream: &mut dyn TokenStream = self.borrow_mut();
+    fn token_mut<'b>(&'b mut self) -> &'b mut Token {
+        let token_stream: &'b mut (dyn TokenStream + 'a) = self.borrow_mut();
         token_stream.token_mut()
     }
 }
 
+/// Simple wrapper of `Box<dyn TokenStream + 'a>`.
+///
+/// See `TokenStream` for more information.
+pub struct BoxTokenStream<'a>(Box<dyn TokenStream + 'a>);
+
+impl<'a, T> From<T> for BoxTokenStream<'a>
+where
+    T: TokenStream + 'a,
+{
+    fn from(token_stream: T) -> BoxTokenStream<'a> {
+        BoxTokenStream(Box::new(token_stream))
+    }
+}
+
+impl<'a> Deref for BoxTokenStream<'a> {
+    type Target = dyn TokenStream + 'a;
+
+    fn deref(&self) -> &Self::Target {
+        &*self.0
+    }
+}
+impl<'a> DerefMut for BoxTokenStream<'a> {
+    fn deref_mut(&mut self) -> &mut Self::Target {
+        &mut *self.0
+    }
+}
+
+/// Simple wrapper of `Box<dyn TokenFilter + 'a>`.
+///
+/// See `TokenStream` for more information.
+pub struct BoxTokenFilter(Box<dyn TokenFilter>);
+
+impl Deref for BoxTokenFilter {
+    type Target = dyn TokenFilter;
+
+    fn deref(&self) -> &dyn TokenFilter {
+        &*self.0
+    }
+}
+
+impl<T: TokenFilter> From<T> for BoxTokenFilter {
+    fn from(tokenizer: T) -> BoxTokenFilter {
+        BoxTokenFilter(Box::new(tokenizer))
+    }
+}
+
 /// `TokenStream` is the result of the tokenization.
 ///
 /// It consists consumable stream of `Token`s.
@@ -186,7 +224,7 @@ impl<'b> TokenStream for Box<dyn TokenStream + 'b> {
 /// ```
 /// use tantivy::tokenizer::*;
 ///
-/// let tokenizer = SimpleTokenizer
+/// let tokenizer = TextAnalyzer::from(SimpleTokenizer)
 ///        .filter(RemoveLongFilter::limit(40))
 ///        .filter(LowerCaser);
 /// let mut token_stream = tokenizer.token_stream("Hello, happy tax payer");
@@ -225,7 +263,7 @@ pub trait TokenStream {
     /// ```
     /// use tantivy::tokenizer::*;
     ///
-    /// let tokenizer = SimpleTokenizer
+    /// let tokenizer = TextAnalyzer::from(SimpleTokenizer)
     ///       .filter(RemoveLongFilter::limit(40))
     ///       .filter(LowerCaser);
     /// let mut token_stream = tokenizer.token_stream("Hello, happy tax payer");
@@ -243,6 +281,8 @@ pub trait TokenStream {
 
     /// Helper function to consume the entire `TokenStream`
     /// and push the tokens to a sink function.
+    ///
+    /// Remove this.
     fn process(&mut self, sink: &mut dyn FnMut(&Token)) -> u32 {
         let mut num_tokens_pushed = 0u32;
         while self.advance() {
@@ -253,33 +293,20 @@ pub trait TokenStream {
     }
 }
 
-#[derive(Clone)]
-pub struct ChainTokenizer<HeadTokenFilterFactory, TailTokenizer> {
-    head: HeadTokenFilterFactory,
-    tail: TailTokenizer,
-}
-
-impl<'a, HeadTokenFilterFactory, TailTokenizer> Tokenizer<'a>
-    for ChainTokenizer<HeadTokenFilterFactory, TailTokenizer>
-where
-    HeadTokenFilterFactory: TokenFilter<TailTokenizer::TokenStreamImpl>,
-    TailTokenizer: Tokenizer<'a>,
-{
-    type TokenStreamImpl = HeadTokenFilterFactory::ResultTokenStream;
-
-    fn token_stream(&self, text: &'a str) -> Self::TokenStreamImpl {
-        let tail_token_stream = self.tail.token_stream(text);
-        self.head.transform(tail_token_stream)
-    }
+pub trait TokenFilterClone {
+    fn box_clone(&self) -> BoxTokenFilter;
 }
 
 /// Trait for the pluggable components of `Tokenizer`s.
-pub trait TokenFilter<TailTokenStream: TokenStream>: Clone {
-    /// The resulting `TokenStream` type.
-    type ResultTokenStream: TokenStream;
-
+pub trait TokenFilter: 'static + Send + Sync + TokenFilterClone {
     /// Wraps a token stream and returns the modified one.
-    fn transform(&self, token_stream: TailTokenStream) -> Self::ResultTokenStream;
+    fn transform<'a>(&self, token_stream: BoxTokenStream<'a>) -> BoxTokenStream<'a>;
+}
+
+impl<T: TokenFilter + Clone> TokenFilterClone for T {
+    fn box_clone(&self) -> BoxTokenFilter {
+        BoxTokenFilter::from(self.clone())
+    }
 }
 
 #[cfg(test)]

src/tokenizer/tokenizer_manager.rs

@@ -1,11 +1,10 @@
 use crate::tokenizer::stemmer::Language;
-use crate::tokenizer::BoxedTokenizer;
+use crate::tokenizer::tokenizer::TextAnalyzer;
 use crate::tokenizer::LowerCaser;
 use crate::tokenizer::RawTokenizer;
 use crate::tokenizer::RemoveLongFilter;
 use crate::tokenizer::SimpleTokenizer;
 use crate::tokenizer::Stemmer;
-use crate::tokenizer::Tokenizer;
 use std::collections::HashMap;
 use std::sync::{Arc, RwLock};
 
@@ -23,16 +22,16 @@ use std::sync::{Arc, RwLock};
 ///  search engine.
 #[derive(Clone)]
 pub struct TokenizerManager {
-    tokenizers: Arc<RwLock<HashMap<String, BoxedTokenizer>>>,
+    tokenizers: Arc<RwLock<HashMap<String, TextAnalyzer>>>,
 }
 
 impl TokenizerManager {
     /// Registers a new tokenizer associated with a given name.
-    pub fn register<A>(&self, tokenizer_name: &str, tokenizer: A)
+    pub fn register<T>(&self, tokenizer_name: &str, tokenizer: T)
     where
-        A: Into<BoxedTokenizer>,
+        TextAnalyzer: From<T>,
     {
-        let boxed_tokenizer = tokenizer.into();
+        let boxed_tokenizer: TextAnalyzer = TextAnalyzer::from(tokenizer);
         self.tokenizers
             .write()
             .expect("Acquiring the lock should never fail")
@@ -40,7 +39,7 @@ impl TokenizerManager {
     }
 
     /// Accessing a tokenizer given its name.
-    pub fn get(&self, tokenizer_name: &str) -> Option<BoxedTokenizer> {
+    pub fn get(&self, tokenizer_name: &str) -> Option<TextAnalyzer> {
         self.tokenizers
             .read()
             .expect("Acquiring the lock should never fail")
@@ -62,13 +61,13 @@ impl Default for TokenizerManager {
         manager.register("raw", RawTokenizer);
         manager.register(
             "default",
-            SimpleTokenizer
+            TextAnalyzer::from(SimpleTokenizer)
                 .filter(RemoveLongFilter::limit(40))
                 .filter(LowerCaser),
         );
         manager.register(
             "en_stem",
-            SimpleTokenizer
+            TextAnalyzer::from(SimpleTokenizer)
                 .filter(RemoveLongFilter::limit(40))
                 .filter(LowerCaser)
                 .filter(Stemmer::new(Language::English)),