Remove BoxTokenStream.

wasm compilation fails with dep only

Cargo.toml

@@ -19,6 +19,7 @@ oneshot = "0.1.5"
 base64 = "0.21.0"
 byteorder = "1.4.3"
 crc32fast = "1.3.2"
+dyn-clone = "1.0.11"
 once_cell = "1.10.0"
 regex = { version = "1.5.5", default-features = false, features = ["std", "unicode"] }
 aho-corasick = "1.0"
@@ -44,7 +45,7 @@ census = "0.4.0"
 rustc-hash = "1.1.0"
 thiserror = "1.0.30"
 htmlescape = "0.3.1"
-fail = "0.5.0"
+fail = { version = "0.5.0", optional = true }
 murmurhash32 = "0.3.0"
 time = { version = "0.3.10", features = ["serde-well-known"] }
 smallvec = "1.8.0"
@@ -112,7 +113,7 @@ lz4-compression = ["lz4_flex"]
 snappy-compression = ["snap"]
 zstd-compression = ["zstd"]
 
-failpoints = ["fail/failpoints"]
+failpoints = ["fail", "fail/failpoints"]
 unstable = [] # useful for benches.
 
 quickwit = ["sstable", "futures-util"]

bitpacker/src/filter_vec/mod.rs

@@ -1,6 +1,6 @@
 use std::ops::RangeInclusive;
 
-#[cfg(any(target_arch = "x86_64"))]
+#[cfg(target_arch = "x86_64")]
 mod avx2;
 
 mod scalar;

columnar/Cargo.toml

@@ -5,7 +5,7 @@ edition = "2021"
 license = "MIT"
 homepage = "https://github.com/quickwit-oss/tantivy"
 repository = "https://github.com/quickwit-oss/tantivy"
-desciption = "column oriented storage for tantivy"
+description = "column oriented storage for tantivy"
 categories = ["database-implementations", "data-structures", "compression"]
 
 [dependencies]

examples/warmer.rs

@@ -6,12 +6,14 @@ use tantivy::collector::TopDocs;
 use tantivy::query::QueryParser;
 use tantivy::schema::{Schema, FAST, TEXT};
 use tantivy::{
-    doc, DocAddress, DocId, Index, IndexReader, Opstamp, Searcher, SearcherGeneration, SegmentId,
-    SegmentReader, Warmer,
+    doc, DocAddress, DocId, Index, Opstamp, Searcher, SearcherGeneration, SegmentId, SegmentReader,
+    Warmer,
 };
 
 // This example shows how warmers can be used to
-// load a values from an external sources using the Warmer API.
+// load values from an external sources and
+// tie their lifecycle to that of the index segments
+// using the Warmer API.
 //
 // In this example, we assume an e-commerce search engine.
 
@@ -23,9 +25,11 @@ pub trait PriceFetcher: Send + Sync + 'static {
     fn fetch_prices(&self, product_ids: &[ProductId]) -> Vec<Price>;
 }
 
+type SegmentKey = (SegmentId, Option<Opstamp>);
+
 struct DynamicPriceColumn {
     field: String,
-    price_cache: RwLock<HashMap<(SegmentId, Option<Opstamp>), Arc<Vec<Price>>>>,
+    price_cache: RwLock<HashMap<SegmentKey, Arc<Vec<Price>>>>,
     price_fetcher: Box<dyn PriceFetcher>,
 }
 
@@ -46,7 +50,6 @@ impl DynamicPriceColumn {
 impl Warmer for DynamicPriceColumn {
     fn warm(&self, searcher: &Searcher) -> tantivy::Result<()> {
         for segment in searcher.segment_readers() {
-            let key = (segment.segment_id(), segment.delete_opstamp());
             let product_id_reader = segment
                 .fast_fields()
                 .u64(&self.field)?
@@ -55,37 +58,40 @@ impl Warmer for DynamicPriceColumn {
                 .doc_ids_alive()
                 .map(|doc| product_id_reader.get_val(doc))
                 .collect();
-            let mut prices_it = self.price_fetcher.fetch_prices(&product_ids).into_iter();
-            let mut price_vals: Vec<Price> = Vec::new();
-            for doc in 0..segment.max_doc() {
-                if segment.is_deleted(doc) {
-                    price_vals.push(0);
-                } else {
-                    price_vals.push(prices_it.next().unwrap())
-                }
-            }
+
+            let mut prices = self.price_fetcher.fetch_prices(&product_ids).into_iter();
+
+            let prices: Vec<Price> = (0..segment.max_doc())
+                .map(|doc| {
+                    if !segment.is_deleted(doc) {
+                        prices.next().unwrap()
+                    } else {
+                        0
+                    }
+                })
+                .collect();
+
+            let key = (segment.segment_id(), segment.delete_opstamp());
             self.price_cache
                 .write()
                 .unwrap()
-                .insert(key, Arc::new(price_vals));
+                .insert(key, Arc::new(prices));
         }
+
         Ok(())
     }
 
     fn garbage_collect(&self, live_generations: &[&SearcherGeneration]) {
-        let live_segment_id_and_delete_ops: HashSet<(SegmentId, Option<Opstamp>)> =
-            live_generations
-                .iter()
-                .flat_map(|gen| gen.segments())
-                .map(|(&segment_id, &opstamp)| (segment_id, opstamp))
-                .collect();
-        let mut price_cache_wrt = self.price_cache.write().unwrap();
-        // let price_cache = std::mem::take(&mut *price_cache_wrt);
-        // Drain would be nicer here.
-        *price_cache_wrt = std::mem::take(&mut *price_cache_wrt)
-            .into_iter()
-            .filter(|(seg_id_and_op, _)| !live_segment_id_and_delete_ops.contains(seg_id_and_op))
+        let live_keys: HashSet<SegmentKey> = live_generations
+            .iter()
+            .flat_map(|gen| gen.segments())
+            .map(|(&segment_id, &opstamp)| (segment_id, opstamp))
             .collect();
+
+        self.price_cache
+            .write()
+            .unwrap()
+            .retain(|key, _| live_keys.contains(key));
     }
 }
 
@@ -100,17 +106,17 @@ pub struct ExternalPriceTable {
 
 impl ExternalPriceTable {
     pub fn update_price(&self, product_id: ProductId, price: Price) {
-        let mut prices_wrt = self.prices.write().unwrap();
-        prices_wrt.insert(product_id, price);
+        self.prices.write().unwrap().insert(product_id, price);
     }
 }
 
 impl PriceFetcher for ExternalPriceTable {
     fn fetch_prices(&self, product_ids: &[ProductId]) -> Vec<Price> {
-        let prices_read = self.prices.read().unwrap();
+        let prices = self.prices.read().unwrap();
+
         product_ids
             .iter()
-            .map(|product_id| prices_read.get(product_id).cloned().unwrap_or(0))
+            .map(|product_id| prices.get(product_id).cloned().unwrap_or(0))
             .collect()
     }
 }
@@ -143,11 +149,8 @@ fn main() -> tantivy::Result<()> {
     writer.add_document(doc!(product_id=>SNEAKERS, text=>"uber sweet sneakers"))?;
     writer.commit()?;
 
-    let warmers: Vec<Weak<dyn Warmer>> = vec![Arc::downgrade(
-        &(price_dynamic_column.clone() as Arc<dyn Warmer>),
-    )];
-    let reader: IndexReader = index.reader_builder().warmers(warmers).try_into()?;
-    reader.reload()?;
+    let warmers = vec![Arc::downgrade(&price_dynamic_column) as Weak<dyn Warmer>];
+    let reader = index.reader_builder().warmers(warmers).try_into()?;
 
     let query_parser = QueryParser::for_index(&index, vec![text]);
     let query = query_parser.parse_query("cooking")?;

src/collector/filter_collector_wrapper.rs

@@ -6,32 +6,35 @@
 //
 // Of course, you can have a look at the tantivy's built-in collectors
 // such as the `CountCollector` for more examples.
-
-// ---
-// Importing tantivy...
+use std::fmt::Debug;
 use std::marker::PhantomData;
-use std::sync::Arc;
 
-use columnar::{ColumnValues, DynamicColumn, HasAssociatedColumnType};
+use columnar::{BytesColumn, Column, DynamicColumn, HasAssociatedColumnType};
 
 use crate::collector::{Collector, SegmentCollector};
 use crate::schema::Field;
-use crate::{Score, SegmentReader, TantivyError};
+use crate::{DocId, Score, SegmentReader, TantivyError};
 
 /// The `FilterCollector` filters docs using a fast field value and a predicate.
-/// Only the documents for which the predicate returned "true" will be passed on to the next
-/// collector.
+///
+/// Only the documents containing at least one value for which the predicate returns `true`
+/// will be passed on to the next collector.
+///
+/// In other words,
+/// - documents with no values are filtered out.
+/// - documents with several values are accepted if at least one value matches the predicate.
+///
 ///
 /// ```rust
 /// use tantivy::collector::{TopDocs, FilterCollector};
 /// use tantivy::query::QueryParser;
-/// use tantivy::schema::{Schema, TEXT, INDEXED, FAST};
+/// use tantivy::schema::{Schema, TEXT, FAST};
 /// use tantivy::{doc, DocAddress, Index};
 ///
 /// # fn main() -> tantivy::Result<()> {
 /// let mut schema_builder = Schema::builder();
 /// let title = schema_builder.add_text_field("title", TEXT);
-/// let price = schema_builder.add_u64_field("price", INDEXED | FAST);
+/// let price = schema_builder.add_u64_field("price", FAST);
 /// let schema = schema_builder.build();
 /// let index = Index::create_in_ram(schema);
 ///
@@ -47,20 +50,24 @@ use crate::{Score, SegmentReader, TantivyError};
 ///
 /// let query_parser = QueryParser::for_index(&index, vec![title]);
 /// let query = query_parser.parse_query("diary")?;
-/// let no_filter_collector = FilterCollector::new(price, &|value: u64| value > 20_120u64, TopDocs::with_limit(2));
+/// let no_filter_collector = FilterCollector::new(price, |value: u64| value > 20_120u64, TopDocs::with_limit(2));
 /// let top_docs = searcher.search(&query, &no_filter_collector)?;
 ///
 /// assert_eq!(top_docs.len(), 1);
 /// assert_eq!(top_docs[0].1, DocAddress::new(0, 1));
 ///
-/// let filter_all_collector: FilterCollector<_, _, u64> = FilterCollector::new(price, &|value| value < 5u64, TopDocs::with_limit(2));
+/// let filter_all_collector: FilterCollector<_, _, u64> = FilterCollector::new(price, |value| value < 5u64, TopDocs::with_limit(2));
 /// let filtered_top_docs = searcher.search(&query, &filter_all_collector)?;
 ///
 /// assert_eq!(filtered_top_docs.len(), 0);
 /// # Ok(())
 /// # }
 /// ```
-pub struct FilterCollector<TCollector, TPredicate, TPredicateValue: Default>
+///
+/// Note that this is limited to fast fields which implement the
+/// [`FastValue`][crate::fastfield::FastValue] trait, e.g. `u64` but not `&[u8]`.
+/// To filter based on a bytes fast field, use a [`BytesFilterCollector`] instead.
+pub struct FilterCollector<TCollector, TPredicate, TPredicateValue>
 where TPredicate: 'static + Clone
 {
     field: Field,
@@ -69,19 +76,15 @@ where TPredicate: 'static + Clone
     t_predicate_value: PhantomData<TPredicateValue>,
 }
 
-impl<TCollector, TPredicate, TPredicateValue: Default>
+impl<TCollector, TPredicate, TPredicateValue>
     FilterCollector<TCollector, TPredicate, TPredicateValue>
 where
     TCollector: Collector + Send + Sync,
     TPredicate: Fn(TPredicateValue) -> bool + Send + Sync + Clone,
 {
-    /// Create a new FilterCollector.
-    pub fn new(
-        field: Field,
-        predicate: TPredicate,
-        collector: TCollector,
-    ) -> FilterCollector<TCollector, TPredicate, TPredicateValue> {
-        FilterCollector {
+    /// Create a new `FilterCollector`.
+    pub fn new(field: Field, predicate: TPredicate, collector: TCollector) -> Self {
+        Self {
             field,
             predicate,
             collector,
@@ -90,7 +93,7 @@ where
     }
 }
 
-impl<TCollector, TPredicate, TPredicateValue: Default> Collector
+impl<TCollector, TPredicate, TPredicateValue> Collector
     for FilterCollector<TCollector, TPredicate, TPredicateValue>
 where
     TCollector: Collector + Send + Sync,
@@ -98,8 +101,6 @@ where
     TPredicateValue: HasAssociatedColumnType,
     DynamicColumn: Into<Option<columnar::Column<TPredicateValue>>>,
 {
-    // That's the type of our result.
-    // Our standard deviation will be a float.
     type Fruit = TCollector::Fruit;
 
     type Child = FilterSegmentCollector<TCollector::Child, TPredicate, TPredicateValue>;
@@ -108,7 +109,7 @@ where
         &self,
         segment_local_id: u32,
         segment_reader: &SegmentReader,
-    ) -> crate::Result<FilterSegmentCollector<TCollector::Child, TPredicate, TPredicateValue>> {
+    ) -> crate::Result<Self::Child> {
         let schema = segment_reader.schema();
         let field_entry = schema.get_field_entry(self.field);
         if !field_entry.is_fast() {
@@ -118,16 +119,16 @@ where
             )));
         }
 
-        let fast_field_reader = segment_reader
+        let column_opt = segment_reader
             .fast_fields()
-            .column_first_or_default(schema.get_field_name(self.field))?;
+            .column_opt(field_entry.name())?;
 
         let segment_collector = self
             .collector
             .for_segment(segment_local_id, segment_reader)?;
 
         Ok(FilterSegmentCollector {
-            fast_field_reader,
+            column_opt,
             segment_collector,
             predicate: self.predicate.clone(),
             t_predicate_value: PhantomData,
@@ -146,35 +147,208 @@ where
     }
 }
 
-pub struct FilterSegmentCollector<TSegmentCollector, TPredicate, TPredicateValue>
-where
-    TPredicate: 'static,
-    DynamicColumn: Into<Option<columnar::Column<TPredicateValue>>>,
-{
-    fast_field_reader: Arc<dyn ColumnValues<TPredicateValue>>,
+pub struct FilterSegmentCollector<TSegmentCollector, TPredicate, TPredicateValue> {
+    column_opt: Option<Column<TPredicateValue>>,
     segment_collector: TSegmentCollector,
     predicate: TPredicate,
     t_predicate_value: PhantomData<TPredicateValue>,
 }
 
+impl<TSegmentCollector, TPredicate, TPredicateValue>
+    FilterSegmentCollector<TSegmentCollector, TPredicate, TPredicateValue>
+where
+    TPredicateValue: PartialOrd + Copy + Debug + Send + Sync + 'static,
+    TPredicate: 'static + Fn(TPredicateValue) -> bool + Send + Sync,
+{
+    #[inline]
+    fn accept_document(&self, doc_id: DocId) -> bool {
+        if let Some(column) = &self.column_opt {
+            for val in column.values_for_doc(doc_id) {
+                if (self.predicate)(val) {
+                    return true;
+                }
+            }
+        }
+        false
+    }
+}
+
 impl<TSegmentCollector, TPredicate, TPredicateValue> SegmentCollector
     for FilterSegmentCollector<TSegmentCollector, TPredicate, TPredicateValue>
 where
     TSegmentCollector: SegmentCollector,
     TPredicateValue: HasAssociatedColumnType,
-    TPredicate: 'static + Fn(TPredicateValue) -> bool + Send + Sync,
-    DynamicColumn: Into<Option<columnar::Column<TPredicateValue>>>,
+    TPredicate: 'static + Fn(TPredicateValue) -> bool + Send + Sync, /* DynamicColumn: Into<Option<columnar::Column<TPredicateValue>>> */
 {
     type Fruit = TSegmentCollector::Fruit;
 
     fn collect(&mut self, doc: u32, score: Score) {
-        let value = self.fast_field_reader.get_val(doc);
-        if (self.predicate)(value) {
-            self.segment_collector.collect(doc, score)
+        if self.accept_document(doc) {
+            self.segment_collector.collect(doc, score);
         }
     }
 
-    fn harvest(self) -> <TSegmentCollector as SegmentCollector>::Fruit {
+    fn harvest(self) -> TSegmentCollector::Fruit {
+        self.segment_collector.harvest()
+    }
+}
+
+/// A variant of the [`FilterCollector`] specialized for bytes fast fields, i.e.
+/// it transparently wraps an inner [`Collector`] but filters documents
+/// based on the result of applying the predicate to the bytes fast field.
+///
+/// A document is accepted if and only if the predicate returns `true` for at least one value.
+///
+/// In other words,
+/// - documents with no values are filtered out.
+/// - documents with several values are accepted if at least one value matches the predicate.
+///
+/// ```rust
+/// use tantivy::collector::{TopDocs, BytesFilterCollector};
+/// use tantivy::query::QueryParser;
+/// use tantivy::schema::{Schema, TEXT, FAST};
+/// use tantivy::{doc, DocAddress, Index};
+///
+/// # fn main() -> tantivy::Result<()> {
+/// let mut schema_builder = Schema::builder();
+/// let title = schema_builder.add_text_field("title", TEXT);
+/// let barcode = schema_builder.add_bytes_field("barcode", FAST);
+/// let schema = schema_builder.build();
+/// let index = Index::create_in_ram(schema);
+///
+/// let mut index_writer = index.writer_with_num_threads(1, 10_000_000)?;
+/// index_writer.add_document(doc!(title => "The Name of the Wind", barcode => &b"010101"[..]))?;
+/// index_writer.add_document(doc!(title => "The Diary of Muadib", barcode => &b"110011"[..]))?;
+/// index_writer.add_document(doc!(title => "A Dairy Cow", barcode => &b"110111"[..]))?;
+/// index_writer.add_document(doc!(title => "The Diary of a Young Girl", barcode => &b"011101"[..]))?;
+/// index_writer.add_document(doc!(title => "Bridget Jones's Diary"))?;
+/// index_writer.commit()?;
+///
+/// let reader = index.reader()?;
+/// let searcher = reader.searcher();
+///
+/// let query_parser = QueryParser::for_index(&index, vec![title]);
+/// let query = query_parser.parse_query("diary")?;
+/// let filter_collector = BytesFilterCollector::new(barcode, |bytes: &[u8]| bytes.starts_with(b"01"), TopDocs::with_limit(2));
+/// let top_docs = searcher.search(&query, &filter_collector)?;
+///
+/// assert_eq!(top_docs.len(), 1);
+/// assert_eq!(top_docs[0].1, DocAddress::new(0, 3));
+/// # Ok(())
+/// # }
+/// ```
+pub struct BytesFilterCollector<TCollector, TPredicate>
+where TPredicate: 'static + Clone
+{
+    field: Field,
+    collector: TCollector,
+    predicate: TPredicate,
+}
+
+impl<TCollector, TPredicate> BytesFilterCollector<TCollector, TPredicate>
+where
+    TCollector: Collector + Send + Sync,
+    TPredicate: Fn(&[u8]) -> bool + Send + Sync + Clone,
+{
+    /// Create a new `BytesFilterCollector`.
+    pub fn new(field: Field, predicate: TPredicate, collector: TCollector) -> Self {
+        Self {
+            field,
+            predicate,
+            collector,
+        }
+    }
+}
+
+impl<TCollector, TPredicate> Collector for BytesFilterCollector<TCollector, TPredicate>
+where
+    TCollector: Collector + Send + Sync,
+    TPredicate: 'static + Fn(&[u8]) -> bool + Send + Sync + Clone,
+{
+    type Fruit = TCollector::Fruit;
+
+    type Child = BytesFilterSegmentCollector<TCollector::Child, TPredicate>;
+
+    fn for_segment(
+        &self,
+        segment_local_id: u32,
+        segment_reader: &SegmentReader,
+    ) -> crate::Result<Self::Child> {
+        let schema = segment_reader.schema();
+        let field_name = schema.get_field_name(self.field);
+
+        let column_opt = segment_reader.fast_fields().bytes(field_name)?;
+
+        let segment_collector = self
+            .collector
+            .for_segment(segment_local_id, segment_reader)?;
+
+        Ok(BytesFilterSegmentCollector {
+            column_opt,
+            segment_collector,
+            predicate: self.predicate.clone(),
+            buffer: Vec::new(),
+        })
+    }
+
+    fn requires_scoring(&self) -> bool {
+        self.collector.requires_scoring()
+    }
+
+    fn merge_fruits(
+        &self,
+        segment_fruits: Vec<<TCollector::Child as SegmentCollector>::Fruit>,
+    ) -> crate::Result<TCollector::Fruit> {
+        self.collector.merge_fruits(segment_fruits)
+    }
+}
+
+pub struct BytesFilterSegmentCollector<TSegmentCollector, TPredicate>
+where TPredicate: 'static
+{
+    column_opt: Option<BytesColumn>,
+    segment_collector: TSegmentCollector,
+    predicate: TPredicate,
+    buffer: Vec<u8>,
+}
+
+impl<TSegmentCollector, TPredicate> BytesFilterSegmentCollector<TSegmentCollector, TPredicate>
+where
+    TSegmentCollector: SegmentCollector,
+    TPredicate: 'static + Fn(&[u8]) -> bool + Send + Sync,
+{
+    #[inline]
+    fn accept_document(&mut self, doc_id: DocId) -> bool {
+        if let Some(column) = &self.column_opt {
+            for ord in column.term_ords(doc_id) {
+                self.buffer.clear();
+
+                let found = column.ord_to_bytes(ord, &mut self.buffer).unwrap_or(false);
+
+                if found && (self.predicate)(&self.buffer) {
+                    return true;
+                }
+            }
+        }
+        false
+    }
+}
+
+impl<TSegmentCollector, TPredicate> SegmentCollector
+    for BytesFilterSegmentCollector<TSegmentCollector, TPredicate>
+where
+    TSegmentCollector: SegmentCollector,
+    TPredicate: 'static + Fn(&[u8]) -> bool + Send + Sync,
+{
+    type Fruit = TSegmentCollector::Fruit;
+
+    fn collect(&mut self, doc: u32, score: Score) {
+        if self.accept_document(doc) {
+            self.segment_collector.collect(doc, score);
+        }
+    }
+
+    fn harvest(self) -> TSegmentCollector::Fruit {
         self.segment_collector.harvest()
     }
 }

src/collector/mod.rs

@@ -112,7 +112,7 @@ mod docset_collector;
 pub use self::docset_collector::DocSetCollector;
 
 mod filter_collector_wrapper;
-pub use self::filter_collector_wrapper::FilterCollector;
+pub use self::filter_collector_wrapper::{BytesFilterCollector, FilterCollector};
 
 /// `Fruit` is the type for the result of our collection.
 /// e.g. `usize` for the `Count` collector.

src/core/segment_reader.rs

@@ -2,8 +2,6 @@ use std::collections::HashMap;
 use std::sync::{Arc, RwLock};
 use std::{fmt, io};
 
-use fail::fail_point;
-
 use crate::core::{InvertedIndexReader, Segment, SegmentComponent, SegmentId};
 use crate::directory::{CompositeFile, FileSlice};
 use crate::error::DataCorruption;
@@ -151,7 +149,7 @@ impl SegmentReader {
 
         let store_file = segment.open_read(SegmentComponent::Store)?;
 
-        fail_point!("SegmentReader::open#middle");
+        crate::fail_point!("SegmentReader::open#middle");
 
         let postings_file = segment.open_read(SegmentComponent::Postings)?;
         let postings_composite = CompositeFile::open(&postings_file)?;

src/directory/ram_directory.rs

@@ -5,7 +5,6 @@ use std::sync::{Arc, RwLock};
 use std::{fmt, result};
 
 use common::HasLen;
-use fail::fail_point;
 
 use super::FileHandle;
 use crate::core::META_FILEPATH;
@@ -184,7 +183,7 @@ impl Directory for RamDirectory {
     }
 
     fn delete(&self, path: &Path) -> result::Result<(), DeleteError> {
-        fail_point!("RamDirectory::delete", |_| {
+        crate::fail_point!("RamDirectory::delete", |_| {
             Err(DeleteError::IoError {
                 io_error: Arc::new(io::Error::from(io::ErrorKind::Other)),
                 filepath: path.to_path_buf(),

src/indexer/segment_updater.rs

@@ -6,7 +6,6 @@ use std::path::PathBuf;
 use std::sync::atomic::{AtomicBool, Ordering};
 use std::sync::{Arc, RwLock};
 
-use fail::fail_point;
 use rayon::{ThreadPool, ThreadPoolBuilder};
 
 use super::segment_manager::SegmentManager;
@@ -43,7 +42,7 @@ pub(crate) fn save_metas(metas: &IndexMeta, directory: &dyn Directory) -> crate:
     let mut buffer = serde_json::to_vec_pretty(metas)?;
     // Just adding a new line at the end of the buffer.
     writeln!(&mut buffer)?;
-    fail_point!("save_metas", |msg| Err(crate::TantivyError::from(
+    crate::fail_point!("save_metas", |msg| Err(crate::TantivyError::from(
         std::io::Error::new(
             std::io::ErrorKind::Other,
             msg.unwrap_or_else(|| "Undefined".to_string())

src/indexer/segment_writer.rs

@@ -15,7 +15,7 @@ use crate::postings::{
 use crate::schema::{FieldEntry, FieldType, Schema, Term, Value, DATE_TIME_PRECISION_INDEXED};
 use crate::store::{StoreReader, StoreWriter};
 use crate::tokenizer::{FacetTokenizer, PreTokenizedStream, TextAnalyzer, Tokenizer};
-use crate::{DocId, Document, Opstamp, SegmentComponent};
+use crate::{DocId, Document, Opstamp, SegmentComponent, TantivyError};
 
 /// Computes the initial size of the hash table.
 ///
@@ -98,14 +98,18 @@ impl SegmentWriter {
                     }
                     _ => None,
                 };
-                text_options
-                    .and_then(|text_index_option| {
-                        let tokenizer_name = &text_index_option.tokenizer();
-                        tokenizer_manager.get(tokenizer_name)
-                    })
-                    .unwrap_or_default()
+                let tokenizer_name = text_options
+                    .map(|text_index_option| text_index_option.tokenizer())
+                    .unwrap_or("default");
+
+                tokenizer_manager.get(tokenizer_name).ok_or_else(|| {
+                    TantivyError::SchemaError(format!(
+                        "Error getting tokenizer for field: {}",
+                        field_entry.name()
+                    ))
+                })
             })
-            .collect();
+            .collect::<Result<Vec<_>, _>>()?;
         Ok(SegmentWriter {
             max_doc: 0,
             ctx: IndexingContext::new(table_size),
@@ -205,7 +209,7 @@ impl SegmentWriter {
                     for value in values {
                         let mut token_stream = match value {
                             Value::PreTokStr(tok_str) => {
-                                PreTokenizedStream::from(tok_str.clone()).into()
+                                Box::new(PreTokenizedStream::from(tok_str.clone()))
                             }
                             Value::Str(ref text) => {
                                 let text_analyzer =
@@ -438,7 +442,9 @@ fn remap_and_write(
 
 #[cfg(test)]
 mod tests {
-    use std::path::Path;
+    use std::path::{Path, PathBuf};
+
+    use tempfile::TempDir;
 
     use super::compute_initial_table_size;
     use crate::collector::Count;
@@ -446,7 +452,9 @@ mod tests {
     use crate::directory::RamDirectory;
     use crate::postings::TermInfo;
     use crate::query::PhraseQuery;
-    use crate::schema::{IndexRecordOption, Schema, Type, STORED, STRING, TEXT};
+    use crate::schema::{
+        IndexRecordOption, Schema, TextFieldIndexing, TextOptions, Type, STORED, STRING, TEXT,
+    };
     use crate::store::{Compressor, StoreReader, StoreWriter};
     use crate::time::format_description::well_known::Rfc3339;
     use crate::time::OffsetDateTime;
@@ -900,4 +908,32 @@ mod tests {
         postings.positions(&mut positions);
         assert_eq!(positions, &[4]); //< as opposed to 3 if we had a position length of 1.
     }
+
+    #[test]
+    fn test_show_error_when_tokenizer_not_registered() {
+        let text_field_indexing = TextFieldIndexing::default()
+            .set_tokenizer("custom_en")
+            .set_index_option(IndexRecordOption::WithFreqsAndPositions);
+        let text_options = TextOptions::default()
+            .set_indexing_options(text_field_indexing)
+            .set_stored();
+        let mut schema_builder = Schema::builder();
+        schema_builder.add_text_field("title", text_options);
+        let schema = schema_builder.build();
+        let tempdir = TempDir::new().unwrap();
+        let tempdir_path = PathBuf::from(tempdir.path());
+        Index::create_in_dir(&tempdir_path, schema).unwrap();
+        let index = Index::open_in_dir(tempdir_path).unwrap();
+        let schema = index.schema();
+        let mut index_writer = index.writer(50_000_000).unwrap();
+        let title = schema.get_field("title").unwrap();
+        let mut document = Document::default();
+        document.add_text(title, "The Old Man and the Sea");
+        index_writer.add_document(document).unwrap();
+        let error = index_writer.commit().unwrap_err();
+        assert_eq!(
+            error.to_string(),
+            "Schema error: 'Error getting tokenizer for field: title'"
+        );
+    }
 }

src/indexer/stamper.rs

@@ -101,6 +101,7 @@ mod test {
 
     use super::Stamper;
 
+    #[allow(clippy::redundant_clone)]
     #[test]
     fn test_stamper() {
         let stamper = Stamper::new(7u64);
@@ -116,6 +117,7 @@ mod test {
         assert_eq!(stamper.stamp(), 15u64);
     }
 
+    #[allow(clippy::redundant_clone)]
     #[test]
     fn test_stamper_revert() {
         let stamper = Stamper::new(7u64);

src/lib.rs

@@ -299,6 +299,35 @@ pub struct DocAddress {
     pub doc_id: DocId,
 }
 
+#[macro_export]
+/// Enable fail_point if feature is enabled.
+macro_rules! fail_point {
+    ($name:expr) => {{
+        #[cfg(feature = "failpoints")]
+        {
+            fail::eval($name, |_| {
+                panic!("Return is not supported for the fail point \"{}\"", $name);
+            });
+        }
+    }};
+    ($name:expr, $e:expr) => {{
+        #[cfg(feature = "failpoints")]
+        {
+            if let Some(res) = fail::eval($name, $e) {
+                return res;
+            }
+        }
+    }};
+    ($name:expr, $cond:expr, $e:expr) => {{
+        #[cfg(feature = "failpoints")]
+        {
+            if $cond {
+                fail::fail_point!($name, $e);
+            }
+        }
+    }};
+}
+
 #[cfg(test)]
 pub mod tests {
     use common::{BinarySerializable, FixedSize};
@@ -876,8 +905,8 @@ pub mod tests {
         }"#,
         )
         .unwrap();
-        let doc = doc!(json_field=>json_val.clone());
-        let index = Index::create_in_ram(schema.clone());
+        let doc = doc!(json_field=>json_val);
+        let index = Index::create_in_ram(schema);
         let mut writer = index.writer_for_tests().unwrap();
         writer.add_document(doc).unwrap();
         writer.commit().unwrap();

src/postings/serializer.rs

@@ -2,7 +2,6 @@ use std::cmp::Ordering;
 use std::io::{self, Write};
 
 use common::{BinarySerializable, CountingWriter, VInt};
-use fail::fail_point;
 
 use super::TermInfo;
 use crate::core::Segment;
@@ -205,7 +204,7 @@ impl<'a> FieldSerializer<'a> {
     /// If the current block is incomplete, it needs to be encoded
     /// using `VInt` encoding.
     pub fn close_term(&mut self) -> io::Result<()> {
-        fail_point!("FieldSerializer::close_term", |msg: Option<String>| {
+        crate::fail_point!("FieldSerializer::close_term", |msg: Option<String>| {
             Err(io::Error::new(io::ErrorKind::Other, format!("{msg:?}")))
         });
         if self.term_open {

src/query/more_like_this/more_like_this.rs

@@ -4,9 +4,7 @@ use std::collections::{BinaryHeap, HashMap};
 use crate::query::bm25::idf;
 use crate::query::{BooleanQuery, BoostQuery, Occur, Query, TermQuery};
 use crate::schema::{Field, FieldType, IndexRecordOption, Term, Value};
-use crate::tokenizer::{
-    BoxTokenStream, FacetTokenizer, PreTokenizedStream, TokenStream, Tokenizer,
-};
+use crate::tokenizer::{FacetTokenizer, PreTokenizedStream, TokenStream, Tokenizer};
 use crate::{DocAddress, Result, Searcher, TantivyError};
 
 #[derive(Debug, PartialEq)]
@@ -206,8 +204,7 @@ impl MoreLikeThis {
                 for value in values {
                     match value {
                         Value::PreTokStr(tok_str) => {
-                            let mut token_stream: BoxTokenStream =
-                                PreTokenizedStream::from(tok_str.clone()).into();
+                            let mut token_stream = PreTokenizedStream::from(tok_str.clone());
                             token_stream.process(&mut |token| {
                                 if !self.is_noise_word(token.text.clone()) {
                                     let term = Term::from_field_text(field, &token.text);

src/tokenizer/mod.rs

@@ -139,7 +139,7 @@ mod tokenizer;
 mod tokenizer_manager;
 mod whitespace_tokenizer;
 
-pub use tokenizer_api::{BoxTokenStream, Token, TokenFilter, TokenStream, Tokenizer};
+pub use tokenizer_api::{Token, TokenFilter, TokenStream, Tokenizer};
 
 pub use self::alphanum_only::AlphaNumOnlyFilter;
 pub use self::ascii_folding_filter::AsciiFoldingFilter;
@@ -154,7 +154,7 @@ pub use self::split_compound_words::SplitCompoundWords;
 pub use self::stemmer::{Language, Stemmer};
 pub use self::stop_word_filter::StopWordFilter;
 pub use self::tokenized_string::{PreTokenizedStream, PreTokenizedString};
-pub use self::tokenizer::TextAnalyzer;
+pub use self::tokenizer::{BoxTokenFilter, TextAnalyzer, TextAnalyzerBuilder};
 pub use self::tokenizer_manager::TokenizerManager;
 pub use self::whitespace_tokenizer::WhitespaceTokenizer;
 

src/tokenizer/tokenizer.rs

@@ -1,36 +1,105 @@
+use dyn_clone::DynClone;
 /// The tokenizer module contains all of the tools used to process
 /// text in `tantivy`.
-use tokenizer_api::{BoxTokenStream, TokenFilter, Tokenizer};
+use tokenizer_api::{TokenFilter, TokenStream, Tokenizer};
 
 use crate::tokenizer::empty_tokenizer::EmptyTokenizer;
 
 /// `TextAnalyzer` tokenizes an input text into tokens and modifies the resulting `TokenStream`.
+#[derive(Clone)]
 pub struct TextAnalyzer {
     tokenizer: Box<dyn BoxableTokenizer>,
 }
 
 /// A boxable `Tokenizer`, with its `TokenStream` type erased.
-trait BoxableTokenizer: 'static + Send + Sync {
+trait BoxableTokenizer: 'static + Send + Sync + DynClone {
     /// Creates a boxed token stream for a given `str`.
-    fn box_token_stream<'a>(&'a mut self, text: &'a str) -> BoxTokenStream<'a>;
-    /// Clone this tokenizer.
-    fn box_clone(&self) -> Box<dyn BoxableTokenizer>;
+    fn box_token_stream<'a>(&'a mut self, text: &'a str) -> Box<dyn TokenStream + 'a>;
 }
 
 impl<T: Tokenizer> BoxableTokenizer for T {
-    fn box_token_stream<'a>(&'a mut self, text: &'a str) -> BoxTokenStream<'a> {
-        self.token_stream(text).into()
-    }
-    fn box_clone(&self) -> Box<dyn BoxableTokenizer> {
-        Box::new(self.clone())
+    fn box_token_stream<'a>(&'a mut self, text: &'a str) -> Box<dyn TokenStream + 'a> {
+        Box::new(self.token_stream(text))
     }
 }
 
-impl Clone for TextAnalyzer {
-    fn clone(&self) -> Self {
-        TextAnalyzer {
-            tokenizer: self.tokenizer.box_clone(),
+dyn_clone::clone_trait_object!(BoxableTokenizer);
+
+/// A boxed `BoxableTokenizer` which is a `Tokenizer` with its `TokenStream` type erased.
+#[derive(Clone)]
+struct BoxTokenizer(Box<dyn BoxableTokenizer>);
+
+impl Tokenizer for BoxTokenizer {
+    type TokenStream<'a> = Box<dyn TokenStream + 'a>;
+
+    fn token_stream<'a>(&'a mut self, text: &'a str) -> Self::TokenStream<'a> {
+        self.0.box_token_stream(text).into()
+    }
+}
+
+/// A boxable `TokenFilter`, with its `Tokenizer` type erased.
+trait BoxableTokenFilter: 'static + Send + Sync {
+    /// Wraps a `BoxedTokenizer` and returns a new one.
+    fn box_transform(&self, tokenizer: BoxTokenizer) -> BoxTokenizer;
+}
+
+impl<T: TokenFilter> BoxableTokenFilter for T {
+    fn box_transform(&self, tokenizer: BoxTokenizer) -> BoxTokenizer {
+        let tokenizer = self.clone().transform(tokenizer);
+        BoxTokenizer(Box::new(tokenizer))
+    }
+}
+
+/// A boxed `BoxableTokenFilter` which is a `TokenFilter` with its `Tokenizer` type erased.
+pub struct BoxTokenFilter(Box<dyn BoxableTokenFilter>);
+
+impl<T: TokenFilter> From<T> for BoxTokenFilter {
+    fn from(tokenizer: T) -> BoxTokenFilter {
+        BoxTokenFilter(Box::new(tokenizer))
+    }
+}
+
+impl TextAnalyzer {
+    /// Builds a new `TextAnalyzer` given a tokenizer and a vector of `BoxTokenFilter`.
+    ///
+    /// When creating a `TextAnalyzer` from a `Tokenizer` and a static set of `TokenFilter`,
+    /// prefer using `TextAnalyzer::builder(tokenizer).filter(token_filter).build()` as it
+    /// will be more performant and create less boxes.
+    ///
+    /// # Example
+    ///
+    /// ```rust
+    /// use tantivy::tokenizer::*;
+    ///
+    /// let en_stem = TextAnalyzer::build(
+    ///     SimpleTokenizer::default(),
+    ///     vec![
+    ///        BoxTokenFilter::from(RemoveLongFilter::limit(40)),
+    ///        BoxTokenFilter::from(LowerCaser),
+    ///        BoxTokenFilter::from(Stemmer::default()),
+    ///     ]);
+    /// ```
+    pub fn build<T: Tokenizer>(
+        tokenizer: T,
+        boxed_token_filters: Vec<BoxTokenFilter>,
+    ) -> TextAnalyzer {
+        let mut boxed_tokenizer = BoxTokenizer(Box::new(tokenizer));
+        for filter in boxed_token_filters.into_iter() {
+            boxed_tokenizer = filter.0.box_transform(boxed_tokenizer);
         }
+        TextAnalyzer {
+            tokenizer: boxed_tokenizer.0,
+        }
+    }
+
+    /// Create a new TextAnalyzerBuilder
+    pub fn builder<T: Tokenizer>(tokenizer: T) -> TextAnalyzerBuilder<T> {
+        TextAnalyzerBuilder { tokenizer }
+    }
+
+    /// Creates a token stream for a given `str`.
+    pub fn token_stream<'a>(&'a mut self, text: &'a str) -> Box<dyn TokenStream + 'a> {
+        self.tokenizer.box_token_stream(text)
     }
 }
 
@@ -46,20 +115,8 @@ impl<T: Tokenizer + Clone> From<T> for TextAnalyzer {
     }
 }
 
-impl TextAnalyzer {
-    /// Create a new TextAnalyzerBuilder
-    pub fn builder<T: Tokenizer>(tokenizer: T) -> TextAnalyzerBuilder<T> {
-        TextAnalyzerBuilder { tokenizer }
-    }
-
-    /// Creates a token stream for a given `str`.
-    pub fn token_stream<'a>(&'a mut self, text: &'a str) -> BoxTokenStream<'a> {
-        self.tokenizer.box_token_stream(text)
-    }
-}
-
 /// Builder helper for [`TextAnalyzer`]
-pub struct TextAnalyzerBuilder<T> {
+pub struct TextAnalyzerBuilder<T: Tokenizer> {
     tokenizer: T,
 }
 
@@ -90,3 +147,37 @@ impl<T: Tokenizer> TextAnalyzerBuilder<T> {
         }
     }
 }
+
+#[cfg(test)]
+mod tests {
+
+    use super::*;
+    use crate::tokenizer::{AlphaNumOnlyFilter, LowerCaser, RemoveLongFilter, WhitespaceTokenizer};
+
+    #[test]
+    fn test_text_analyzer_builder() {
+        let mut analyzer = TextAnalyzer::builder(WhitespaceTokenizer::default())
+            .filter(AlphaNumOnlyFilter)
+            .filter(RemoveLongFilter::limit(6))
+            .filter(LowerCaser)
+            .build();
+        let mut stream = analyzer.token_stream("- first bullet point");
+        assert_eq!(stream.next().unwrap().text, "first");
+        assert_eq!(stream.next().unwrap().text, "point");
+    }
+
+    #[test]
+    fn test_text_analyzer_with_filters_boxed() {
+        let mut analyzer = TextAnalyzer::build(
+            WhitespaceTokenizer::default(),
+            vec![
+                BoxTokenFilter::from(AlphaNumOnlyFilter),
+                BoxTokenFilter::from(LowerCaser),
+                BoxTokenFilter::from(RemoveLongFilter::limit(6)),
+            ],
+        );
+        let mut stream = analyzer.token_stream("- first bullet point");
+        assert_eq!(stream.next().unwrap().text, "first");
+        assert_eq!(stream.next().unwrap().text, "point");
+    }
+}

sstable/Cargo.toml

@@ -7,7 +7,7 @@ homepage = "https://github.com/quickwit-oss/tantivy"
 repository = "https://github.com/quickwit-oss/tantivy"
 keywords = ["search", "information", "retrieval", "sstable"]
 categories = ["database-implementations", "data-structures", "compression"]
-desciption = "sstables for tantivy"
+description = "sstables for tantivy"
 
 [dependencies]
 common = {version= "0.5", path="../common", package="tantivy-common"}

stacker/src/fastcpy.rs

@@ -44,7 +44,7 @@ pub fn fast_short_slice_copy(src: &[u8], dst: &mut [u8]) {
         return;
     }
 
-    /// The code will use the vmovdqu instruction to copy 32 bytes at a time.
+    // The code will use the vmovdqu instruction to copy 32 bytes at a time.
     #[cfg(target_feature = "avx")]
     {
         if len <= 64 {

tokenizer-api/src/lib.rs

@@ -6,7 +6,6 @@
 //! Checkout the [tantivy repo](https://github.com/quickwit-oss/tantivy/tree/main/src/tokenizer) for some examples.
 
 use std::borrow::{Borrow, BorrowMut};
-use std::ops::{Deref, DerefMut};
 
 use serde::{Deserialize, Serialize};
 
@@ -60,30 +59,6 @@ pub trait Tokenizer: 'static + Clone + Send + Sync {
     fn token_stream<'a>(&'a mut self, text: &'a str) -> Self::TokenStream<'a>;
 }
 
-/// Simple wrapper of `Box<dyn TokenStream + 'a>`.
-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
-    }
-}
-
 impl<'a> TokenStream for Box<dyn TokenStream + 'a> {
     fn advance(&mut self) -> bool {
         let token_stream: &mut dyn TokenStream = self.borrow_mut();
@@ -137,7 +112,7 @@ pub trait TokenStream {
 }
 
 /// Trait for the pluggable components of `Tokenizer`s.
-pub trait TokenFilter: 'static + Send + Sync {
+pub trait TokenFilter: 'static + Send + Sync + Clone {
     /// The Tokenizer type returned by this filter, typically parametrized by the underlying
     /// Tokenizer.
     type Tokenizer<T: Tokenizer>: Tokenizer;