extra commit

CHANGELOG.md

@@ -1,16 +1,3 @@
-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)

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "tantivy"
-version = "0.12.0"
+version = "0.11.3"
 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 = "0.2.1"
+tantivy-fst = "0.1"
 memmap = {version = "0.7", optional=true}
 lz4 = {version="1.20", optional=true}
 snap = {version="0.2"}
@@ -60,6 +60,7 @@ winapi = "0.3"
 rand = "0.7"
 maplit = "1"
 matches = "0.1.8"
+time = "0.1.42"
 
 [dev-dependencies.fail]
 version = "0.3"

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, Tokenizer};
+use tantivy::tokenizer::{PreTokenizedString, SimpleTokenizer, Token, TokenStream, 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 = TextAnalyzer::from(SimpleTokenizer)
+    let tokenizer = SimpleTokenizer
         .filter(LowerCaser)
         .filter(StopWordFilter::remove(vec![
             "the".to_string(),

src/collector/count_collector.rs

@@ -1,6 +1,7 @@
 use super::Collector;
 use crate::collector::SegmentCollector;
 use crate::DocId;
+use crate::Result;
 use crate::Score;
 use crate::SegmentLocalId;
 use crate::SegmentReader;
@@ -43,11 +44,7 @@ impl Collector for Count {
 
     type Child = SegmentCountCollector;
 
-    fn for_segment(
-        &self,
-        _: SegmentLocalId,
-        _: &SegmentReader,
-    ) -> crate::Result<SegmentCountCollector> {
+    fn for_segment(&self, _: SegmentLocalId, _: &SegmentReader) -> Result<SegmentCountCollector> {
         Ok(SegmentCountCollector::default())
     }
 
@@ -55,7 +52,7 @@ impl Collector for Count {
         false
     }
 
-    fn merge_fruits(&self, segment_counts: Vec<usize>) -> crate::Result<usize> {
+    fn merge_fruits(&self, segment_counts: Vec<usize>) -> Result<usize> {
         Ok(segment_counts.into_iter().sum())
     }
 }

src/collector/custom_score_top_collector.rs

@@ -1,5 +1,6 @@
 use crate::collector::top_collector::{TopCollector, TopSegmentCollector};
 use crate::collector::{Collector, SegmentCollector};
+use crate::Result;
 use crate::{DocAddress, DocId, Score, SegmentReader};
 
 pub(crate) struct CustomScoreTopCollector<TCustomScorer, TScore = Score> {
@@ -41,7 +42,7 @@ pub trait CustomScorer<TScore>: Sync {
     type Child: CustomSegmentScorer<TScore>;
     /// Builds a child scorer for a specific segment. The child scorer is associated to
     /// a specific segment.
-    fn segment_scorer(&self, segment_reader: &SegmentReader) -> crate::Result<Self::Child>;
+    fn segment_scorer(&self, segment_reader: &SegmentReader) -> Result<Self::Child>;
 }
 
 impl<TCustomScorer, TScore> Collector for CustomScoreTopCollector<TCustomScorer, TScore>
@@ -57,7 +58,7 @@ where
         &self,
         segment_local_id: u32,
         segment_reader: &SegmentReader,
-    ) -> crate::Result<Self::Child> {
+    ) -> Result<Self::Child> {
         let segment_scorer = self.custom_scorer.segment_scorer(segment_reader)?;
         let segment_collector = self
             .collector
@@ -72,7 +73,7 @@ where
         false
     }
 
-    fn merge_fruits(&self, segment_fruits: Vec<Self::Fruit>) -> crate::Result<Self::Fruit> {
+    fn merge_fruits(&self, segment_fruits: Vec<Self::Fruit>) -> Result<Self::Fruit> {
         self.collector.merge_fruits(segment_fruits)
     }
 }
@@ -110,7 +111,7 @@ where
 {
     type Child = T;
 
-    fn segment_scorer(&self, segment_reader: &SegmentReader) -> crate::Result<Self::Child> {
+    fn segment_scorer(&self, segment_reader: &SegmentReader) -> Result<Self::Child> {
         Ok((self)(segment_reader))
     }
 }

src/collector/facet_collector.rs

@@ -5,6 +5,7 @@ use crate::fastfield::FacetReader;
 use crate::schema::Facet;
 use crate::schema::Field;
 use crate::DocId;
+use crate::Result;
 use crate::Score;
 use crate::SegmentLocalId;
 use crate::SegmentReader;
@@ -83,9 +84,9 @@ fn facet_depth(facet_bytes: &[u8]) -> usize {
 /// use tantivy::collector::FacetCollector;
 /// use tantivy::query::AllQuery;
 /// use tantivy::schema::{Facet, Schema, TEXT};
-/// use tantivy::{doc, Index};
+/// use tantivy::{doc, Index, Result};
 ///
-/// fn example() -> tantivy::Result<()> {
+/// fn example() -> Result<()> {
 ///     let mut schema_builder = Schema::builder();
 ///
 ///     // Facet have their own specific type.
@@ -261,7 +262,7 @@ impl Collector for FacetCollector {
         &self,
         _: SegmentLocalId,
         reader: &SegmentReader,
-    ) -> crate::Result<FacetSegmentCollector> {
+    ) -> Result<FacetSegmentCollector> {
         let field_name = reader.schema().get_field_name(self.field);
         let facet_reader = reader.facet_reader(self.field).ok_or_else(|| {
             TantivyError::SchemaError(format!("Field {:?} is not a facet field.", field_name))
@@ -327,7 +328,7 @@ impl Collector for FacetCollector {
         false
     }
 
-    fn merge_fruits(&self, segments_facet_counts: Vec<FacetCounts>) -> crate::Result<FacetCounts> {
+    fn merge_fruits(&self, segments_facet_counts: Vec<FacetCounts>) -> Result<FacetCounts> {
         let mut facet_counts: BTreeMap<Facet, u64> = BTreeMap::new();
         for segment_facet_counts in segments_facet_counts {
             for (facet, count) in segment_facet_counts.facet_counts {

src/collector/mod.rs

@@ -85,6 +85,7 @@ See the `custom_collector` example.
 */
 
 use crate::DocId;
+use crate::Result;
 use crate::Score;
 use crate::SegmentLocalId;
 use crate::SegmentReader;
@@ -146,14 +147,14 @@ pub trait Collector: Sync {
         &self,
         segment_local_id: SegmentLocalId,
         segment: &SegmentReader,
-    ) -> crate::Result<Self::Child>;
+    ) -> Result<Self::Child>;
 
     /// Returns true iff the collector requires to compute scores for documents.
     fn requires_scoring(&self) -> bool;
 
     /// Combines the fruit associated to the collection of each segments
     /// into one fruit.
-    fn merge_fruits(&self, segment_fruits: Vec<Self::Fruit>) -> crate::Result<Self::Fruit>;
+    fn merge_fruits(&self, segment_fruits: Vec<Self::Fruit>) -> Result<Self::Fruit>;
 }
 
 /// The `SegmentCollector` is the trait in charge of defining the
@@ -184,11 +185,7 @@ where
     type Fruit = (Left::Fruit, Right::Fruit);
     type Child = (Left::Child, Right::Child);
 
-    fn for_segment(
-        &self,
-        segment_local_id: u32,
-        segment: &SegmentReader,
-    ) -> crate::Result<Self::Child> {
+    fn for_segment(&self, segment_local_id: u32, segment: &SegmentReader) -> Result<Self::Child> {
         let left = self.0.for_segment(segment_local_id, segment)?;
         let right = self.1.for_segment(segment_local_id, segment)?;
         Ok((left, right))
@@ -201,7 +198,7 @@ where
     fn merge_fruits(
         &self,
         children: Vec<(Left::Fruit, Right::Fruit)>,
-    ) -> crate::Result<(Left::Fruit, Right::Fruit)> {
+    ) -> Result<(Left::Fruit, Right::Fruit)> {
         let mut left_fruits = vec![];
         let mut right_fruits = vec![];
         for (left_fruit, right_fruit) in children {
@@ -243,11 +240,7 @@ where
     type Fruit = (One::Fruit, Two::Fruit, Three::Fruit);
     type Child = (One::Child, Two::Child, Three::Child);
 
-    fn for_segment(
-        &self,
-        segment_local_id: u32,
-        segment: &SegmentReader,
-    ) -> crate::Result<Self::Child> {
+    fn for_segment(&self, segment_local_id: u32, segment: &SegmentReader) -> Result<Self::Child> {
         let one = self.0.for_segment(segment_local_id, segment)?;
         let two = self.1.for_segment(segment_local_id, segment)?;
         let three = self.2.for_segment(segment_local_id, segment)?;
@@ -258,7 +251,7 @@ where
         self.0.requires_scoring() || self.1.requires_scoring() || self.2.requires_scoring()
     }
 
-    fn merge_fruits(&self, children: Vec<Self::Fruit>) -> crate::Result<Self::Fruit> {
+    fn merge_fruits(&self, children: Vec<Self::Fruit>) -> Result<Self::Fruit> {
         let mut one_fruits = vec![];
         let mut two_fruits = vec![];
         let mut three_fruits = vec![];
@@ -306,11 +299,7 @@ where
     type Fruit = (One::Fruit, Two::Fruit, Three::Fruit, Four::Fruit);
     type Child = (One::Child, Two::Child, Three::Child, Four::Child);
 
-    fn for_segment(
-        &self,
-        segment_local_id: u32,
-        segment: &SegmentReader,
-    ) -> crate::Result<Self::Child> {
+    fn for_segment(&self, segment_local_id: u32, segment: &SegmentReader) -> Result<Self::Child> {
         let one = self.0.for_segment(segment_local_id, segment)?;
         let two = self.1.for_segment(segment_local_id, segment)?;
         let three = self.2.for_segment(segment_local_id, segment)?;
@@ -325,7 +314,7 @@ where
             || self.3.requires_scoring()
     }
 
-    fn merge_fruits(&self, children: Vec<Self::Fruit>) -> crate::Result<Self::Fruit> {
+    fn merge_fruits(&self, children: Vec<Self::Fruit>) -> Result<Self::Fruit> {
         let mut one_fruits = vec![];
         let mut two_fruits = vec![];
         let mut three_fruits = vec![];

src/collector/multi_collector.rs

@@ -2,6 +2,7 @@ use super::Collector;
 use super::SegmentCollector;
 use crate::collector::Fruit;
 use crate::DocId;
+use crate::Result;
 use crate::Score;
 use crate::SegmentLocalId;
 use crate::SegmentReader;
@@ -23,7 +24,7 @@ impl<TCollector: Collector> Collector for CollectorWrapper<TCollector> {
         &self,
         segment_local_id: u32,
         reader: &SegmentReader,
-    ) -> crate::Result<Box<dyn BoxableSegmentCollector>> {
+    ) -> Result<Box<dyn BoxableSegmentCollector>> {
         let child = self.0.for_segment(segment_local_id, reader)?;
         Ok(Box::new(SegmentCollectorWrapper(child)))
     }
@@ -32,10 +33,7 @@ impl<TCollector: Collector> Collector for CollectorWrapper<TCollector> {
         self.0.requires_scoring()
     }
 
-    fn merge_fruits(
-        &self,
-        children: Vec<<Self as Collector>::Fruit>,
-    ) -> crate::Result<Box<dyn Fruit>> {
+    fn merge_fruits(&self, children: Vec<<Self as Collector>::Fruit>) -> Result<Box<dyn Fruit>> {
         let typed_fruit: Vec<TCollector::Fruit> = children
             .into_iter()
             .map(|untyped_fruit| {
@@ -46,7 +44,7 @@ impl<TCollector: Collector> Collector for CollectorWrapper<TCollector> {
                         TantivyError::InvalidArgument("Failed to cast child fruit.".to_string())
                     })
             })
-            .collect::<crate::Result<_>>()?;
+            .collect::<Result<_>>()?;
         let merged_fruit = self.0.merge_fruits(typed_fruit)?;
         Ok(Box::new(merged_fruit))
     }
@@ -177,12 +175,12 @@ impl<'a> Collector for MultiCollector<'a> {
         &self,
         segment_local_id: SegmentLocalId,
         segment: &SegmentReader,
-    ) -> crate::Result<MultiCollectorChild> {
+    ) -> Result<MultiCollectorChild> {
         let children = self
             .collector_wrappers
             .iter()
             .map(|collector_wrapper| collector_wrapper.for_segment(segment_local_id, segment))
-            .collect::<crate::Result<Vec<_>>>()?;
+            .collect::<Result<Vec<_>>>()?;
         Ok(MultiCollectorChild { children })
     }
 
@@ -193,7 +191,7 @@ impl<'a> Collector for MultiCollector<'a> {
             .any(Collector::requires_scoring)
     }
 
-    fn merge_fruits(&self, segments_multifruits: Vec<MultiFruit>) -> crate::Result<MultiFruit> {
+    fn merge_fruits(&self, segments_multifruits: Vec<MultiFruit>) -> Result<MultiFruit> {
         let mut segment_fruits_list: Vec<Vec<Box<dyn Fruit>>> = (0..self.collector_wrappers.len())
             .map(|_| Vec::with_capacity(segments_multifruits.len()))
             .collect::<Vec<_>>();
@@ -211,7 +209,7 @@ impl<'a> Collector for MultiCollector<'a> {
             .map(|(child_collector, segment_fruits)| {
                 Ok(Some(child_collector.merge_fruits(segment_fruits)?))
             })
-            .collect::<crate::Result<_>>()?;
+            .collect::<Result<_>>()?;
         Ok(MultiFruit { sub_fruits })
     }
 }

src/collector/tests.rs

@@ -55,7 +55,7 @@ impl Collector for TestCollector {
         &self,
         segment_id: SegmentLocalId,
         _reader: &SegmentReader,
-    ) -> crate::Result<TestSegmentCollector> {
+    ) -> Result<TestSegmentCollector> {
         Ok(TestSegmentCollector {
             segment_id,
             fruit: TestFruit::default(),
@@ -66,7 +66,7 @@ impl Collector for TestCollector {
         self.compute_score
     }
 
-    fn merge_fruits(&self, mut children: Vec<TestFruit>) -> crate::Result<TestFruit> {
+    fn merge_fruits(&self, mut children: Vec<TestFruit>) -> Result<TestFruit> {
         children.sort_by_key(|fruit| {
             if fruit.docs().is_empty() {
                 0
@@ -124,7 +124,7 @@ impl Collector for FastFieldTestCollector {
         &self,
         _: SegmentLocalId,
         segment_reader: &SegmentReader,
-    ) -> crate::Result<FastFieldSegmentCollector> {
+    ) -> Result<FastFieldSegmentCollector> {
         let reader = segment_reader
             .fast_fields()
             .u64(self.field)
@@ -139,7 +139,7 @@ impl Collector for FastFieldTestCollector {
         false
     }
 
-    fn merge_fruits(&self, children: Vec<Vec<u64>>) -> crate::Result<Vec<u64>> {
+    fn merge_fruits(&self, children: Vec<Vec<u64>>) -> Result<Vec<u64>> {
         Ok(children.into_iter().flat_map(|v| v.into_iter()).collect())
     }
 }
@@ -184,7 +184,7 @@ impl Collector for BytesFastFieldTestCollector {
         &self,
         _segment_local_id: u32,
         segment_reader: &SegmentReader,
-    ) -> crate::Result<BytesFastFieldSegmentCollector> {
+    ) -> Result<BytesFastFieldSegmentCollector> {
         Ok(BytesFastFieldSegmentCollector {
             vals: Vec::new(),
             reader: segment_reader
@@ -198,7 +198,7 @@ impl Collector for BytesFastFieldTestCollector {
         false
     }
 
-    fn merge_fruits(&self, children: Vec<Vec<u8>>) -> crate::Result<Vec<u8>> {
+    fn merge_fruits(&self, children: Vec<Vec<u8>>) -> Result<Vec<u8>> {
         Ok(children.into_iter().flat_map(|c| c.into_iter()).collect())
     }
 }

src/collector/top_collector.rs

@@ -1,5 +1,6 @@
 use crate::DocAddress;
 use crate::DocId;
+use crate::Result;
 use crate::SegmentLocalId;
 use crate::SegmentReader;
 use serde::export::PhantomData;
@@ -85,7 +86,7 @@ where
     pub fn merge_fruits(
         &self,
         children: Vec<Vec<(T, DocAddress)>>,
-    ) -> crate::Result<Vec<(T, DocAddress)>> {
+    ) -> Result<Vec<(T, DocAddress)>> {
         if self.limit == 0 {
             return Ok(Vec::new());
         }
@@ -112,7 +113,7 @@ where
         &self,
         segment_id: SegmentLocalId,
         _: &SegmentReader,
-    ) -> crate::Result<TopSegmentCollector<F>> {
+    ) -> Result<TopSegmentCollector<F>> {
         Ok(TopSegmentCollector::new(segment_id, self.limit))
     }
 }

src/collector/top_score_collector.rs

@@ -6,10 +6,10 @@ 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;
+use crate::Result;
 use crate::Score;
 use crate::SegmentLocalId;
 use crate::SegmentReader;
@@ -61,36 +61,6 @@ 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::TantivyError::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".
     ///
@@ -104,7 +74,7 @@ impl TopDocs {
     ///
     /// ```rust
     /// # use tantivy::schema::{Schema, FAST, TEXT};
-    /// # use tantivy::{doc, Index, DocAddress};
+    /// # use tantivy::{doc, Index, Result, DocAddress};
     /// # use tantivy::query::{Query, QueryParser};
     /// use tantivy::Searcher;
     /// use tantivy::collector::TopDocs;
@@ -141,7 +111,7 @@ impl TopDocs {
     /// fn docs_sorted_by_rating(searcher: &Searcher,
     ///                          query: &dyn Query,
     ///                          sort_by_field: Field)
-    ///     -> tantivy::Result<Vec<(u64, DocAddress)>> {
+    ///     -> Result<Vec<(u64, DocAddress)>> {
     ///
     ///     // This is where we build our topdocs collector
     ///     //
@@ -173,7 +143,14 @@ impl TopDocs {
         self,
         field: Field,
     ) -> impl Collector<Fruit = Vec<(u64, DocAddress)>> {
-        self.custom_score(ScorerByField { field })
+        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)
+        })
     }
 
     /// Ranks the documents using a custom score.
@@ -407,7 +384,7 @@ impl Collector for TopDocs {
         &self,
         segment_local_id: SegmentLocalId,
         reader: &SegmentReader,
-    ) -> crate::Result<Self::Child> {
+    ) -> Result<Self::Child> {
         let collector = self.0.for_segment(segment_local_id, reader)?;
         Ok(TopScoreSegmentCollector(collector))
     }
@@ -416,10 +393,7 @@ impl Collector for TopDocs {
         true
     }
 
-    fn merge_fruits(
-        &self,
-        child_fruits: Vec<Vec<(Score, DocAddress)>>,
-    ) -> crate::Result<Self::Fruit> {
+    fn merge_fruits(&self, child_fruits: Vec<Vec<(Score, DocAddress)>>) -> Result<Self::Fruit> {
         self.0.merge_fruits(child_fruits)
     }
 }
@@ -598,6 +572,7 @@ 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);
@@ -612,12 +587,7 @@ 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);
-        let err = top_collector.for_segment(0, segment);
-        if let Err(crate::TantivyError::SchemaError(msg)) = err {
-            assert_eq!(msg, "Field requested is not a i64/u64 fast field.");
-        } else {
-            assert!(false);
-        }
+        assert!(top_collector.for_segment(0, segment).is_ok());
     }
 
     fn index(

src/common/composite_file.rs

@@ -186,7 +186,7 @@ mod test {
     use super::{CompositeFile, CompositeWrite};
     use crate::common::BinarySerializable;
     use crate::common::VInt;
-    use crate::directory::{Directory, RAMDirectory};
+    use crate::directory::{Directory, RAMDirectory, ReadOnlyDirectory};
     use crate::schema::Field;
     use std::io::Write;
     use std::path::Path;

src/core/executor.rs

@@ -1,3 +1,4 @@
+use crate::Result;
 use crossbeam::channel;
 use rayon::{ThreadPool, ThreadPoolBuilder};
 
@@ -9,9 +10,7 @@ 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),
 }
 
@@ -21,8 +20,8 @@ impl Executor {
         Executor::SingleThread
     }
 
-    /// Creates an Executor that dispatches the tasks in a thread pool.
-    pub fn multi_thread(num_threads: usize, prefix: &'static str) -> crate::Result<Executor> {
+    // 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)
             .thread_name(move |num| format!("{}{}", prefix, num))
@@ -30,22 +29,22 @@ 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,
         AIterator: Iterator<Item = A>,
-        F: Sized + Sync + Fn(A) -> crate::Result<R>,
+        F: Sized + Sync + Fn(A) -> Result<R>,
     >(
         &self,
         f: F,
         args: AIterator,
-    ) -> crate::Result<Vec<R>> {
+    ) -> Result<Vec<R>> {
         match self {
-            Executor::SingleThread => args.map(f).collect::<crate::Result<_>>(),
+            Executor::SingleThread => args.map(f).collect::<Result<_>>(),
             Executor::ThreadPool(pool) => {
                 let args_with_indices: Vec<(usize, A)> = args.enumerate().collect();
                 let num_fruits = args_with_indices.len();

src/core/index.rs

@@ -20,8 +20,10 @@ use crate::reader::IndexReaderBuilder;
 use crate::schema::Field;
 use crate::schema::FieldType;
 use crate::schema::Schema;
-use crate::tokenizer::{TextAnalyzer, TokenizerManager};
+use crate::tokenizer::BoxedTokenizer;
+use crate::tokenizer::TokenizerManager;
 use crate::IndexWriter;
+use crate::Result;
 use num_cpus;
 use std::borrow::BorrowMut;
 use std::collections::HashSet;
@@ -30,10 +32,7 @@ use std::fmt;
 use std::path::{Path, PathBuf};
 use std::sync::Arc;
 
-fn load_metas(
-    directory: &dyn Directory,
-    inventory: &SegmentMetaInventory,
-) -> crate::Result<IndexMeta> {
+fn load_metas(directory: &dyn Directory, inventory: &SegmentMetaInventory) -> Result<IndexMeta> {
     let meta_data = directory.atomic_read(&META_FILEPATH)?;
     let meta_string = String::from_utf8_lossy(&meta_data);
     IndexMeta::deserialize(&meta_string, &inventory)
@@ -74,14 +73,14 @@ impl Index {
 
     /// Replace the default single thread search executor pool
     /// by a thread pool with a given number of threads.
-    pub fn set_multithread_executor(&mut self, num_threads: usize) -> crate::Result<()> {
+    pub fn set_multithread_executor(&mut self, num_threads: usize) -> Result<()> {
         self.executor = Arc::new(Executor::multi_thread(num_threads, "thrd-tantivy-search-")?);
         Ok(())
     }
 
     /// Replace the default single thread search executor pool
     /// by a thread pool with a given number of threads.
-    pub fn set_default_multithread_executor(&mut self) -> crate::Result<()> {
+    pub fn set_default_multithread_executor(&mut self) -> Result<()> {
         let default_num_threads = num_cpus::get();
         self.set_multithread_executor(default_num_threads)
     }
@@ -100,10 +99,7 @@ impl Index {
     ///
     /// If a previous index was in this directory, then its meta file will be destroyed.
     #[cfg(feature = "mmap")]
-    pub fn create_in_dir<P: AsRef<Path>>(
-        directory_path: P,
-        schema: Schema,
-    ) -> crate::Result<Index> {
+    pub fn create_in_dir<P: AsRef<Path>>(directory_path: P, schema: Schema) -> Result<Index> {
         let mmap_directory = MmapDirectory::open(directory_path)?;
         if Index::exists(&mmap_directory) {
             return Err(TantivyError::IndexAlreadyExists);
@@ -112,7 +108,7 @@ impl Index {
     }
 
     /// Opens or creates a new index in the provided directory
-    pub fn open_or_create<Dir: Directory>(dir: Dir, schema: Schema) -> crate::Result<Index> {
+    pub fn open_or_create<Dir: Directory>(dir: Dir, schema: Schema) -> Result<Index> {
         if !Index::exists(&dir) {
             return Index::create(dir, schema);
         }
@@ -135,13 +131,13 @@ impl Index {
     /// The temp directory is only used for testing the `MmapDirectory`.
     /// For other unit tests, prefer the `RAMDirectory`, see: `create_in_ram`.
     #[cfg(feature = "mmap")]
-    pub fn create_from_tempdir(schema: Schema) -> crate::Result<Index> {
+    pub fn create_from_tempdir(schema: Schema) -> Result<Index> {
         let mmap_directory = MmapDirectory::create_from_tempdir()?;
         Index::create(mmap_directory, schema)
     }
 
     /// Creates a new index given an implementation of the trait `Directory`
-    pub fn create<Dir: Directory>(dir: Dir, schema: Schema) -> crate::Result<Index> {
+    pub fn create<Dir: Directory>(dir: Dir, schema: Schema) -> Result<Index> {
         let directory = ManagedDirectory::wrap(dir)?;
         Index::from_directory(directory, schema)
     }
@@ -149,7 +145,7 @@ impl Index {
     /// Create a new index from a directory.
     ///
     /// This will overwrite existing meta.json
-    fn from_directory(mut directory: ManagedDirectory, schema: Schema) -> crate::Result<Index> {
+    fn from_directory(mut directory: ManagedDirectory, schema: Schema) -> Result<Index> {
         save_new_metas(schema.clone(), directory.borrow_mut())?;
         let metas = IndexMeta::with_schema(schema);
         Index::create_from_metas(directory, &metas, SegmentMetaInventory::default())
@@ -160,7 +156,7 @@ impl Index {
         directory: ManagedDirectory,
         metas: &IndexMeta,
         inventory: SegmentMetaInventory,
-    ) -> crate::Result<Index> {
+    ) -> Result<Index> {
         let schema = metas.schema.clone();
         Ok(Index {
             directory,
@@ -177,11 +173,11 @@ impl Index {
     }
 
     /// Helper to access the tokenizer associated to a specific field.
-    pub fn tokenizer_for_field(&self, field: Field) -> crate::Result<TextAnalyzer> {
+    pub fn tokenizer_for_field(&self, field: Field) -> Result<BoxedTokenizer> {
         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<TextAnalyzer> = match field_type {
+        let tokenizer_name_opt: Option<BoxedTokenizer> = match field_type {
             FieldType::Str(text_options) => text_options
                 .get_indexing_options()
                 .map(|text_indexing_options| text_indexing_options.tokenizer().to_string())
@@ -200,7 +196,7 @@ impl Index {
     /// Create a default `IndexReader` for the given index.
     ///
     /// See [`Index.reader_builder()`](#method.reader_builder).
-    pub fn reader(&self) -> crate::Result<IndexReader> {
+    pub fn reader(&self) -> Result<IndexReader> {
         self.reader_builder().try_into()
     }
 
@@ -215,7 +211,7 @@ impl Index {
 
     /// Opens a new directory from an index path.
     #[cfg(feature = "mmap")]
-    pub fn open_in_dir<P: AsRef<Path>>(directory_path: P) -> crate::Result<Index> {
+    pub fn open_in_dir<P: AsRef<Path>>(directory_path: P) -> Result<Index> {
         let mmap_directory = MmapDirectory::open(directory_path)?;
         Index::open(mmap_directory)
     }
@@ -239,7 +235,7 @@ impl Index {
     }
 
     /// Open the index using the provided directory
-    pub fn open<D: Directory>(directory: D) -> crate::Result<Index> {
+    pub fn open<D: Directory>(directory: D) -> Result<Index> {
         let directory = ManagedDirectory::wrap(directory)?;
         let inventory = SegmentMetaInventory::default();
         let metas = load_metas(&directory, &inventory)?;
@@ -247,7 +243,7 @@ impl Index {
     }
 
     /// Reads the index meta file from the directory.
-    pub fn load_metas(&self) -> crate::Result<IndexMeta> {
+    pub fn load_metas(&self) -> Result<IndexMeta> {
         load_metas(self.directory(), &self.inventory)
     }
 
@@ -275,7 +271,7 @@ impl Index {
         &self,
         num_threads: usize,
         overall_heap_size_in_bytes: usize,
-    ) -> crate::Result<IndexWriter> {
+    ) -> Result<IndexWriter> {
         let directory_lock = self
             .directory
             .acquire_lock(&INDEX_WRITER_LOCK)
@@ -310,7 +306,7 @@ impl Index {
     /// If the lockfile already exists, returns `Error::FileAlreadyExists`.
     /// # Panics
     /// If the heap size per thread is too small, panics.
-    pub fn writer(&self, overall_heap_size_in_bytes: usize) -> crate::Result<IndexWriter> {
+    pub fn writer(&self, overall_heap_size_in_bytes: usize) -> Result<IndexWriter> {
         let mut num_threads = num_cpus::get();
         let heap_size_in_bytes_per_thread = overall_heap_size_in_bytes / num_threads;
         if heap_size_in_bytes_per_thread < HEAP_SIZE_MIN {
@@ -327,7 +323,7 @@ impl Index {
     }
 
     /// Returns the list of segments that are searchable
-    pub fn searchable_segments(&self) -> crate::Result<Vec<Segment>> {
+    pub fn searchable_segments(&self) -> Result<Vec<Segment>> {
         Ok(self
             .searchable_segment_metas()?
             .into_iter()
@@ -342,7 +338,7 @@ impl Index {
 
     /// Creates a new segment.
     pub fn new_segment(&self) -> Segment {
-        let segment_meta = self
+        let mut segment_meta = self
             .inventory
             .new_segment_meta(SegmentId::generate_random(), 0);
         self.segment(segment_meta)
@@ -360,12 +356,12 @@ impl Index {
 
     /// Reads the meta.json and returns the list of
     /// `SegmentMeta` from the last commit.
-    pub fn searchable_segment_metas(&self) -> crate::Result<Vec<SegmentMeta>> {
+    pub fn searchable_segment_metas(&self) -> Result<Vec<SegmentMeta>> {
         Ok(self.load_metas()?.segments)
     }
 
     /// Returns the list of segment ids that are searchable.
-    pub fn searchable_segment_ids(&self) -> crate::Result<Vec<SegmentId>> {
+    pub fn searchable_segment_ids(&self) -> Result<Vec<SegmentId>> {
         Ok(self
             .searchable_segment_metas()?
             .iter()
@@ -374,7 +370,7 @@ impl Index {
     }
 
     /// Returns the set of corrupted files
-    pub fn validate_checksum(&self) -> crate::Result<HashSet<PathBuf>> {
+    pub fn validate_checksum(&self) -> Result<HashSet<PathBuf>> {
         self.directory.list_damaged().map_err(Into::into)
     }
 }

src/core/index_meta.rs

@@ -35,6 +35,7 @@ impl SegmentMetaInventory {
             segment_id,
             max_doc,
             deletes: None,
+            bundled: false,
         };
         SegmentMeta::from(self.inventory.track(inner))
     }
@@ -81,6 +82,19 @@ impl SegmentMeta {
         self.tracked.segment_id
     }
 
+    pub fn with_bundled(self) -> SegmentMeta {
+        SegmentMeta::from(self.tracked.map(|inner| InnerSegmentMeta {
+            segment_id: inner.segment_id,
+            max_doc: inner.max_doc,
+            deletes: inner.deletes.clone(),
+            bundled: true,
+        }))
+    }
+
+    pub fn is_bundled(&self) -> bool {
+        self.tracked.bundled
+    }
+
     /// Returns the number of deleted documents.
     pub fn num_deleted_docs(&self) -> u32 {
         self.tracked
@@ -107,8 +121,12 @@ impl SegmentMeta {
     /// It just joins the segment id with the extension
     /// associated to a segment component.
     pub fn relative_path(&self, component: SegmentComponent) -> PathBuf {
-        let mut path = self.id().uuid_string();
-        path.push_str(&*match component {
+        let suffix = self.suffix(component);
+        self.relative_path_from_suffix(&suffix)
+    }
+
+    fn suffix(&self, component: SegmentComponent) -> String {
+        match component {
             SegmentComponent::POSTINGS => ".idx".to_string(),
             SegmentComponent::POSITIONS => ".pos".to_string(),
             SegmentComponent::POSITIONSSKIP => ".posidx".to_string(),
@@ -117,7 +135,17 @@ impl SegmentMeta {
             SegmentComponent::FASTFIELDS => ".fast".to_string(),
             SegmentComponent::FIELDNORMS => ".fieldnorm".to_string(),
             SegmentComponent::DELETE => format!(".{}.del", self.delete_opstamp().unwrap_or(0)),
-        });
+        }
+    }
+
+    /// Returns the relative path of a component of our segment.
+    ///
+    /// It just joins the segment id with the extension
+    /// associated to a segment component.
+    pub fn relative_path_from_suffix(&self, suffix: &str) -> PathBuf {
+        let mut path = self.id().uuid_string();
+        path.push_str(".");
+        path.push_str(&suffix);
         PathBuf::from(path)
     }
 
@@ -161,6 +189,7 @@ impl SegmentMeta {
             segment_id: inner_meta.segment_id,
             max_doc,
             deletes: None,
+            bundled: inner_meta.bundled,
         });
         SegmentMeta { tracked }
     }
@@ -175,6 +204,7 @@ impl SegmentMeta {
             segment_id: inner_meta.segment_id,
             max_doc: inner_meta.max_doc,
             deletes: Some(delete_meta),
+            bundled: inner_meta.bundled,
         });
         SegmentMeta { tracked }
     }
@@ -185,6 +215,7 @@ struct InnerSegmentMeta {
     segment_id: SegmentId,
     max_doc: u32,
     deletes: Option<DeleteMeta>,
+    bundled: bool,
 }
 
 impl InnerSegmentMeta {

src/core/inverted_index_reader.rs

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

src/core/searcher.rs

@@ -14,6 +14,7 @@ use crate::store::StoreReader;
 use crate::termdict::TermMerger;
 use crate::DocAddress;
 use crate::Index;
+use crate::Result;
 use std::fmt;
 use std::sync::Arc;
 
@@ -22,7 +23,7 @@ fn collect_segment<C: Collector>(
     weight: &dyn Weight,
     segment_ord: u32,
     segment_reader: &SegmentReader,
-) -> crate::Result<C::Fruit> {
+) -> Result<C::Fruit> {
     let mut scorer = weight.scorer(segment_reader)?;
     let mut segment_collector = collector.for_segment(segment_ord as u32, segment_reader)?;
     if let Some(delete_bitset) = segment_reader.delete_bitset() {
@@ -77,7 +78,7 @@ impl Searcher {
     ///
     /// The searcher uses the segment ordinal to route the
     /// the request to the right `Segment`.
-    pub fn doc(&self, doc_address: DocAddress) -> crate::Result<Document> {
+    pub fn doc(&self, doc_address: DocAddress) -> Result<Document> {
         let DocAddress(segment_local_id, doc_id) = doc_address;
         let store_reader = &self.store_readers[segment_local_id as usize];
         store_reader.get(doc_id)
@@ -131,11 +132,7 @@ impl Searcher {
     ///
     ///  Finally, the Collector merges each of the child collectors into itself for result usability
     ///  by the caller.
-    pub fn search<C: Collector>(
-        &self,
-        query: &dyn Query,
-        collector: &C,
-    ) -> crate::Result<C::Fruit> {
+    pub fn search<C: Collector>(&self, query: &dyn Query, collector: &C) -> Result<C::Fruit> {
         let executor = self.index.search_executor();
         self.search_with_executor(query, collector, executor)
     }
@@ -157,7 +154,7 @@ impl Searcher {
         query: &dyn Query,
         collector: &C,
         executor: &Executor,
-    ) -> crate::Result<C::Fruit> {
+    ) -> Result<C::Fruit> {
         let scoring_enabled = collector.requires_scoring();
         let weight = query.weight(self, scoring_enabled)?;
         let segment_readers = self.segment_readers();

src/core/segment.rs

@@ -4,7 +4,7 @@ use crate::core::SegmentId;
 use crate::core::SegmentMeta;
 use crate::directory::error::{OpenReadError, OpenWriteError};
 use crate::directory::Directory;
-use crate::directory::{ReadOnlySource, WritePtr};
+use crate::directory::{ReadOnlyDirectory, ReadOnlySource, WritePtr};
 use crate::indexer::segment_serializer::SegmentSerializer;
 use crate::schema::Schema;
 use crate::Opstamp;
@@ -90,8 +90,21 @@ impl Segment {
     /// Open one of the component file for *regular* write.
     pub fn open_write(&mut self, component: SegmentComponent) -> Result<WritePtr, OpenWriteError> {
         let path = self.relative_path(component);
-        let write = self.index.directory_mut().open_write(&path)?;
-        Ok(write)
+        self.index.directory_mut().open_write(&path)
+    }
+
+    pub fn open_bundle_writer(&mut self) -> Result<WritePtr, OpenWriteError> {
+        let path = self.meta.relative_path_from_suffix("bundle");
+        self.index.directory_mut().open_write(&path)
+    }
+
+    pub(crate) fn open_write_in_directory(
+        &mut self,
+        component: SegmentComponent,
+        directory: &mut dyn Directory,
+    ) -> Result<WritePtr, OpenWriteError> {
+        let path = self.relative_path(component);
+        directory.open_write(&path)
     }
 }
 

src/core/segment_reader.rs

@@ -16,6 +16,7 @@ use crate::space_usage::SegmentSpaceUsage;
 use crate::store::StoreReader;
 use crate::termdict::TermDictionary;
 use crate::DocId;
+use crate::Result;
 use fail::fail_point;
 use std::collections::HashMap;
 use std::fmt;
@@ -144,7 +145,7 @@ impl SegmentReader {
     }
 
     /// Open a new segment for reading.
-    pub fn open(segment: &Segment) -> crate::Result<SegmentReader> {
+    pub fn open(segment: &Segment) -> Result<SegmentReader> {
         let termdict_source = segment.open_read(SegmentComponent::TERMS)?;
         let termdict_composite = CompositeFile::open(&termdict_source)?;
 

src/directory/bundle_directory.rs

@@ -0,0 +1,97 @@
+use crate::directory::directory::ReadOnlyDirectory;
+use crate::directory::error::OpenReadError;
+use crate::directory::ReadOnlySource;
+use crate::error::DataCorruption;
+use std::collections::HashMap;
+use std::path::{Path, PathBuf};
+use std::sync::Arc;
+
+#[derive(Clone)]
+struct BundleDirectory {
+    source_map: Arc<HashMap<PathBuf, ReadOnlySource>>,
+}
+
+impl BundleDirectory {
+    pub fn from_source(source: ReadOnlySource) -> Result<BundleDirectory, DataCorruption> {
+        let mut index_offset_buf = [0u8; 8];
+        let (body_idx, footer_offset) = source.split_from_end(8);
+        index_offset_buf.copy_from_slice(footer_offset.as_slice());
+        let offset = u64::from_le_bytes(index_offset_buf);
+        let (body_source, idx_source) = body_idx.split(offset as usize);
+        let idx: HashMap<PathBuf, (u64, u64)> = serde_json::from_slice(idx_source.as_slice())
+            .map_err(|err| {
+                let msg = format!("Failed to read index from bundle. {:?}", err);
+                DataCorruption::comment_only(msg)
+            })?;
+        let source_map: HashMap<PathBuf, ReadOnlySource> = idx
+            .into_iter()
+            .map(|(path, (start, stop))| {
+                let source = body_source.slice(start as usize, stop as usize);
+                (path, source)
+            })
+            .collect();
+        Ok(BundleDirectory {
+            source_map: Arc::new(source_map),
+        })
+    }
+}
+
+impl ReadOnlyDirectory for BundleDirectory {
+    fn open_read(&self, path: &Path) -> Result<ReadOnlySource, OpenReadError> {
+        self.source_map
+            .get(path)
+            .cloned()
+            .ok_or_else(|| OpenReadError::FileDoesNotExist(path.to_path_buf()))
+    }
+
+    fn exists(&self, path: &Path) -> bool {
+        self.source_map.contains_key(path)
+    }
+
+    fn atomic_read(&self, path: &Path) -> Result<Vec<u8>, OpenReadError> {
+        let source = self
+            .source_map
+            .get(path)
+            .ok_or_else(|| OpenReadError::FileDoesNotExist(path.to_path_buf()))?;
+        Ok(source.as_slice().to_vec())
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::BundleDirectory;
+    use crate::directory::{RAMDirectory, ReadOnlyDirectory, TerminatingWrite};
+    use crate::Directory;
+    use std::io::Write;
+    use std::path::Path;
+
+    #[test]
+    fn test_bundle_directory() {
+        let mut ram_directory = RAMDirectory::default();
+        let test_path_atomic = Path::new("testpath_atomic");
+        let test_path_wrt = Path::new("testpath_wrt");
+        assert!(ram_directory
+            .atomic_write(test_path_atomic, b"titi")
+            .is_ok());
+        {
+            let mut test_wrt = ram_directory.open_write(test_path_wrt).unwrap();
+            assert!(test_wrt.write_all(b"toto").is_ok());
+            assert!(test_wrt.terminate().is_ok());
+        }
+        let mut dest_directory = RAMDirectory::default();
+        let bundle_path = Path::new("bundle");
+        let mut wrt = dest_directory.open_write(bundle_path).unwrap();
+        assert!(ram_directory.serialize_bundle(&mut wrt).is_ok());
+        assert!(wrt.terminate().is_ok());
+        let source = dest_directory.open_read(bundle_path).unwrap();
+        let bundle_directory = BundleDirectory::from_source(source).unwrap();
+        assert_eq!(
+            &bundle_directory.atomic_read(test_path_atomic).unwrap()[..],
+            b"titi"
+        );
+        assert_eq!(
+            &bundle_directory.open_read(test_path_wrt).unwrap()[..],
+            b"toto"
+        );
+    }
+}

src/directory/directory.rs

@@ -100,17 +100,7 @@ fn retry_policy(is_blocking: bool) -> RetryPolicy {
     }
 }
 
-/// Write-once read many (WORM) abstraction for where
-/// tantivy's data should be stored.
-///
-/// There are currently two implementations of `Directory`
-///
-/// - The [`MMapDirectory`](struct.MmapDirectory.html), this
-/// should be your default choice.
-/// - The [`RAMDirectory`](struct.RAMDirectory.html), which
-/// should be used mostly for tests.
-///
-pub trait Directory: DirectoryClone + fmt::Debug + Send + Sync + 'static {
+pub trait ReadOnlyDirectory {
     /// Opens a virtual file for read.
     ///
     /// Once a virtual file is open, its data may not
@@ -122,6 +112,31 @@ pub trait Directory: DirectoryClone + fmt::Debug + Send + Sync + 'static {
     /// You should only use this to read files create with [Directory::open_write].
     fn open_read(&self, path: &Path) -> result::Result<ReadOnlySource, OpenReadError>;
 
+    /// Returns true iff the file exists
+    fn exists(&self, path: &Path) -> bool;
+
+    /// Reads the full content file that has been written using
+    /// atomic_write.
+    ///
+    /// This should only be used for small files.
+    ///
+    /// You should only use this to read files create with [Directory::atomic_write].
+    fn atomic_read(&self, path: &Path) -> Result<Vec<u8>, OpenReadError>;
+}
+
+/// Write-once read many (WORM) abstraction for where
+/// tantivy's data should be stored.
+///
+/// There are currently two implementations of `Directory`
+///
+/// - The [`MMapDirectory`](struct.MmapDirectory.html), this
+/// should be your default choice.
+/// - The [`RAMDirectory`](struct.RAMDirectory.html), which
+/// should be used mostly for tests.
+///
+pub trait Directory:
+    DirectoryClone + ReadOnlyDirectory + fmt::Debug + Send + Sync + 'static
+{
     /// Removes a file
     ///
     /// Removing a file will not affect an eventual
@@ -131,9 +146,6 @@ pub trait Directory: DirectoryClone + fmt::Debug + Send + Sync + 'static {
     /// `DeleteError::DoesNotExist`.
     fn delete(&self, path: &Path) -> result::Result<(), DeleteError>;
 
-    /// Returns true iff the file exists
-    fn exists(&self, path: &Path) -> bool;
-
     /// Opens a writer for the *virtual file* associated with
     /// a Path.
     ///
@@ -155,14 +167,6 @@ pub trait Directory: DirectoryClone + fmt::Debug + Send + Sync + 'static {
     /// The file may not previously exist.
     fn open_write(&mut self, path: &Path) -> Result<WritePtr, OpenWriteError>;
 
-    /// Reads the full content file that has been written using
-    /// atomic_write.
-    ///
-    /// This should only be used for small files.
-    ///
-    /// 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.
     ///
     /// This calls ensure that reads can never *observe*

src/directory/managed_directory.rs

@@ -10,6 +10,7 @@ use crate::directory::{WatchCallback, WatchHandle};
 use crate::error::DataCorruption;
 use crate::Directory;
 
+use crate::directory::directory::ReadOnlyDirectory;
 use crc32fast::Hasher;
 use serde_json;
 use std::collections::HashSet;
@@ -150,7 +151,7 @@ impl ManagedDirectory {
                 }
                 Err(err) => {
                     error!("Failed to acquire lock for GC");
-                    return Err(crate::TantivyError::from(err));
+                    return Err(crate::Error::from(err));
                 }
             }
         }
@@ -264,14 +265,6 @@ impl ManagedDirectory {
 }
 
 impl Directory for ManagedDirectory {
-    fn open_read(&self, path: &Path) -> result::Result<ReadOnlySource, OpenReadError> {
-        let read_only_source = self.directory.open_read(path)?;
-        let (footer, reader) = Footer::extract_footer(read_only_source)
-            .map_err(|err| IOError::with_path(path.to_path_buf(), err))?;
-        footer.is_compatible()?;
-        Ok(reader)
-    }
-
     fn open_write(&mut self, path: &Path) -> result::Result<WritePtr, OpenWriteError> {
         self.register_file_as_managed(path)
             .map_err(|e| IOError::with_path(path.to_owned(), e))?;
@@ -289,18 +282,10 @@ impl Directory for ManagedDirectory {
         self.directory.atomic_write(path, data)
     }
 
-    fn atomic_read(&self, path: &Path) -> result::Result<Vec<u8>, OpenReadError> {
-        self.directory.atomic_read(path)
-    }
-
     fn delete(&self, path: &Path) -> result::Result<(), DeleteError> {
         self.directory.delete(path)
     }
 
-    fn exists(&self, path: &Path) -> bool {
-        self.directory.exists(path)
-    }
-
     fn acquire_lock(&self, lock: &Lock) -> result::Result<DirectoryLock, LockError> {
         self.directory.acquire_lock(lock)
     }
@@ -310,6 +295,24 @@ impl Directory for ManagedDirectory {
     }
 }
 
+impl ReadOnlyDirectory for ManagedDirectory {
+    fn open_read(&self, path: &Path) -> result::Result<ReadOnlySource, OpenReadError> {
+        let read_only_source = self.directory.open_read(path)?;
+        let (footer, reader) = Footer::extract_footer(read_only_source)
+            .map_err(|err| IOError::with_path(path.to_path_buf(), err))?;
+        footer.is_compatible()?;
+        Ok(reader)
+    }
+
+    fn exists(&self, path: &Path) -> bool {
+        self.directory.exists(path)
+    }
+
+    fn atomic_read(&self, path: &Path) -> result::Result<Vec<u8>, OpenReadError> {
+        self.directory.atomic_read(path)
+    }
+}
+
 impl Clone for ManagedDirectory {
     fn clone(&self) -> ManagedDirectory {
         ManagedDirectory {
@@ -323,7 +326,9 @@ impl Clone for ManagedDirectory {
 #[cfg(test)]
 mod tests_mmap_specific {
 
-    use crate::directory::{Directory, ManagedDirectory, MmapDirectory, TerminatingWrite};
+    use crate::directory::{
+        Directory, ManagedDirectory, MmapDirectory, ReadOnlyDirectory, TerminatingWrite,
+    };
     use std::collections::HashSet;
     use std::fs::OpenOptions;
     use std::io::Write;

src/directory/mmap_directory.rs

@@ -6,6 +6,7 @@ use self::notify::RawEvent;
 use self::notify::RecursiveMode;
 use self::notify::Watcher;
 use crate::core::META_FILEPATH;
+use crate::directory::directory::ReadOnlyDirectory;
 use crate::directory::error::LockError;
 use crate::directory::error::{
     DeleteError, IOError, OpenDirectoryError, OpenReadError, OpenWriteError,
@@ -142,7 +143,7 @@ impl MmapCache {
 }
 
 struct WatcherWrapper {
-    _watcher: Mutex<notify::PollWatcher>,
+    _watcher: Mutex<notify::RecommendedWatcher>,
     watcher_router: Arc<WatchCallbackList>,
 }
 
@@ -150,7 +151,7 @@ impl WatcherWrapper {
     pub fn new(path: &Path) -> Result<Self, OpenDirectoryError> {
         let (tx, watcher_recv): (Sender<RawEvent>, Receiver<RawEvent>) = channel();
         // We need to initialize the
-        let watcher = notify::poll::PollWatcher::with_delay_ms(tx, 1)
+        let watcher = notify::raw_watcher(tx)
             .and_then(|mut watcher| {
                 watcher.watch(path, RecursiveMode::Recursive)?;
                 Ok(watcher)
@@ -407,24 +408,6 @@ impl TerminatingWrite for SafeFileWriter {
 }
 
 impl Directory for MmapDirectory {
-    fn open_read(&self, path: &Path) -> result::Result<ReadOnlySource, OpenReadError> {
-        debug!("Open Read {:?}", path);
-        let full_path = self.resolve_path(path);
-
-        let mut mmap_cache = self.inner.mmap_cache.write().map_err(|_| {
-            let msg = format!(
-                "Failed to acquired write lock \
-                 on mmap cache while reading {:?}",
-                path
-            );
-            IOError::with_path(path.to_owned(), make_io_err(msg))
-        })?;
-        Ok(mmap_cache
-            .get_mmap(&full_path)?
-            .map(ReadOnlySource::from)
-            .unwrap_or_else(ReadOnlySource::empty))
-    }
-
     /// Any entry associated to the path in the mmap will be
     /// removed before the file is deleted.
     fn delete(&self, path: &Path) -> result::Result<(), DeleteError> {
@@ -443,11 +426,6 @@ impl Directory for MmapDirectory {
         }
     }
 
-    fn exists(&self, path: &Path) -> bool {
-        let full_path = self.resolve_path(path);
-        full_path.exists()
-    }
-
     fn open_write(&mut self, path: &Path) -> Result<WritePtr, OpenWriteError> {
         debug!("Open Write {:?}", path);
         let full_path = self.resolve_path(path);
@@ -478,25 +456,6 @@ impl Directory for MmapDirectory {
         Ok(BufWriter::new(Box::new(writer)))
     }
 
-    fn atomic_read(&self, path: &Path) -> Result<Vec<u8>, OpenReadError> {
-        let full_path = self.resolve_path(path);
-        let mut buffer = Vec::new();
-        match File::open(&full_path) {
-            Ok(mut file) => {
-                file.read_to_end(&mut buffer)
-                    .map_err(|e| IOError::with_path(path.to_owned(), e))?;
-                Ok(buffer)
-            }
-            Err(e) => {
-                if e.kind() == io::ErrorKind::NotFound {
-                    Err(OpenReadError::FileDoesNotExist(path.to_owned()))
-                } else {
-                    Err(IOError::with_path(path.to_owned(), e).into())
-                }
-            }
-        }
-    }
-
     fn atomic_write(&mut self, path: &Path, data: &[u8]) -> io::Result<()> {
         debug!("Atomic Write {:?}", path);
         let full_path = self.resolve_path(path);
@@ -530,6 +489,50 @@ impl Directory for MmapDirectory {
     }
 }
 
+impl ReadOnlyDirectory for MmapDirectory {
+    fn open_read(&self, path: &Path) -> result::Result<ReadOnlySource, OpenReadError> {
+        debug!("Open Read {:?}", path);
+        let full_path = self.resolve_path(path);
+
+        let mut mmap_cache = self.inner.mmap_cache.write().map_err(|_| {
+            let msg = format!(
+                "Failed to acquired write lock \
+                 on mmap cache while reading {:?}",
+                path
+            );
+            IOError::with_path(path.to_owned(), make_io_err(msg))
+        })?;
+        Ok(mmap_cache
+            .get_mmap(&full_path)?
+            .map(ReadOnlySource::from)
+            .unwrap_or_else(ReadOnlySource::empty))
+    }
+
+    fn exists(&self, path: &Path) -> bool {
+        let full_path = self.resolve_path(path);
+        full_path.exists()
+    }
+
+    fn atomic_read(&self, path: &Path) -> Result<Vec<u8>, OpenReadError> {
+        let full_path = self.resolve_path(path);
+        let mut buffer = Vec::new();
+        match File::open(&full_path) {
+            Ok(mut file) => {
+                file.read_to_end(&mut buffer)
+                    .map_err(|e| IOError::with_path(path.to_owned(), e))?;
+                Ok(buffer)
+            }
+            Err(e) => {
+                if e.kind() == io::ErrorKind::NotFound {
+                    Err(OpenReadError::FileDoesNotExist(path.to_owned()))
+                } else {
+                    Err(IOError::with_path(path.to_owned(), e).into())
+                }
+            }
+        }
+    }
+}
+
 #[cfg(test)]
 mod tests {
 

src/directory/mod.rs

@@ -7,6 +7,7 @@ WORM directory abstraction.
 #[cfg(feature = "mmap")]
 mod mmap_directory;
 
+mod bundle_directory;
 mod directory;
 mod directory_lock;
 mod footer;
@@ -19,7 +20,7 @@ mod watch_event_router;
 pub mod error;
 
 pub use self::directory::DirectoryLock;
-pub use self::directory::{Directory, DirectoryClone};
+pub use self::directory::{Directory, DirectoryClone, ReadOnlyDirectory};
 pub use self::directory_lock::{Lock, INDEX_WRITER_LOCK, META_LOCK};
 pub use self::ram_directory::RAMDirectory;
 pub use self::read_only_source::ReadOnlySource;

src/directory/ram_directory.rs

@@ -1,4 +1,6 @@
+use crate::common::CountingWriter;
 use crate::core::META_FILEPATH;
+use crate::directory::directory::ReadOnlyDirectory;
 use crate::directory::error::{DeleteError, OpenReadError, OpenWriteError};
 use crate::directory::AntiCallToken;
 use crate::directory::WatchCallbackList;
@@ -115,6 +117,22 @@ impl InnerDirectory {
     fn total_mem_usage(&self) -> usize {
         self.fs.values().map(|f| f.len()).sum()
     }
+
+    fn serialize_bundle(&self, wrt: &mut WritePtr) -> io::Result<()> {
+        let mut counting_writer = CountingWriter::wrap(wrt);
+        let mut file_index: HashMap<PathBuf, (u64, u64)> = HashMap::default();
+        for (path, source) in &self.fs {
+            let start = counting_writer.written_bytes();
+            counting_writer.write_all(source.as_slice())?;
+            let stop = counting_writer.written_bytes();
+            file_index.insert(path.to_path_buf(), (start, stop));
+        }
+        let index_offset = counting_writer.written_bytes();
+        serde_json::to_writer(&mut counting_writer, &file_index)?;
+        let index_offset_buffer = index_offset.to_le_bytes();
+        counting_writer.write_all(&index_offset_buffer[..])?;
+        Ok(())
+    }
 }
 
 impl fmt::Debug for RAMDirectory {
@@ -144,13 +162,18 @@ impl RAMDirectory {
     pub fn total_mem_usage(&self) -> usize {
         self.fs.read().unwrap().total_mem_usage()
     }
+
+    /// Serialize the RAMDirectory into a bundle.
+    ///
+    /// This method will fail, write nothing, and return an error if a
+    /// clone of this repository exists.
+    pub fn serialize_bundle(self, wrt: &mut WritePtr) -> io::Result<()> {
+        let inner_directory_rlock = self.fs.read().unwrap();
+        inner_directory_rlock.serialize_bundle(wrt)
+    }
 }
 
 impl Directory for RAMDirectory {
-    fn open_read(&self, path: &Path) -> result::Result<ReadOnlySource, OpenReadError> {
-        self.fs.read().unwrap().open_read(path)
-    }
-
     fn delete(&self, path: &Path) -> result::Result<(), DeleteError> {
         fail_point!("RAMDirectory::delete", |_| {
             use crate::directory::error::IOError;
@@ -160,10 +183,6 @@ impl Directory for RAMDirectory {
         self.fs.write().unwrap().delete(path)
     }
 
-    fn exists(&self, path: &Path) -> bool {
-        self.fs.read().unwrap().exists(path)
-    }
-
     fn open_write(&mut self, path: &Path) -> Result<WritePtr, OpenWriteError> {
         let mut fs = self.fs.write().unwrap();
         let path_buf = PathBuf::from(path);
@@ -177,10 +196,6 @@ impl Directory for RAMDirectory {
         }
     }
 
-    fn atomic_read(&self, path: &Path) -> Result<Vec<u8>, OpenReadError> {
-        Ok(self.open_read(path)?.as_slice().to_owned())
-    }
-
     fn atomic_write(&mut self, path: &Path, data: &[u8]) -> io::Result<()> {
         fail_point!("RAMDirectory::atomic_write", |msg| Err(io::Error::new(
             io::ErrorKind::Other,
@@ -204,3 +219,17 @@ impl Directory for RAMDirectory {
         Ok(self.fs.write().unwrap().watch(watch_callback))
     }
 }
+
+impl ReadOnlyDirectory for RAMDirectory {
+    fn open_read(&self, path: &Path) -> result::Result<ReadOnlySource, OpenReadError> {
+        self.fs.read().unwrap().open_read(path)
+    }
+
+    fn exists(&self, path: &Path) -> bool {
+        self.fs.read().unwrap().exists(path)
+    }
+
+    fn atomic_read(&self, path: &Path) -> Result<Vec<u8>, OpenReadError> {
+        Ok(self.open_read(path)?.as_slice().to_owned())
+    }
+}

src/error.rs

@@ -25,10 +25,10 @@ impl DataCorruption {
         }
     }
 
-    pub fn comment_only(comment: String) -> DataCorruption {
+    pub fn comment_only<TS: ToString>(comment: TS) -> DataCorruption {
         DataCorruption {
             filepath: None,
-            comment,
+            comment: comment.to_string(),
         }
     }
 }

src/fastfield/mod.rs

@@ -179,7 +179,7 @@ mod tests {
 
     use super::*;
     use crate::common::CompositeFile;
-    use crate::directory::{Directory, RAMDirectory, WritePtr};
+    use crate::directory::{Directory, RAMDirectory, ReadOnlyDirectory, WritePtr};
     use crate::fastfield::FastFieldReader;
     use crate::merge_policy::NoMergePolicy;
     use crate::schema::Field;

src/fastfield/multivalued/mod.rs

@@ -7,6 +7,9 @@ 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;
@@ -14,7 +17,6 @@ mod tests {
     use crate::schema::IntOptions;
     use crate::schema::Schema;
     use crate::Index;
-    use chrono::Duration;
 
     #[test]
     fn test_multivalued_u64() {

src/fastfield/reader.rs

@@ -4,7 +4,7 @@ use crate::common::compute_num_bits;
 use crate::common::BinarySerializable;
 use crate::common::CompositeFile;
 use crate::directory::ReadOnlySource;
-use crate::directory::{Directory, RAMDirectory, WritePtr};
+use crate::directory::{Directory, RAMDirectory, ReadOnlyDirectory, WritePtr};
 use crate::fastfield::{FastFieldSerializer, FastFieldsWriter};
 use crate::schema::Schema;
 use crate::schema::FAST;

src/fastfield/readers.rs

@@ -4,6 +4,7 @@ use crate::fastfield::MultiValueIntFastFieldReader;
 use crate::fastfield::{FastFieldNotAvailableError, FastFieldReader};
 use crate::schema::{Cardinality, Field, FieldType, Schema};
 use crate::space_usage::PerFieldSpaceUsage;
+use crate::Result;
 use std::collections::HashMap;
 
 /// Provides access to all of the FastFieldReader.
@@ -53,7 +54,7 @@ impl FastFieldReaders {
     pub(crate) fn load_all(
         schema: &Schema,
         fast_fields_composite: &CompositeFile,
-    ) -> crate::Result<FastFieldReaders> {
+    ) -> Result<FastFieldReaders> {
         let mut fast_field_readers = FastFieldReaders {
             fast_field_i64: Default::default(),
             fast_field_u64: Default::default(),

src/indexer/index_writer.rs

@@ -897,7 +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, max_merge_size: 10000000, min_layer_size: 10000, \
+            "LogMergePolicy { min_merge_size: 8, min_layer_size: 10000, \
              level_log_size: 0.75 }"
         );
         let merge_policy = Box::new(NoMergePolicy::default());

src/indexer/log_merge_policy.rs

@@ -6,14 +6,12 @@ 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,
 }
@@ -28,12 +26,6 @@ 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) {
@@ -61,7 +53,6 @@ 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)>>();
 
@@ -95,7 +86,6 @@ 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,
         }
@@ -114,7 +104,6 @@ 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
     }
@@ -152,11 +141,11 @@ mod tests {
             create_random_segment_meta(10),
             create_random_segment_meta(10),
             create_random_segment_meta(10),
-            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(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(10),
             create_random_segment_meta(10),
             create_random_segment_meta(10),
@@ -193,19 +182,4 @@ 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/merger.rs

@@ -21,6 +21,8 @@ use crate::store::StoreWriter;
 use crate::termdict::TermMerger;
 use crate::termdict::TermOrdinal;
 use crate::DocId;
+use crate::Result;
+use crate::TantivyError;
 use itertools::Itertools;
 use std::cmp;
 use std::collections::HashMap;
@@ -141,7 +143,7 @@ impl DeltaComputer {
 }
 
 impl IndexMerger {
-    pub fn open(schema: Schema, segments: &[Segment]) -> crate::Result<IndexMerger> {
+    pub fn open(schema: Schema, segments: &[Segment]) -> Result<IndexMerger> {
         let mut readers = vec![];
         let mut max_doc: u32 = 0u32;
         for segment in segments {
@@ -157,7 +159,7 @@ impl IndexMerger {
                  which exceeds the limit {}.",
                 max_doc, MAX_DOC_LIMIT
             );
-            return Err(crate::TantivyError::InvalidArgument(err_msg));
+            return Err(TantivyError::InvalidArgument(err_msg));
         }
         Ok(IndexMerger {
             schema,
@@ -166,10 +168,7 @@ impl IndexMerger {
         })
     }
 
-    fn write_fieldnorms(
-        &self,
-        fieldnorms_serializer: &mut FieldNormsSerializer,
-    ) -> crate::Result<()> {
+    fn write_fieldnorms(&self, fieldnorms_serializer: &mut FieldNormsSerializer) -> Result<()> {
         let fields = FieldNormsWriter::fields_with_fieldnorm(&self.schema);
         let mut fieldnorms_data = Vec::with_capacity(self.max_doc as usize);
         for field in fields {
@@ -190,7 +189,7 @@ impl IndexMerger {
         &self,
         fast_field_serializer: &mut FastFieldSerializer,
         mut term_ord_mappings: HashMap<Field, TermOrdinalMapping>,
-    ) -> crate::Result<()> {
+    ) -> Result<()> {
         for (field, field_entry) in self.schema.fields() {
             let field_type = field_entry.field_type();
             match *field_type {
@@ -235,7 +234,7 @@ impl IndexMerger {
         &self,
         field: Field,
         fast_field_serializer: &mut FastFieldSerializer,
-    ) -> crate::Result<()> {
+    ) -> Result<()> {
         let mut u64_readers = vec![];
         let mut min_value = u64::max_value();
         let mut max_value = u64::min_value();
@@ -285,7 +284,7 @@ impl IndexMerger {
         &self,
         field: Field,
         fast_field_serializer: &mut FastFieldSerializer,
-    ) -> crate::Result<()> {
+    ) -> Result<()> {
         let mut total_num_vals = 0u64;
         let mut u64s_readers: Vec<MultiValueIntFastFieldReader<u64>> = Vec::new();
 
@@ -332,7 +331,7 @@ impl IndexMerger {
         field: Field,
         term_ordinal_mappings: &TermOrdinalMapping,
         fast_field_serializer: &mut FastFieldSerializer,
-    ) -> crate::Result<()> {
+    ) -> Result<()> {
         // Multifastfield consists in 2 fastfields.
         // The first serves as an index into the second one and is stricly increasing.
         // The second contains the actual values.
@@ -372,7 +371,7 @@ impl IndexMerger {
         &self,
         field: Field,
         fast_field_serializer: &mut FastFieldSerializer,
-    ) -> crate::Result<()> {
+    ) -> Result<()> {
         // Multifastfield consists in 2 fastfields.
         // The first serves as an index into the second one and is stricly increasing.
         // The second contains the actual values.
@@ -437,7 +436,7 @@ impl IndexMerger {
         &self,
         field: Field,
         fast_field_serializer: &mut FastFieldSerializer,
-    ) -> crate::Result<()> {
+    ) -> Result<()> {
         let mut total_num_vals = 0u64;
         let mut bytes_readers: Vec<BytesFastFieldReader> = Vec::new();
 
@@ -493,7 +492,7 @@ impl IndexMerger {
         indexed_field: Field,
         field_type: &FieldType,
         serializer: &mut InvertedIndexSerializer,
-    ) -> crate::Result<Option<TermOrdinalMapping>> {
+    ) -> Result<Option<TermOrdinalMapping>> {
         let mut positions_buffer: Vec<u32> = Vec::with_capacity(1_000);
         let mut delta_computer = DeltaComputer::new();
         let field_readers = self
@@ -647,7 +646,7 @@ impl IndexMerger {
     fn write_postings(
         &self,
         serializer: &mut InvertedIndexSerializer,
-    ) -> crate::Result<HashMap<Field, TermOrdinalMapping>> {
+    ) -> Result<HashMap<Field, TermOrdinalMapping>> {
         let mut term_ordinal_mappings = HashMap::new();
         for (field, field_entry) in self.schema.fields() {
             if field_entry.is_indexed() {
@@ -661,7 +660,7 @@ impl IndexMerger {
         Ok(term_ordinal_mappings)
     }
 
-    fn write_storable_fields(&self, store_writer: &mut StoreWriter) -> crate::Result<()> {
+    fn write_storable_fields(&self, store_writer: &mut StoreWriter) -> Result<()> {
         for reader in &self.readers {
             let store_reader = reader.get_store_reader();
             if reader.num_deleted_docs() > 0 {
@@ -678,7 +677,7 @@ impl IndexMerger {
 }
 
 impl SerializableSegment for IndexMerger {
-    fn write(&self, mut serializer: SegmentSerializer) -> crate::Result<u32> {
+    fn write(&self, mut serializer: SegmentSerializer) -> Result<u32> {
         let term_ord_mappings = self.write_postings(serializer.get_postings_serializer())?;
         self.write_fieldnorms(serializer.get_fieldnorms_serializer())?;
         self.write_fast_fields(serializer.get_fast_field_serializer(), term_ord_mappings)?;

src/indexer/operation.rs

@@ -19,8 +19,6 @@ 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/prepared_commit.rs

@@ -1,5 +1,6 @@
 use super::IndexWriter;
 use crate::Opstamp;
+use crate::Result;
 use futures::executor::block_on;
 
 /// A prepared commit
@@ -26,11 +27,11 @@ impl<'a> PreparedCommit<'a> {
         self.payload = Some(payload.to_string())
     }
 
-    pub fn abort(self) -> crate::Result<Opstamp> {
+    pub fn abort(self) -> Result<Opstamp> {
         self.index_writer.rollback()
     }
 
-    pub fn commit(self) -> crate::Result<Opstamp> {
+    pub fn commit(self) -> Result<Opstamp> {
         info!("committing {}", self.opstamp);
         let _ = block_on(
             self.index_writer

src/indexer/segment_manager.rs

@@ -4,6 +4,7 @@ use crate::core::SegmentMeta;
 use crate::error::TantivyError;
 use crate::indexer::delete_queue::DeleteCursor;
 use crate::indexer::SegmentEntry;
+use crate::Result as TantivyResult;
 use std::collections::hash_set::HashSet;
 use std::fmt::{self, Debug, Formatter};
 use std::sync::RwLock;
@@ -48,7 +49,7 @@ pub struct SegmentManager {
 }
 
 impl Debug for SegmentManager {
-    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
+    fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), fmt::Error> {
         let lock = self.read();
         write!(
             f,
@@ -144,7 +145,7 @@ impl SegmentManager {
     /// Returns an error if some segments are missing, or if
     /// the `segment_ids` are not either all committed or all
     /// uncommitted.
-    pub fn start_merge(&self, segment_ids: &[SegmentId]) -> crate::Result<Vec<SegmentEntry>> {
+    pub fn start_merge(&self, segment_ids: &[SegmentId]) -> TantivyResult<Vec<SegmentEntry>> {
         let registers_lock = self.read();
         let mut segment_entries = vec![];
         if registers_lock.uncommitted.contains_all(segment_ids) {
@@ -187,7 +188,7 @@ impl SegmentManager {
             .segments_status(before_merge_segment_ids)
             .ok_or_else(|| {
                 warn!("couldn't find segment in SegmentManager");
-                crate::TantivyError::InvalidArgument(
+                crate::Error::InvalidArgument(
                     "The segments that were merged could not be found in the SegmentManager. \
                      This is not necessarily a bug, and can happen after a rollback for instance."
                         .to_string(),

src/indexer/segment_serializer.rs

@@ -1,8 +1,13 @@
+use crate::Directory;
+
 use crate::core::Segment;
 use crate::core::SegmentComponent;
+use crate::directory::error::OpenWriteError;
+use crate::directory::{DirectoryClone, RAMDirectory, TerminatingWrite, WritePtr};
 use crate::fastfield::FastFieldSerializer;
 use crate::fieldnorm::FieldNormsSerializer;
 use crate::postings::InvertedIndexSerializer;
+use crate::schema::Schema;
 use crate::store::StoreWriter;
 
 /// Segment serializer is in charge of laying out on disk
@@ -12,25 +17,50 @@ pub struct SegmentSerializer {
     fast_field_serializer: FastFieldSerializer,
     fieldnorms_serializer: FieldNormsSerializer,
     postings_serializer: InvertedIndexSerializer,
+    bundle_writer: Option<(RAMDirectory, WritePtr)>,
+}
+
+pub(crate) struct SegmentSerializerWriters {
+    postings_wrt: WritePtr,
+    positions_skip_wrt: WritePtr,
+    positions_wrt: WritePtr,
+    terms_wrt: WritePtr,
+    fast_field_wrt: WritePtr,
+    fieldnorms_wrt: WritePtr,
+    store_wrt: WritePtr,
+}
+
+impl SegmentSerializerWriters {
+    pub(crate) fn for_segment(segment: &mut Segment) -> Result<Self, OpenWriteError> {
+        Ok(SegmentSerializerWriters {
+            postings_wrt: segment.open_write(SegmentComponent::POSTINGS)?,
+            positions_skip_wrt: segment.open_write(SegmentComponent::POSITIONS)?,
+            positions_wrt: segment.open_write(SegmentComponent::POSITIONSSKIP)?,
+            terms_wrt: segment.open_write(SegmentComponent::TERMS)?,
+            fast_field_wrt: segment.open_write(SegmentComponent::FASTFIELDS)?,
+            fieldnorms_wrt: segment.open_write(SegmentComponent::FIELDNORMS)?,
+            store_wrt: segment.open_write(SegmentComponent::STORE)?,
+        })
+    }
 }
 
 impl SegmentSerializer {
-    /// Creates a new `SegmentSerializer`.
-    pub fn for_segment(segment: &mut Segment) -> crate::Result<SegmentSerializer> {
-        let store_write = segment.open_write(SegmentComponent::STORE)?;
-
-        let fast_field_write = segment.open_write(SegmentComponent::FASTFIELDS)?;
-        let fast_field_serializer = FastFieldSerializer::from_write(fast_field_write)?;
-
-        let fieldnorms_write = segment.open_write(SegmentComponent::FIELDNORMS)?;
-        let fieldnorms_serializer = FieldNormsSerializer::from_write(fieldnorms_write)?;
-
-        let postings_serializer = InvertedIndexSerializer::open(segment)?;
+    pub(crate) fn new(schema: Schema, writers: SegmentSerializerWriters) -> crate::Result<Self> {
+        let fast_field_serializer = FastFieldSerializer::from_write(writers.fast_field_wrt)?;
+        let fieldnorms_serializer = FieldNormsSerializer::from_write(writers.fieldnorms_wrt)?;
+        let postings_serializer = InvertedIndexSerializer::open(
+            schema,
+            writers.terms_wrt,
+            writers.postings_wrt,
+            writers.positions_wrt,
+            writers.positions_skip_wrt,
+        );
         Ok(SegmentSerializer {
-            store_writer: StoreWriter::new(store_write),
+            store_writer: StoreWriter::new(writers.store_wrt),
             fast_field_serializer,
             fieldnorms_serializer,
             postings_serializer,
+            bundle_writer: None,
         })
     }
 
@@ -55,11 +85,15 @@ impl SegmentSerializer {
     }
 
     /// Finalize the segment serialization.
-    pub fn close(self) -> crate::Result<()> {
+    pub fn close(mut self) -> crate::Result<()> {
         self.fast_field_serializer.close()?;
         self.postings_serializer.close()?;
         self.store_writer.close()?;
         self.fieldnorms_serializer.close()?;
+        if let Some((ram_directory, mut bundle_wrt)) = self.bundle_writer.take() {
+            ram_directory.serialize_bundle(&mut bundle_wrt)?;
+            bundle_wrt.terminate()?;
+        }
         Ok(())
     }
 }

src/indexer/segment_updater.rs

@@ -12,6 +12,7 @@ use crate::indexer::index_writer::advance_deletes;
 use crate::indexer::merge_operation::MergeOperationInventory;
 use crate::indexer::merger::IndexMerger;
 use crate::indexer::segment_manager::SegmentsStatus;
+use crate::indexer::segment_serializer::SegmentSerializerWriters;
 use crate::indexer::stamper::Stamper;
 use crate::indexer::SegmentEntry;
 use crate::indexer::SegmentSerializer;
@@ -132,7 +133,9 @@ fn merge(
     let merger: IndexMerger = IndexMerger::open(index.schema(), &segments[..])?;
 
     // ... we just serialize this index merger in our new segment to merge the two segments.
-    let segment_serializer = SegmentSerializer::for_segment(&mut merged_segment)?;
+    let segment_serializer_wrts = SegmentSerializerWriters::for_segment(&mut merged_segment)?;
+    let segment_serializer =
+        SegmentSerializer::new(merged_segment.schema(), segment_serializer_wrts)?;
 
     let num_docs = merger.write(segment_serializer)?;
 
@@ -173,18 +176,14 @@ impl SegmentUpdater {
             .pool_size(1)
             .create()
             .map_err(|_| {
-                crate::TantivyError::SystemError(
-                    "Failed to spawn segment updater thread".to_string(),
-                )
+                crate::Error::SystemError("Failed to spawn segment updater thread".to_string())
             })?;
         let merge_thread_pool = ThreadPoolBuilder::new()
             .name_prefix("merge_thread")
             .pool_size(NUM_MERGE_THREADS)
             .create()
             .map_err(|_| {
-                crate::TantivyError::SystemError(
-                    "Failed to spawn segment merging thread".to_string(),
-                )
+                crate::Error::SystemError("Failed to spawn segment merging thread".to_string())
             })?;
         let index_meta = index.load_metas()?;
         Ok(SegmentUpdater(Arc::new(InnerSegmentUpdater {
@@ -226,7 +225,7 @@ impl SegmentUpdater {
         receiver.unwrap_or_else(|_| {
             let err_msg =
                 "A segment_updater future did not success. This should never happen.".to_string();
-            Err(crate::TantivyError::SystemError(err_msg))
+            Err(crate::Error::SystemError(err_msg))
         })
     }
 
@@ -423,7 +422,7 @@ impl SegmentUpdater {
         });
 
         Ok(merging_future_recv
-            .unwrap_or_else(|_| Err(crate::TantivyError::SystemError("Merge failed".to_string()))))
+            .unwrap_or_else(|_| Err(crate::Error::SystemError("Merge failed".to_string()))))
     }
 
     async fn consider_merge_options(&self) {

src/indexer/segment_writer.rs

@@ -3,7 +3,7 @@ use crate::core::Segment;
 use crate::core::SerializableSegment;
 use crate::fastfield::FastFieldsWriter;
 use crate::fieldnorm::FieldNormsWriter;
-use crate::indexer::segment_serializer::SegmentSerializer;
+use crate::indexer::segment_serializer::{SegmentSerializer, SegmentSerializerWriters};
 use crate::postings::compute_table_size;
 use crate::postings::MultiFieldPostingsWriter;
 use crate::schema::FieldType;
@@ -11,18 +11,21 @@ use crate::schema::Schema;
 use crate::schema::Term;
 use crate::schema::Value;
 use crate::schema::{Field, FieldEntry};
-use crate::tokenizer::{BoxTokenStream, PreTokenizedStream};
-use crate::tokenizer::{FacetTokenizer, TextAnalyzer};
-use crate::tokenizer::{TokenStreamChain, Tokenizer};
+use crate::tokenizer::BoxedTokenizer;
+use crate::tokenizer::FacetTokenizer;
+use crate::tokenizer::PreTokenizedStream;
+use crate::tokenizer::{TokenStream, TokenStreamChain, Tokenizer};
 use crate::DocId;
 use crate::Opstamp;
+use crate::Result;
+use crate::TantivyError;
 use std::io;
 use std::str;
 
 /// Computes the initial size of the hash table.
 ///
 /// Returns a number of bit `b`, such that the recommended initial table size is 2^b.
-fn initial_table_size(per_thread_memory_budget: usize) -> crate::Result<usize> {
+fn initial_table_size(per_thread_memory_budget: usize) -> Result<usize> {
     let table_memory_upper_bound = per_thread_memory_budget / 3;
     if let Some(limit) = (10..)
         .take_while(|num_bits: &usize| compute_table_size(*num_bits) < table_memory_upper_bound)
@@ -30,7 +33,7 @@ fn initial_table_size(per_thread_memory_budget: usize) -> crate::Result<usize> {
     {
         Ok(limit.min(19)) // we cap it at 2^19 = 512K.
     } else {
-        Err(crate::TantivyError::InvalidArgument(
+        Err(TantivyError::InvalidArgument(
             format!("per thread memory budget (={}) is too small. Raise the memory budget or lower the number of threads.", per_thread_memory_budget)))
     }
 }
@@ -47,7 +50,7 @@ pub struct SegmentWriter {
     fast_field_writers: FastFieldsWriter,
     fieldnorms_writer: FieldNormsWriter,
     doc_opstamps: Vec<Opstamp>,
-    tokenizers: Vec<Option<TextAnalyzer>>,
+    tokenizers: Vec<Option<BoxedTokenizer>>,
 }
 
 impl SegmentWriter {
@@ -64,9 +67,10 @@ impl SegmentWriter {
         memory_budget: usize,
         mut segment: Segment,
         schema: &Schema,
-    ) -> crate::Result<SegmentWriter> {
+    ) -> Result<SegmentWriter> {
         let table_num_bits = initial_table_size(memory_budget)?;
-        let segment_serializer = SegmentSerializer::for_segment(&mut segment)?;
+        let segment_serializer_wrts = SegmentSerializerWriters::for_segment(&mut segment)?;
+        let segment_serializer = SegmentSerializer::new(segment.schema(), segment_serializer_wrts)?;
         let multifield_postings = MultiFieldPostingsWriter::new(schema, table_num_bits);
         let tokenizers = schema
             .fields()
@@ -97,7 +101,7 @@ impl SegmentWriter {
     ///
     /// Finalize consumes the `SegmentWriter`, so that it cannot
     /// be used afterwards.
-    pub fn finalize(mut self) -> crate::Result<Vec<u64>> {
+    pub fn finalize(mut self) -> Result<Vec<u64>> {
         self.fieldnorms_writer.fill_up_to_max_doc(self.max_doc);
         write(
             &self.multifield_postings,
@@ -156,7 +160,7 @@ impl SegmentWriter {
                     }
                 }
                 FieldType::Str(_) => {
-                    let mut token_streams: Vec<BoxTokenStream> = vec![];
+                    let mut token_streams: Vec<Box<dyn TokenStream>> = vec![];
                     let mut offsets = vec![];
                     let mut total_offset = 0;
 
@@ -169,7 +173,7 @@ impl SegmentWriter {
                                 }
 
                                 token_streams
-                                    .push(PreTokenizedStream::from(tok_str.clone()).into());
+                                    .push(Box::new(PreTokenizedStream::from(tok_str.clone())));
                             }
                             Value::Str(ref text) => {
                                 if let Some(ref mut tokenizer) =
@@ -188,7 +192,8 @@ impl SegmentWriter {
                     let num_tokens = if token_streams.is_empty() {
                         0
                     } else {
-                        let mut token_stream = TokenStreamChain::new(offsets, token_streams);
+                        let mut token_stream: Box<dyn TokenStream> =
+                            Box::new(TokenStreamChain::new(offsets, token_streams));
                         self.multifield_postings
                             .index_text(doc_id, field, &mut token_stream)
                     };
@@ -279,7 +284,7 @@ fn write(
     fast_field_writers: &FastFieldsWriter,
     fieldnorms_writer: &FieldNormsWriter,
     mut serializer: SegmentSerializer,
-) -> crate::Result<()> {
+) -> Result<()> {
     let term_ord_map = multifield_postings.serialize(serializer.get_postings_serializer())?;
     fast_field_writers.serialize(serializer.get_fast_field_serializer(), &term_ord_map)?;
     fieldnorms_writer.serialize(serializer.get_fieldnorms_serializer())?;
@@ -288,7 +293,7 @@ fn write(
 }
 
 impl SerializableSegment for SegmentWriter {
-    fn write(&self, serializer: SegmentSerializer) -> crate::Result<u32> {
+    fn write(&self, serializer: SegmentSerializer) -> Result<u32> {
         let max_doc = self.max_doc;
         write(
             &self.multifield_postings,

src/indexer/stamper.rs

@@ -1,76 +1,18 @@
 use crate::Opstamp;
 use std::ops::Range;
-use std::sync::atomic::Ordering;
+use std::sync::atomic::{AtomicU64, 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<AtomicU64Wrapper>);
+pub struct Stamper(Arc<AtomicU64>);
 
 impl Stamper {
     pub fn new(first_opstamp: Opstamp) -> Stamper {
-        Stamper(Arc::new(AtomicU64Wrapper::new(first_opstamp)))
+        Stamper(Arc::new(AtomicU64::new(first_opstamp)))
     }
 
     pub fn stamp(&self) -> Opstamp {
@@ -89,7 +31,8 @@ impl Stamper {
 
     /// Reverts the stamper to a given `Opstamp` value and returns it
     pub fn revert(&self, to_opstamp: Opstamp) -> Opstamp {
-        self.0.revert(to_opstamp, Ordering::SeqCst)
+        self.0.store(to_opstamp, Ordering::SeqCst);
+        to_opstamp
     }
 }
 

src/lib.rs

@@ -121,13 +121,13 @@ mod functional_test;
 mod macros;
 
 pub use crate::error::TantivyError;
+
+#[deprecated(since = "0.7.0", note = "please use `tantivy::TantivyError` instead")]
+pub use crate::error::TantivyError as Error;
 pub use chrono;
 
 /// Tantivy result.
-///
-/// Within tantivy, please avoid importing `Result` using `use crate::Result`
-/// and instead, refer to this as `crate::Result<T>`.
-pub type Result<T> = std::result::Result<T, TantivyError>;
+pub type Result<T> = std::result::Result<T, error::TantivyError>;
 
 /// Tantivy DateTime
 pub type DateTime = chrono::DateTime<chrono::Utc>;
@@ -161,11 +161,10 @@ 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::{Executor, SegmentComponent};
+pub use crate::core::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;

src/postings/mod.rs

@@ -75,7 +75,7 @@ pub mod tests {
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
         let mut segment = index.new_segment();
-        let mut posting_serializer = InvertedIndexSerializer::open(&mut segment).unwrap();
+        let mut posting_serializer = InvertedIndexSerializer::for_segment(&mut segment).unwrap();
         {
             let mut field_serializer = posting_serializer.new_field(text_field, 120 * 4).unwrap();
             field_serializer.new_term("abc".as_bytes()).unwrap();

src/postings/postings_writer.rs

@@ -11,6 +11,7 @@ use crate::termdict::TermOrdinal;
 use crate::tokenizer::TokenStream;
 use crate::tokenizer::{Token, MAX_TOKEN_LEN};
 use crate::DocId;
+use crate::Result;
 use fnv::FnvHashMap;
 use std::collections::HashMap;
 use std::io;
@@ -128,7 +129,7 @@ impl MultiFieldPostingsWriter {
     pub fn serialize(
         &self,
         serializer: &mut InvertedIndexSerializer,
-    ) -> crate::Result<HashMap<Field, FnvHashMap<UnorderedTermId, TermOrdinal>>> {
+    ) -> Result<HashMap<Field, FnvHashMap<UnorderedTermId, TermOrdinal>>> {
         let mut term_offsets: Vec<(&[u8], Addr, UnorderedTermId)> =
             self.term_index.iter().collect();
         term_offsets.sort_unstable_by_key(|&(k, _, _)| k);

src/postings/serializer.rs

@@ -10,7 +10,8 @@ use crate::postings::USE_SKIP_INFO_LIMIT;
 use crate::schema::Schema;
 use crate::schema::{Field, FieldEntry, FieldType};
 use crate::termdict::{TermDictionaryBuilder, TermOrdinal};
-use crate::DocId;
+use crate::Result;
+use crate::{Directory, DocId};
 use std::io::{self, Write};
 
 /// `InvertedIndexSerializer` is in charge of serializing
@@ -53,33 +54,36 @@ pub struct InvertedIndexSerializer {
 }
 
 impl InvertedIndexSerializer {
-    /// Open a new `InvertedIndexSerializer` for the given segment
-    fn create(
-        terms_write: CompositeWrite<WritePtr>,
-        postings_write: CompositeWrite<WritePtr>,
-        positions_write: CompositeWrite<WritePtr>,
-        positionsidx_write: CompositeWrite<WritePtr>,
-        schema: Schema,
-    ) -> crate::Result<InvertedIndexSerializer> {
-        Ok(InvertedIndexSerializer {
-            terms_write,
-            postings_write,
-            positions_write,
-            positionsidx_write,
+    pub(crate) fn for_segment(segment: &mut Segment) -> crate::Result<Self> {
+        let schema = segment.schema();
+        use crate::core::SegmentComponent;
+        let terms_wrt = segment.open_write(SegmentComponent::TERMS)?;
+        let postings_wrt = segment.open_write(SegmentComponent::POSTINGS)?;
+        let positions_wrt = segment.open_write(SegmentComponent::POSITIONS)?;
+        let positions_idx_wrt = segment.open_write(SegmentComponent::POSITIONSSKIP)?;
+        Ok(Self::open(
             schema,
-        })
+            terms_wrt,
+            postings_wrt,
+            positions_wrt,
+            positions_idx_wrt,
+        ))
     }
-
     /// Open a new `PostingsSerializer` for the given segment
-    pub fn open(segment: &mut Segment) -> crate::Result<InvertedIndexSerializer> {
-        use crate::SegmentComponent::{POSITIONS, POSITIONSSKIP, POSTINGS, TERMS};
-        InvertedIndexSerializer::create(
-            CompositeWrite::wrap(segment.open_write(TERMS)?),
-            CompositeWrite::wrap(segment.open_write(POSTINGS)?),
-            CompositeWrite::wrap(segment.open_write(POSITIONS)?),
-            CompositeWrite::wrap(segment.open_write(POSITIONSSKIP)?),
-            segment.schema(),
-        )
+    pub(crate) fn open(
+        schema: Schema,
+        terms_wrt: WritePtr,
+        postings_wrt: WritePtr,
+        positions_wrt: WritePtr,
+        positions_idx_wrt: WritePtr,
+    ) -> InvertedIndexSerializer {
+        InvertedIndexSerializer {
+            terms_write: CompositeWrite::wrap(terms_wrt),
+            postings_write: CompositeWrite::wrap(postings_wrt),
+            positions_write: CompositeWrite::wrap(positions_wrt),
+            positionsidx_write: CompositeWrite::wrap(positions_idx_wrt),
+            schema,
+        }
     }
 
     /// Must be called before starting pushing terms of
@@ -147,7 +151,8 @@ impl<'a> FieldSerializer<'a> {
             }
             _ => (false, false),
         };
-        let term_dictionary_builder = TermDictionaryBuilder::create(term_dictionary_write)?;
+        let term_dictionary_builder =
+            TermDictionaryBuilder::create(term_dictionary_write, &field_type)?;
         let postings_serializer =
             PostingsSerializer::new(postings_write, term_freq_enabled, position_enabled);
         let positions_serializer_opt = if position_enabled {

src/query/all_query.rs

@@ -4,6 +4,7 @@ use crate::docset::DocSet;
 use crate::query::explanation::does_not_match;
 use crate::query::{Explanation, Query, Scorer, Weight};
 use crate::DocId;
+use crate::Result;
 use crate::Score;
 
 /// Query that matches all of the documents.
@@ -13,7 +14,7 @@ use crate::Score;
 pub struct AllQuery;
 
 impl Query for AllQuery {
-    fn weight(&self, _: &Searcher, _: bool) -> crate::Result<Box<dyn Weight>> {
+    fn weight(&self, _: &Searcher, _: bool) -> Result<Box<dyn Weight>> {
         Ok(Box::new(AllWeight))
     }
 }
@@ -22,7 +23,7 @@ impl Query for AllQuery {
 pub struct AllWeight;
 
 impl Weight for AllWeight {
-    fn scorer(&self, reader: &SegmentReader) -> crate::Result<Box<dyn Scorer>> {
+    fn scorer(&self, reader: &SegmentReader) -> Result<Box<dyn Scorer>> {
         Ok(Box::new(AllScorer {
             state: State::NotStarted,
             doc: 0u32,
@@ -30,7 +31,7 @@ impl Weight for AllWeight {
         }))
     }
 
-    fn explain(&self, reader: &SegmentReader, doc: DocId) -> crate::Result<Explanation> {
+    fn explain(&self, reader: &SegmentReader, doc: DocId) -> Result<Explanation> {
         if doc >= reader.max_doc() {
             return Err(does_not_match(doc));
         }

src/query/boolean_query/boolean_query.rs

@@ -5,6 +5,7 @@ use crate::query::TermQuery;
 use crate::query::Weight;
 use crate::schema::IndexRecordOption;
 use crate::schema::Term;
+use crate::Result;
 use crate::Searcher;
 use std::collections::BTreeSet;
 
@@ -29,9 +30,9 @@ use std::collections::BTreeSet;
 ///use tantivy::query::{BooleanQuery, Occur, PhraseQuery, Query, TermQuery};
 ///use tantivy::schema::{IndexRecordOption, Schema, TEXT};
 ///use tantivy::Term;
-///use tantivy::Index;
+///use tantivy::{Index, Result};
 ///
-///fn main() -> tantivy::Result<()> {
+///fn main() -> Result<()> {
 ///    let mut schema_builder = Schema::builder();
 ///    let title = schema_builder.add_text_field("title", TEXT);
 ///    let body = schema_builder.add_text_field("body", TEXT);
@@ -148,14 +149,14 @@ impl From<Vec<(Occur, Box<dyn Query>)>> for BooleanQuery {
 }
 
 impl Query for BooleanQuery {
-    fn weight(&self, searcher: &Searcher, scoring_enabled: bool) -> crate::Result<Box<dyn Weight>> {
+    fn weight(&self, searcher: &Searcher, scoring_enabled: bool) -> Result<Box<dyn Weight>> {
         let sub_weights = self
             .subqueries
             .iter()
             .map(|&(ref occur, ref subquery)| {
                 Ok((*occur, subquery.weight(searcher, scoring_enabled)?))
             })
-            .collect::<crate::Result<_>>()?;
+            .collect::<Result<_>>()?;
         Ok(Box::new(BooleanWeight::new(sub_weights, scoring_enabled)))
     }
 

src/query/boolean_query/boolean_weight.rs

@@ -10,6 +10,7 @@ use crate::query::Scorer;
 use crate::query::Union;
 use crate::query::Weight;
 use crate::query::{intersect_scorers, Explanation};
+use crate::Result;
 use crate::{DocId, SkipResult};
 use std::collections::HashMap;
 
@@ -55,7 +56,7 @@ impl BooleanWeight {
     fn per_occur_scorers(
         &self,
         reader: &SegmentReader,
-    ) -> crate::Result<HashMap<Occur, Vec<Box<dyn Scorer>>>> {
+    ) -> Result<HashMap<Occur, Vec<Box<dyn Scorer>>>> {
         let mut per_occur_scorers: HashMap<Occur, Vec<Box<dyn Scorer>>> = HashMap::new();
         for &(ref occur, ref subweight) in &self.weights {
             let sub_scorer: Box<dyn Scorer> = subweight.scorer(reader)?;
@@ -70,7 +71,7 @@ impl BooleanWeight {
     fn complex_scorer<TScoreCombiner: ScoreCombiner>(
         &self,
         reader: &SegmentReader,
-    ) -> crate::Result<Box<dyn Scorer>> {
+    ) -> Result<Box<dyn Scorer>> {
         let mut per_occur_scorers = self.per_occur_scorers(reader)?;
 
         let should_scorer_opt: Option<Box<dyn Scorer>> = per_occur_scorers
@@ -112,7 +113,7 @@ impl BooleanWeight {
 }
 
 impl Weight for BooleanWeight {
-    fn scorer(&self, reader: &SegmentReader) -> crate::Result<Box<dyn Scorer>> {
+    fn scorer(&self, reader: &SegmentReader) -> Result<Box<dyn Scorer>> {
         if self.weights.is_empty() {
             Ok(Box::new(EmptyScorer))
         } else if self.weights.len() == 1 {
@@ -129,7 +130,7 @@ impl Weight for BooleanWeight {
         }
     }
 
-    fn explain(&self, reader: &SegmentReader, doc: DocId) -> crate::Result<Explanation> {
+    fn explain(&self, reader: &SegmentReader, doc: DocId) -> Result<Explanation> {
         let mut scorer = self.scorer(reader)?;
         if scorer.skip_next(doc) != SkipResult::Reached {
             return Err(does_not_match(doc));

src/query/empty_query.rs

@@ -4,6 +4,7 @@ use crate::query::Weight;
 use crate::query::{Explanation, Query};
 use crate::DocId;
 use crate::DocSet;
+use crate::Result;
 use crate::Score;
 use crate::Searcher;
 use crate::SegmentReader;
@@ -15,15 +16,11 @@ use crate::SegmentReader;
 pub struct EmptyQuery;
 
 impl Query for EmptyQuery {
-    fn weight(
-        &self,
-        _searcher: &Searcher,
-        _scoring_enabled: bool,
-    ) -> crate::Result<Box<dyn Weight>> {
+    fn weight(&self, _searcher: &Searcher, _scoring_enabled: bool) -> Result<Box<dyn Weight>> {
         Ok(Box::new(EmptyWeight))
     }
 
-    fn count(&self, _searcher: &Searcher) -> crate::Result<usize> {
+    fn count(&self, _searcher: &Searcher) -> Result<usize> {
         Ok(0)
     }
 }
@@ -33,11 +30,11 @@ impl Query for EmptyQuery {
 /// It is useful for tests and handling edge cases.
 pub struct EmptyWeight;
 impl Weight for EmptyWeight {
-    fn scorer(&self, _reader: &SegmentReader) -> crate::Result<Box<dyn Scorer>> {
+    fn scorer(&self, _reader: &SegmentReader) -> Result<Box<dyn Scorer>> {
         Ok(Box::new(EmptyScorer))
     }
 
-    fn explain(&self, _reader: &SegmentReader, doc: DocId) -> crate::Result<Explanation> {
+    fn explain(&self, _reader: &SegmentReader, doc: DocId) -> Result<Explanation> {
         Err(does_not_match(doc))
     }
 }

src/query/fuzzy_query.rs

@@ -1,7 +1,8 @@
+use crate::error::TantivyError::InvalidArgument;
 use crate::query::{AutomatonWeight, Query, Weight};
 use crate::schema::Term;
+use crate::Result;
 use crate::Searcher;
-use crate::TantivyError::InvalidArgument;
 use levenshtein_automata::{LevenshteinAutomatonBuilder, DFA};
 use once_cell::sync::Lazy;
 use std::collections::HashMap;
@@ -30,9 +31,9 @@ static LEV_BUILDER: Lazy<HashMap<(u8, bool), LevenshteinAutomatonBuilder>> = Laz
 /// use tantivy::collector::{Count, TopDocs};
 /// use tantivy::query::FuzzyTermQuery;
 /// use tantivy::schema::{Schema, TEXT};
-/// use tantivy::{doc, Index, Term};
+/// use tantivy::{doc, Index, Result, Term};
 ///
-/// fn example() -> tantivy::Result<()> {
+/// fn example() -> Result<()> {
 ///     let mut schema_builder = Schema::builder();
 ///     let title = schema_builder.add_text_field("title", TEXT);
 ///     let schema = schema_builder.build();
@@ -101,7 +102,7 @@ impl FuzzyTermQuery {
         }
     }
 
-    fn specialized_weight(&self) -> crate::Result<AutomatonWeight<DFA>> {
+    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)
@@ -118,11 +119,7 @@ impl FuzzyTermQuery {
 }
 
 impl Query for FuzzyTermQuery {
-    fn weight(
-        &self,
-        _searcher: &Searcher,
-        _scoring_enabled: bool,
-    ) -> crate::Result<Box<dyn Weight>> {
+    fn weight(&self, _searcher: &Searcher, _scoring_enabled: bool) -> Result<Box<dyn Weight>> {
         Ok(Box::new(self.specialized_weight()?))
     }
 }

src/query/phrase_query/phrase_query.rs

@@ -1,10 +1,12 @@
 use super::PhraseWeight;
 use crate::core::searcher::Searcher;
+use crate::error::TantivyError;
 use crate::query::bm25::BM25Weight;
 use crate::query::Query;
 use crate::query::Weight;
 use crate::schema::IndexRecordOption;
 use crate::schema::{Field, Term};
+use crate::Result;
 use std::collections::BTreeSet;
 
 /// `PhraseQuery` matches a specific sequence of words.
@@ -79,7 +81,7 @@ impl PhraseQuery {
         &self,
         searcher: &Searcher,
         scoring_enabled: bool,
-    ) -> crate::Result<PhraseWeight> {
+    ) -> Result<PhraseWeight> {
         let schema = searcher.schema();
         let field_entry = schema.get_field_entry(self.field);
         let has_positions = field_entry
@@ -89,7 +91,7 @@ impl PhraseQuery {
             .unwrap_or(false);
         if !has_positions {
             let field_name = field_entry.name();
-            return Err(crate::TantivyError::SchemaError(format!(
+            return Err(TantivyError::SchemaError(format!(
                 "Applied phrase query on field {:?}, which does not have positions indexed",
                 field_name
             )));
@@ -108,7 +110,7 @@ impl Query for PhraseQuery {
     /// Create the weight associated to a query.
     ///
     /// See [`Weight`](./trait.Weight.html).
-    fn weight(&self, searcher: &Searcher, scoring_enabled: bool) -> crate::Result<Box<dyn Weight>> {
+    fn weight(&self, searcher: &Searcher, scoring_enabled: bool) -> Result<Box<dyn Weight>> {
         let phrase_weight = self.phrase_weight(searcher, scoring_enabled)?;
         Ok(Box::new(phrase_weight))
     }

src/query/query.rs

@@ -2,6 +2,7 @@ use super::Weight;
 use crate::core::searcher::Searcher;
 use crate::query::Explanation;
 use crate::DocAddress;
+use crate::Result;
 use crate::Term;
 use downcast_rs::impl_downcast;
 use std::collections::BTreeSet;
@@ -47,17 +48,17 @@ pub trait Query: QueryClone + downcast_rs::Downcast + fmt::Debug {
     /// can increase performances.
     ///
     /// See [`Weight`](./trait.Weight.html).
-    fn weight(&self, searcher: &Searcher, scoring_enabled: bool) -> crate::Result<Box<dyn Weight>>;
+    fn weight(&self, searcher: &Searcher, scoring_enabled: bool) -> Result<Box<dyn Weight>>;
 
     /// Returns an `Explanation` for the score of the document.
-    fn explain(&self, searcher: &Searcher, doc_address: DocAddress) -> crate::Result<Explanation> {
+    fn explain(&self, searcher: &Searcher, doc_address: DocAddress) -> Result<Explanation> {
         let reader = searcher.segment_reader(doc_address.segment_ord());
         let weight = self.weight(searcher, true)?;
         weight.explain(reader, doc_address.doc())
     }
 
     /// Returns the number of documents matching the query.
-    fn count(&self, searcher: &Searcher) -> crate::Result<usize> {
+    fn count(&self, searcher: &Searcher) -> Result<usize> {
         let weight = self.weight(searcher, false)?;
         let mut result = 0;
         for reader in searcher.segment_readers() {
@@ -85,11 +86,11 @@ where
 }
 
 impl Query for Box<dyn Query> {
-    fn weight(&self, searcher: &Searcher, scoring_enabled: bool) -> crate::Result<Box<dyn Weight>> {
+    fn weight(&self, searcher: &Searcher, scoring_enabled: bool) -> Result<Box<dyn Weight>> {
         self.as_ref().weight(searcher, scoring_enabled)
     }
 
-    fn count(&self, searcher: &Searcher) -> crate::Result<usize> {
+    fn count(&self, searcher: &Searcher) -> Result<usize> {
         self.as_ref().count(searcher)
     }
 

src/query/query_parser/query_parser.rs

@@ -533,7 +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, TextAnalyzer, TokenizerManager,
+        LowerCaser, SimpleTokenizer, StopWordFilter, Tokenizer, TokenizerManager,
     };
     use crate::Index;
     use matches::assert_matches;
@@ -563,7 +563,7 @@ mod test {
         let tokenizer_manager = TokenizerManager::default();
         tokenizer_manager.register(
             "en_with_stop_words",
-            TextAnalyzer::from(SimpleTokenizer)
+            SimpleTokenizer
                 .filter(LowerCaser)
                 .filter(StopWordFilter::remove(vec!["the".to_string()])),
         );

src/query/regex_query.rs

@@ -1,6 +1,7 @@
 use crate::error::TantivyError;
 use crate::query::{AutomatonWeight, Query, Weight};
 use crate::schema::Field;
+use crate::Result;
 use crate::Searcher;
 use std::clone::Clone;
 use std::sync::Arc;
@@ -57,7 +58,7 @@ pub struct RegexQuery {
 
 impl RegexQuery {
     /// Creates a new RegexQuery from a given pattern
-    pub fn from_pattern(regex_pattern: &str, field: Field) -> crate::Result<Self> {
+    pub fn from_pattern(regex_pattern: &str, field: Field) -> Result<Self> {
         let regex = Regex::new(&regex_pattern)
             .map_err(|_| TantivyError::InvalidArgument(regex_pattern.to_string()))?;
         Ok(RegexQuery::from_regex(regex, field))
@@ -77,11 +78,7 @@ impl RegexQuery {
 }
 
 impl Query for RegexQuery {
-    fn weight(
-        &self,
-        _searcher: &Searcher,
-        _scoring_enabled: bool,
-    ) -> crate::Result<Box<dyn Weight>> {
+    fn weight(&self, _searcher: &Searcher, _scoring_enabled: bool) -> Result<Box<dyn Weight>> {
         Ok(Box::new(self.specialized_weight()))
     }
 }

src/query/term_query/term_query.rs

@@ -3,6 +3,7 @@ use crate::query::bm25::BM25Weight;
 use crate::query::Query;
 use crate::query::Weight;
 use crate::schema::IndexRecordOption;
+use crate::Result;
 use crate::Searcher;
 use crate::Term;
 use std::collections::BTreeSet;
@@ -100,7 +101,7 @@ impl TermQuery {
 }
 
 impl Query for TermQuery {
-    fn weight(&self, searcher: &Searcher, scoring_enabled: bool) -> crate::Result<Box<dyn Weight>> {
+    fn weight(&self, searcher: &Searcher, scoring_enabled: bool) -> Result<Box<dyn Weight>> {
         Ok(Box::new(self.specialized_weight(searcher, scoring_enabled)))
     }
     fn query_terms(&self, term_set: &mut BTreeSet<Term>) {

src/reader/mod.rs

@@ -7,6 +7,7 @@ use crate::directory::Directory;
 use crate::directory::WatchHandle;
 use crate::directory::META_LOCK;
 use crate::Index;
+use crate::Result;
 use crate::Searcher;
 use crate::SegmentReader;
 use std::sync::Arc;
@@ -61,7 +62,7 @@ impl IndexReaderBuilder {
     /// to open different segment readers. It may take hundreds of milliseconds
     /// of time and it may return an error.
     /// TODO(pmasurel) Use the `TryInto` trait once it is available in stable.
-    pub fn try_into(self) -> crate::Result<IndexReader> {
+    pub fn try_into(self) -> Result<IndexReader> {
         let inner_reader = InnerIndexReader {
             index: self.index,
             num_searchers: self.num_searchers,
@@ -120,14 +121,14 @@ struct InnerIndexReader {
 }
 
 impl InnerIndexReader {
-    fn reload(&self) -> crate::Result<()> {
+    fn reload(&self) -> Result<()> {
         let segment_readers: Vec<SegmentReader> = {
             let _meta_lock = self.index.directory().acquire_lock(&META_LOCK)?;
             let searchable_segments = self.searchable_segments()?;
             searchable_segments
                 .iter()
                 .map(SegmentReader::open)
-                .collect::<crate::Result<_>>()?
+                .collect::<Result<_>>()?
         };
         let schema = self.index.schema();
         let searchers = (0..self.num_searchers)
@@ -138,7 +139,7 @@ impl InnerIndexReader {
     }
 
     /// Returns the list of segments that are searchable
-    fn searchable_segments(&self) -> crate::Result<Vec<Segment>> {
+    fn searchable_segments(&self) -> Result<Vec<Segment>> {
         self.index.searchable_segments()
     }
 
@@ -175,7 +176,7 @@ impl IndexReader {
     ///
     /// This automatic reload can take 10s of milliseconds to kick in however, and in unit tests
     /// it can be nice to deterministically force the reload of searchers.
-    pub fn reload(&self) -> crate::Result<()> {
+    pub fn reload(&self) -> Result<()> {
         self.inner.reload()
     }
 

src/schema/facet.rs

@@ -122,11 +122,6 @@ 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 {
@@ -270,21 +265,4 @@ 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/snippet/mod.rs

@@ -1,8 +1,10 @@
 use crate::query::Query;
 use crate::schema::Field;
 use crate::schema::Value;
-use crate::tokenizer::{TextAnalyzer, Token};
+use crate::tokenizer::BoxedTokenizer;
+use crate::tokenizer::{Token, TokenStream};
 use crate::Document;
+use crate::Result;
 use crate::Searcher;
 use htmlescape::encode_minimal;
 use std::cmp::Ordering;
@@ -140,7 +142,7 @@ 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: &TextAnalyzer,
+    tokenizer: &BoxedTokenizer,
     text: &'a str,
     terms: &BTreeMap<String, f32>,
     max_num_chars: usize,
@@ -249,7 +251,7 @@ fn select_best_fragment_combination(fragments: &[FragmentCandidate], text: &str)
 /// ```
 pub struct SnippetGenerator {
     terms_text: BTreeMap<String, f32>,
-    tokenizer: TextAnalyzer,
+    tokenizer: BoxedTokenizer,
     field: Field,
     max_num_chars: usize,
 }
@@ -260,7 +262,7 @@ impl SnippetGenerator {
         searcher: &Searcher,
         query: &dyn Query,
         field: Field,
-    ) -> crate::Result<SnippetGenerator> {
+    ) -> Result<SnippetGenerator> {
         let mut terms = BTreeSet::new();
         query.query_terms(&mut terms);
         let terms_text: BTreeMap<String, f32> = terms
@@ -345,11 +347,12 @@ 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(&From::from(SimpleTokenizer), TEST_TEXT, &terms, 100);
+        let fragments = search_fragments(&boxed_tokenizer, TEST_TEXT, &terms, 100);
         assert_eq!(fragments.len(), 7);
         {
             let first = &fragments[0];
@@ -371,12 +374,13 @@ 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(&From::from(SimpleTokenizer), TEST_TEXT, &terms, 20);
+            let fragments = search_fragments(&boxed_tokenizer, TEST_TEXT, &terms, 20);
             {
                 let first = &fragments[0];
                 assert_eq!(first.score, 1.0);
@@ -385,12 +389,13 @@ 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(&From::from(SimpleTokenizer), TEST_TEXT, &terms, 20);
+            let fragments = search_fragments(&boxed_tokenizer, TEST_TEXT, &terms, 20);
             //assert_eq!(fragments.len(), 7);
             {
                 let first = &fragments[0];
@@ -404,12 +409,14 @@ 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(&From::from(SimpleTokenizer), &text, &terms, 3);
+        let fragments = search_fragments(&boxed_tokenizer, &text, &terms, 3);
 
         assert_eq!(fragments.len(), 1);
         {
@@ -426,12 +433,14 @@ 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(&From::from(SimpleTokenizer), &text, &terms, 3);
+        let fragments = search_fragments(&boxed_tokenizer, &text, &terms, 3);
 
         assert_eq!(fragments.len(), 2);
         {
@@ -448,13 +457,15 @@ 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(&From::from(SimpleTokenizer), &text, &terms, 7);
+        let fragments = search_fragments(&boxed_tokenizer, &text, &terms, 7);
 
         assert_eq!(fragments.len(), 2);
         {
@@ -471,12 +482,14 @@ 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(&From::from(SimpleTokenizer), &text, &terms, 3);
+        let fragments = search_fragments(&boxed_tokenizer, &text, &terms, 3);
 
         assert_eq!(fragments.len(), 0);
 
@@ -487,10 +500,12 @@ 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(&From::from(SimpleTokenizer), &text, &terms, 3);
+        let fragments = search_fragments(&boxed_tokenizer, &text, &terms, 3);
         assert_eq!(fragments.len(), 0);
 
         let snippet = select_best_fragment_combination(&fragments[..], &text);

src/store/mod.rs

@@ -57,7 +57,7 @@ use self::compression_snap::{compress, decompress};
 pub mod tests {
 
     use super::*;
-    use crate::directory::{Directory, RAMDirectory, WritePtr};
+    use crate::directory::{Directory, RAMDirectory, ReadOnlyDirectory, WritePtr};
     use crate::schema::Document;
     use crate::schema::FieldValue;
     use crate::schema::Schema;

src/store/reader.rs

@@ -1,3 +1,5 @@
+use crate::Result;
+
 use super::decompress;
 use super::skiplist::SkipList;
 use crate::common::BinarySerializable;
@@ -73,7 +75,7 @@ impl StoreReader {
     ///
     /// It should not be called to score documents
     /// for instance.
-    pub fn get(&self, doc_id: DocId) -> crate::Result<Document> {
+    pub fn get(&self, doc_id: DocId) -> Result<Document> {
         let (first_doc_id, block_offset) = self.block_offset(doc_id);
         self.read_block(block_offset as usize)?;
         let current_block_mut = self.current_block.borrow_mut();

src/termdict/mod.rs

@@ -36,9 +36,9 @@ pub use self::termdict::{TermDictionary, TermDictionaryBuilder};
 mod tests {
     use super::{TermDictionary, TermDictionaryBuilder, TermStreamer};
     use crate::core::Index;
-    use crate::directory::{Directory, RAMDirectory, ReadOnlySource};
+    use crate::directory::{Directory, RAMDirectory, ReadOnlyDirectory, ReadOnlySource};
     use crate::postings::TermInfo;
-    use crate::schema::{Document, Schema, TEXT};
+    use crate::schema::{Document, FieldType, Schema, TEXT};
     use std::path::PathBuf;
     use std::str;
 
@@ -52,12 +52,6 @@ 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] = [
@@ -73,7 +67,9 @@ mod tests {
         let path = PathBuf::from("TermDictionary");
         {
             let write = directory.open_write(&path).unwrap();
-            let mut term_dictionary_builder = TermDictionaryBuilder::create(write).unwrap();
+            let field_type = FieldType::Str(TEXT);
+            let mut term_dictionary_builder =
+                TermDictionaryBuilder::create(write, &field_type).unwrap();
             for term in COUNTRIES.iter() {
                 term_dictionary_builder
                     .insert(term.as_bytes(), &make_term_info(0u64))
@@ -97,7 +93,9 @@ mod tests {
         let path = PathBuf::from("TermDictionary");
         {
             let write = directory.open_write(&path).unwrap();
-            let mut term_dictionary_builder = TermDictionaryBuilder::create(write).unwrap();
+            let field_type = FieldType::Str(TEXT);
+            let mut term_dictionary_builder =
+                TermDictionaryBuilder::create(write, &field_type).unwrap();
             term_dictionary_builder
                 .insert("abc".as_bytes(), &make_term_info(34u64))
                 .unwrap();
@@ -181,8 +179,10 @@ 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![]).unwrap();
+            let mut term_dictionary_builder =
+                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
             for &(ref id, ref i) in &ids {
                 term_dictionary_builder
                     .insert(id.as_bytes(), &make_term_info(*i as u64))
@@ -209,8 +209,10 @@ 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![]).unwrap();
+            let mut term_dictionary_builder =
+                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
             // term requires more than 16bits
             term_dictionary_builder
                 .insert("abcdefghijklmnopqrstuvwxy", &make_term_info(1))
@@ -242,8 +244,10 @@ 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![]).unwrap();
+            let mut term_dictionary_builder =
+                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
             for &(ref id, ref i) in &ids {
                 term_dictionary_builder
                     .insert(id.as_bytes(), &make_term_info(*i as u64))
@@ -309,8 +313,10 @@ 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![]).unwrap();
+            let mut term_dictionary_builder =
+                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
             term_dictionary_builder
                 .insert(&[], &make_term_info(1 as u64))
                 .unwrap();
@@ -331,8 +337,10 @@ 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![]).unwrap();
+            let mut term_dictionary_builder =
+                TermDictionaryBuilder::create(vec![], &field_type).unwrap();
             for i in 0u8..10u8 {
                 let number_arr = [i; 1];
                 term_dictionary_builder
@@ -344,91 +352,41 @@ mod tests {
         let source = ReadOnlySource::from(buffer);
         let term_dictionary: TermDictionary = TermDictionary::from_source(&source);
 
-        let value_list = |mut streamer: TermStreamer<'_>, backwards: bool| {
+        let value_list = |mut streamer: TermStreamer<'_>| {
             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, 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),
+                value_list(range),
                 vec![2u32, 3u32, 4u32, 5u32, 6u32, 7u32, 8u32, 9u32]
             );
         }
         {
             let range = term_dictionary.range().gt([2u8]).into_stream();
             assert_eq!(
-                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),
+                value_list(range),
                 vec![3u32, 4u32, 5u32, 6u32, 7u32, 8u32, 9u32]
             );
         }
         {
             let range = term_dictionary.range().lt([6u8]).into_stream();
-            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]
-            );
+            assert_eq!(value_list(range), vec![0u32, 1u32, 2u32, 3u32, 4u32, 5u32]);
         }
         {
             let range = term_dictionary.range().le([6u8]).into_stream();
             assert_eq!(
-                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),
+                value_list(range),
                 vec![0u32, 1u32, 2u32, 3u32, 4u32, 5u32, 6u32]
             );
         }
         {
             let range = term_dictionary.range().ge([0u8]).lt([5u8]).into_stream();
-            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]);
+            assert_eq!(value_list(range), vec![0u32, 1u32, 2u32, 3u32, 4u32]);
         }
     }
 
@@ -450,7 +408,9 @@ mod tests {
         let path = PathBuf::from("TermDictionary");
         {
             let write = directory.open_write(&path).unwrap();
-            let mut term_dictionary_builder = TermDictionaryBuilder::create(write).unwrap();
+            let field_type = FieldType::Str(TEXT);
+            let mut term_dictionary_builder =
+                TermDictionaryBuilder::create(write, &field_type).unwrap();
             for term in COUNTRIES.iter() {
                 term_dictionary_builder
                     .insert(term.as_bytes(), &make_term_info(0u64))

src/termdict/streamer.rs

@@ -51,12 +51,6 @@ 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) -> io::Result<Self> {
+    pub fn create(w: W, _field_type: &FieldType) -> io::Result<Self> {
         let fst_builder = tantivy_fst::MapBuilder::new(w).map_err(convert_fst_error)?;
         Ok(TermDictionaryBuilder {
             fst_builder,
@@ -92,14 +92,6 @@ 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.
 ///
@@ -130,8 +122,14 @@ impl TermDictionary {
     }
 
     /// Creates an empty term dictionary which contains no terms.
-    pub fn empty() -> Self {
-        TermDictionary::from_source(&*EMPTY_DATA_SOURCE)
+    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)
     }
 
     /// Returns the number of terms in the dictionary.

src/tokenizer/alphanum_only.rs

@@ -2,7 +2,7 @@
 //! ```rust
 //! use tantivy::tokenizer::*;
 //!
-//! let tokenizer = TextAnalyzer::from(RawTokenizer)
+//! let tokenizer = 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 = TextAnalyzer::from(SimpleTokenizer)
+//! let tokenizer = SimpleTokenizer
 //!   .filter(AlphaNumOnlyFilter);
 //!
 //! let mut stream = tokenizer.token_stream("hello there 💣");
@@ -19,30 +19,56 @@
 //! // the "emoji" is dropped because its not an alphanum
 //! assert!(stream.next().is_none());
 //! ```
-use super::{BoxTokenStream, Token, TokenFilter, TokenStream};
+use super::{Token, TokenFilter, TokenStream};
 
 /// `TokenFilter` that removes all tokens that contain non
 /// ascii alphanumeric characters.
 #[derive(Clone)]
 pub struct AlphaNumOnlyFilter;
 
-pub struct AlphaNumOnlyFilterStream<'a> {
-    tail: BoxTokenStream<'a>,
+pub struct AlphaNumOnlyFilterStream<TailTokenStream>
+where
+    TailTokenStream: TokenStream,
+{
+    tail: TailTokenStream,
 }
 
-impl<'a> AlphaNumOnlyFilterStream<'a> {
+impl<TailTokenStream> AlphaNumOnlyFilterStream<TailTokenStream>
+where
+    TailTokenStream: TokenStream,
+{
     fn predicate(&self, token: &Token) -> bool {
         token.text.chars().all(|c| c.is_ascii_alphanumeric())
     }
-}
 
-impl TokenFilter for AlphaNumOnlyFilter {
-    fn transform<'a>(&self, token_stream: BoxTokenStream<'a>) -> BoxTokenStream<'a> {
-        BoxTokenStream::from(AlphaNumOnlyFilterStream { tail: token_stream })
+    fn wrap(tail: TailTokenStream) -> AlphaNumOnlyFilterStream<TailTokenStream> {
+        AlphaNumOnlyFilterStream { tail }
     }
 }
 
-impl<'a> TokenStream for AlphaNumOnlyFilterStream<'a> {
+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()
+    }
+
     fn advance(&mut self) -> bool {
         while self.tail.advance() {
             if self.predicate(self.tail.token()) {
@@ -52,12 +78,4 @@ impl<'a> TokenStream for AlphaNumOnlyFilterStream<'a> {
 
         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::{BoxTokenStream, Token, TokenFilter, TokenStream};
+use super::{Token, TokenFilter, TokenStream};
 use std::mem;
 
 /// This class converts alphabetic, numeric, and symbolic Unicode characters
@@ -7,21 +7,26 @@ use std::mem;
 #[derive(Clone)]
 pub struct AsciiFoldingFilter;
 
-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),
-        })
+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)
     }
 }
 
-pub struct AsciiFoldingFilterTokenStream<'a> {
+pub struct AsciiFoldingFilterTokenStream<TailTokenStream> {
     buffer: String,
-    tail: BoxTokenStream<'a>,
+    tail: TailTokenStream,
 }
 
-impl<'a> TokenStream for AsciiFoldingFilterTokenStream<'a> {
+impl<TailTokenStream> TokenStream for AsciiFoldingFilterTokenStream<TailTokenStream>
+where
+    TailTokenStream: TokenStream,
+{
     fn advance(&mut self) -> bool {
         if !self.tail.advance() {
             return false;
@@ -43,6 +48,18 @@ impl<'a> TokenStream for AsciiFoldingFilterTokenStream<'a> {
     }
 }
 
+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`
@@ -1544,7 +1561,8 @@ mod tests {
     use crate::tokenizer::AsciiFoldingFilter;
     use crate::tokenizer::RawTokenizer;
     use crate::tokenizer::SimpleTokenizer;
-    use crate::tokenizer::TextAnalyzer;
+    use crate::tokenizer::TokenStream;
+    use crate::tokenizer::Tokenizer;
     use std::iter;
 
     #[test]
@@ -1561,7 +1579,7 @@ mod tests {
 
     fn folding_helper(text: &str) -> Vec<String> {
         let mut tokens = Vec::new();
-        TextAnalyzer::from(SimpleTokenizer)
+        SimpleTokenizer
             .filter(AsciiFoldingFilter)
             .token_stream(text)
             .process(&mut |token| {
@@ -1571,9 +1589,7 @@ mod tests {
     }
 
     fn folding_using_raw_tokenizer_helper(text: &str) -> String {
-        let mut token_stream = TextAnalyzer::from(RawTokenizer)
-            .filter(AsciiFoldingFilter)
-            .token_stream(text);
+        let mut token_stream = 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::{BoxTokenStream, Token, TokenStream, Tokenizer};
+use super::{Token, TokenStream, Tokenizer};
 use crate::schema::FACET_SEP_BYTE;
 
 /// The `FacetTokenizer` process a `Facet` binary representation
@@ -25,14 +25,15 @@ pub struct FacetTokenStream<'a> {
     token: Token,
 }
 
-impl Tokenizer for FacetTokenizer {
-    fn token_stream<'a>(&self, text: &'a str) -> BoxTokenStream<'a> {
+impl<'a> Tokenizer<'a> for FacetTokenizer {
+    type TokenStreamImpl = FacetTokenStream<'a>;
+
+    fn token_stream(&self, text: &'a str) -> Self::TokenStreamImpl {
         FacetTokenStream {
             text,
             state: State::RootFacetNotEmitted, //< pos is the first char that has not been processed yet.
             token: Token::default(),
         }
-        .into()
     }
 }
 
@@ -83,7 +84,7 @@ mod tests {
 
     use super::FacetTokenizer;
     use crate::schema::Facet;
-    use crate::tokenizer::{Token, Tokenizer};
+    use crate::tokenizer::{Token, TokenStream, Tokenizer};
 
     #[test]
     fn test_facet_tokenizer() {

src/tokenizer/lower_caser.rs

@@ -1,23 +1,24 @@
 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;
 
-pub struct LowerCaserTokenStream<'a> {
+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> {
     buffer: String,
-    tail: BoxTokenStream<'a>,
+    tail: TailTokenStream,
 }
 
 // writes a lowercased version of text into output.
@@ -30,7 +31,18 @@ fn to_lowercase_unicode(text: &mut String, output: &mut String) {
     }
 }
 
-impl<'a> TokenStream for LowerCaserTokenStream<'a> {
+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()
+    }
+
     fn advance(&mut self) -> bool {
         if !self.tail.advance() {
             return false;
@@ -44,19 +56,26 @@ impl<'a> TokenStream for LowerCaserTokenStream<'a> {
         }
         true
     }
+}
 
-    fn token(&self) -> &Token {
-        self.tail.token()
-    }
-
-    fn token_mut(&mut self) -> &mut Token {
-        self.tail.token_mut()
+impl<TailTokenStream> LowerCaserTokenStream<TailTokenStream>
+where
+    TailTokenStream: TokenStream,
+{
+    fn wrap(tail: TailTokenStream) -> LowerCaserTokenStream<TailTokenStream> {
+        LowerCaserTokenStream {
+            tail,
+            buffer: String::with_capacity(100),
+        }
     }
 }
 
 #[cfg(test)]
 mod tests {
-    use crate::tokenizer::{LowerCaser, SimpleTokenizer, TextAnalyzer};
+    use crate::tokenizer::LowerCaser;
+    use crate::tokenizer::SimpleTokenizer;
+    use crate::tokenizer::TokenStream;
+    use crate::tokenizer::Tokenizer;
 
     #[test]
     fn test_to_lower_case() {
@@ -68,9 +87,7 @@ mod tests {
 
     fn lowercase_helper(text: &str) -> Vec<String> {
         let mut tokens = vec![];
-        let mut token_stream = TextAnalyzer::from(SimpleTokenizer)
-            .filter(LowerCaser)
-            .token_stream(text);
+        let mut token_stream = 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 = TextAnalyzer::from(SimpleTokenizer)
+//! let en_stem = 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 = TextAnalyzer::from(SimpleTokenizer)
+//! let custom_en_tokenizer = SimpleTokenizer
 //!     .filter(RemoveLongFilter::limit(40))
 //!     .filter(LowerCaser);
 //! index
@@ -143,11 +143,10 @@ 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::{
-    BoxTokenFilter, BoxTokenStream, TextAnalyzer, Token, TokenFilter, TokenStream, Tokenizer,
-};
+pub use self::tokenizer::{Token, TokenFilter, TokenStream, Tokenizer};
 
 pub use self::tokenizer_manager::TokenizerManager;
 
@@ -161,9 +160,9 @@ 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, TokenizerManager,
+        Language, LowerCaser, RemoveLongFilter, SimpleTokenizer, Stemmer, Token, Tokenizer,
+        TokenizerManager,
     };
-    use crate::tokenizer::TextAnalyzer;
 
     /// This is a function that can be used in tests and doc tests
     /// to assert a token's correctness.
@@ -230,7 +229,7 @@ pub mod tests {
         let tokenizer_manager = TokenizerManager::default();
         tokenizer_manager.register(
             "el_stem",
-            TextAnalyzer::from(SimpleTokenizer)
+            SimpleTokenizer
                 .filter(RemoveLongFilter::limit(40))
                 .filter(LowerCaser)
                 .filter(Stemmer::new(Language::Greek)),

src/tokenizer/ngram_tokenizer.rs

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

src/tokenizer/raw_tokenizer.rs

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

src/tokenizer/remove_long.rs

@@ -2,7 +2,7 @@
 //! ```rust
 //! use tantivy::tokenizer::*;
 //!
-//! let tokenizer = TextAnalyzer::from(SimpleTokenizer)
+//! let tokenizer = SimpleTokenizer
 //!   .filter(RemoveLongFilter::limit(5));
 //!
 //! let mut stream = tokenizer.token_stream("toolong nice");
@@ -13,7 +13,6 @@
 //! ```
 //!
 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).
@@ -32,27 +31,56 @@ impl RemoveLongFilter {
     }
 }
 
-impl<'a> RemoveLongFilterStream<'a> {
+impl<TailTokenStream> RemoveLongFilterStream<TailTokenStream>
+where
+    TailTokenStream: TokenStream,
+{
     fn predicate(&self, token: &Token) -> bool {
         token.text.len() < self.token_length_limit
     }
-}
 
-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,
-        })
+    fn wrap(
+        token_length_limit: usize,
+        tail: TailTokenStream,
+    ) -> RemoveLongFilterStream<TailTokenStream> {
+        RemoveLongFilterStream {
+            token_length_limit,
+            tail,
+        }
     }
 }
 
-pub struct RemoveLongFilterStream<'a> {
-    token_length_limit: usize,
-    tail: BoxTokenStream<'a>,
+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)
+    }
 }
 
-impl<'a> TokenStream for RemoveLongFilterStream<'a> {
+pub struct RemoveLongFilterStream<TailTokenStream>
+where
+    TailTokenStream: TokenStream,
+{
+    token_length_limit: usize,
+    tail: TailTokenStream,
+}
+
+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()
+    }
+
     fn advance(&mut self) -> bool {
         while self.tail.advance() {
             if self.predicate(self.tail.token()) {
@@ -61,12 +89,4 @@ impl<'a> TokenStream for RemoveLongFilterStream<'a> {
         }
         false
     }
-
-    fn token(&self) -> &Token {
-        self.tail.token()
-    }
-
-    fn token_mut(&mut self) -> &mut Token {
-        self.tail.token_mut()
-    }
 }

src/tokenizer/simple_tokenizer.rs

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

src/tokenizer/stemmer.rs

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

src/tokenizer/stop_word_filter.rs

@@ -2,7 +2,7 @@
 //! ```rust
 //! use tantivy::tokenizer::*;
 //!
-//! let tokenizer = TextAnalyzer::from(SimpleTokenizer)
+//! let tokenizer = SimpleTokenizer
 //!   .filter(StopWordFilter::remove(vec!["the".to_string(), "is".to_string()]));
 //!
 //! let mut stream = tokenizer.token_stream("the fox is crafty");
@@ -11,7 +11,6 @@
 //! 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;
@@ -49,27 +48,53 @@ impl StopWordFilter {
     }
 }
 
-pub struct StopWordFilterStream<'a> {
+pub struct StopWordFilterStream<TailTokenStream>
+where
+    TailTokenStream: TokenStream,
+{
     words: StopWordHashSet,
-    tail: BoxTokenStream<'a>,
+    tail: TailTokenStream,
 }
 
-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> 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<'a> StopWordFilterStream<'a> {
+impl<TailTokenStream> StopWordFilterStream<TailTokenStream>
+where
+    TailTokenStream: TokenStream,
+{
     fn predicate(&self, token: &Token) -> bool {
         !self.words.contains(&token.text)
     }
+
+    fn wrap(
+        words: StopWordHashSet,
+        tail: TailTokenStream,
+    ) -> StopWordFilterStream<TailTokenStream> {
+        StopWordFilterStream { words, tail }
+    }
 }
 
-impl<'a> TokenStream for StopWordFilterStream<'a> {
+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()
+    }
+
     fn advance(&mut self) -> bool {
         while self.tail.advance() {
             if self.predicate(self.tail.token()) {
@@ -78,14 +103,6 @@ impl<'a> TokenStream for StopWordFilterStream<'a> {
         }
         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

@@ -1,21 +1,23 @@
-use crate::tokenizer::{BoxTokenStream, Token, TokenStream};
-use std::ops::DerefMut;
+use crate::tokenizer::{Token, TokenStream};
 
 const POSITION_GAP: usize = 2;
 
-pub(crate) struct TokenStreamChain<'a> {
+pub(crate) struct TokenStreamChain<TTokenStream: TokenStream> {
     offsets: Vec<usize>,
-    token_streams: Vec<BoxTokenStream<'a>>,
+    token_streams: Vec<TTokenStream>,
     position_shift: usize,
     stream_idx: usize,
     token: Token,
 }
 
-impl<'a> TokenStreamChain<'a> {
+impl<'a, TTokenStream> TokenStreamChain<TTokenStream>
+where
+    TTokenStream: TokenStream,
+{
     pub fn new(
         offsets: Vec<usize>,
-        token_streams: Vec<BoxTokenStream<'a>>,
-    ) -> TokenStreamChain<'a> {
+        token_streams: Vec<TTokenStream>,
+    ) -> TokenStreamChain<TTokenStream> {
         TokenStreamChain {
             offsets,
             stream_idx: 0,
@@ -26,10 +28,13 @@ impl<'a> TokenStreamChain<'a> {
     }
 }
 
-impl<'a> TokenStream for TokenStreamChain<'a> {
+impl<'a, TTokenStream> TokenStream for TokenStreamChain<TTokenStream>
+where
+    TTokenStream: TokenStream,
+{
     fn advance(&mut self) -> bool {
         while self.stream_idx < self.token_streams.len() {
-            let token_stream = self.token_streams[self.stream_idx].deref_mut();
+            let token_stream = &mut self.token_streams[self.stream_idx];
             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::{BoxTokenStream, Token, TokenStream, TokenStreamChain};
+use crate::tokenizer::{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],
-    ) -> BoxTokenStream {
+    ) -> Box<dyn TokenStream + 'a> {
         if tok_strings.len() == 1 {
-            PreTokenizedStream::from((*tok_strings[0]).clone()).into()
+            Box::new(PreTokenizedStream::from((*tok_strings[0]).clone()))
         } else {
             let mut offsets = vec![];
             let mut total_offset = 0;
@@ -53,12 +53,11 @@ impl PreTokenizedStream {
                     total_offset += last_token.offset_to;
                 }
             }
-            // TODO remove the string cloning.
-            let token_streams: Vec<BoxTokenStream<'static>> = tok_strings
+            let token_streams: Vec<_> = tok_strings
                 .iter()
-                .map(|&tok_string| PreTokenizedStream::from((*tok_string).clone()).into())
+                .map(|tok_string| PreTokenizedStream::from((*tok_string).clone()))
                 .collect();
-            TokenStreamChain::new(offsets, token_streams).into()
+            Box::new(TokenStreamChain::new(offsets, token_streams))
         }
     }
 }

src/tokenizer/tokenizer.rs

@@ -2,7 +2,6 @@ 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)]
@@ -34,31 +33,20 @@ impl Default for Token {
     }
 }
 
-/// `TextAnalyzer` tokenizes an input text into tokens and modifies the resulting `TokenStream`.
+/// `Tokenizer` are in charge of splitting text into a stream of token
+/// before indexing.
 ///
-/// 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>,
-}
+/// 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;
 
-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,
-        }
-    }
+    /// Creates a token stream for a given `str`.
+    fn token_stream(&self, text: &'a str) -> Self::TokenStreamImpl;
 
     /// Appends a token filter to the current tokenizer.
     ///
@@ -70,26 +58,90 @@ impl TextAnalyzer {
     /// ```rust
     /// use tantivy::tokenizer::*;
     ///
-    /// let en_stem = TextAnalyzer::from(SimpleTokenizer)
+    /// let en_stem = SimpleTokenizer
     ///     .filter(RemoveLongFilter::limit(40))
     ///     .filter(LowerCaser)
     ///     .filter(Stemmer::default());
     /// ```
     ///
-    pub fn filter<F: Into<BoxTokenFilter>>(mut self, token_filter: F) -> Self {
-        self.token_filters.push(token_filter.into());
-        self
+    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)
     }
 
     /// Tokenize an array`&str`
     ///
-    /// The resulting `BoxTokenStream` is equivalent to what would be obtained if the &str were
+    /// 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.
-    pub fn token_stream_texts<'a>(&self, texts: &'a [&'a str]) -> BoxTokenStream<'a> {
+    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> {
         assert!(!texts.is_empty());
         if texts.len() == 1 {
-            self.token_stream(texts[0])
+            Box::new(self.0.token_stream(texts[0]))
         } else {
             let mut offsets = vec![];
             let mut total_offset = 0;
@@ -97,124 +149,34 @@ impl TextAnalyzer {
                 offsets.push(total_offset);
                 total_offset += text.len();
             }
-            let token_streams: Vec<BoxTokenStream<'a>> = texts
-                .iter()
-                .cloned()
-                .map(|text| self.token_stream(text))
-                .collect();
-            From::from(TokenStreamChain::new(offsets, token_streams))
+            let token_streams: Vec<_> =
+                texts.iter().map(|text| self.0.token_stream(text)).collect();
+            Box::new(TokenStreamChain::new(offsets, token_streams))
         }
     }
 
-    /// 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
+    fn boxed_clone(&self) -> BoxedTokenizer {
+        self.0.clone().into()
     }
 }
 
-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> {
+impl<'b> TokenStream for Box<dyn TokenStream + 'b> {
     fn advance(&mut self) -> bool {
         let token_stream: &mut dyn TokenStream = self.borrow_mut();
         token_stream.advance()
     }
 
-    fn token<'b>(&'b self) -> &'b Token {
-        let token_stream: &'b (dyn TokenStream + 'a) = self.borrow();
+    fn token(&self) -> &Token {
+        let token_stream: &dyn TokenStream = self.borrow();
         token_stream.token()
     }
 
-    fn token_mut<'b>(&'b mut self) -> &'b mut Token {
-        let token_stream: &'b mut (dyn TokenStream + 'a) = self.borrow_mut();
+    fn token_mut(&mut self) -> &mut Token {
+        let token_stream: &mut dyn TokenStream = 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.
@@ -224,7 +186,7 @@ impl<T: TokenFilter> From<T> for BoxTokenFilter {
 /// ```
 /// use tantivy::tokenizer::*;
 ///
-/// let tokenizer = TextAnalyzer::from(SimpleTokenizer)
+/// let tokenizer = SimpleTokenizer
 ///        .filter(RemoveLongFilter::limit(40))
 ///        .filter(LowerCaser);
 /// let mut token_stream = tokenizer.token_stream("Hello, happy tax payer");
@@ -263,7 +225,7 @@ pub trait TokenStream {
     /// ```
     /// use tantivy::tokenizer::*;
     ///
-    /// let tokenizer = TextAnalyzer::from(SimpleTokenizer)
+    /// let tokenizer = SimpleTokenizer
     ///       .filter(RemoveLongFilter::limit(40))
     ///       .filter(LowerCaser);
     /// let mut token_stream = tokenizer.token_stream("Hello, happy tax payer");
@@ -281,8 +243,6 @@ 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() {
@@ -293,20 +253,33 @@ pub trait TokenStream {
     }
 }
 
-pub trait TokenFilterClone {
-    fn box_clone(&self) -> BoxTokenFilter;
+#[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)
+    }
 }
 
 /// Trait for the pluggable components of `Tokenizer`s.
-pub trait TokenFilter: 'static + Send + Sync + TokenFilterClone {
-    /// Wraps a token stream and returns the modified one.
-    fn transform<'a>(&self, token_stream: BoxTokenStream<'a>) -> BoxTokenStream<'a>;
-}
+pub trait TokenFilter<TailTokenStream: TokenStream>: Clone {
+    /// The resulting `TokenStream` type.
+    type ResultTokenStream: TokenStream;
 
-impl<T: TokenFilter + Clone> TokenFilterClone for T {
-    fn box_clone(&self) -> BoxTokenFilter {
-        BoxTokenFilter::from(self.clone())
-    }
+    /// Wraps a token stream and returns the modified one.
+    fn transform(&self, token_stream: TailTokenStream) -> Self::ResultTokenStream;
 }
 
 #[cfg(test)]

src/tokenizer/tokenizer_manager.rs

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

tests/failpoints/mod.rs

@@ -1,6 +1,8 @@
 use fail;
 use std::path::Path;
-use tantivy::directory::{Directory, ManagedDirectory, RAMDirectory, TerminatingWrite};
+use tantivy::directory::{
+    Directory, ManagedDirectory, RAMDirectory, ReadOnlyDirectory, TerminatingWrite,
+};
 use tantivy::doc;
 use tantivy::schema::{Schema, TEXT};
 use tantivy::{Index, Term};