Extract and index data

milli/src/update/index_documents/extract/extract_docid_word_positions.rs

@@ -1,9 +1,9 @@
-use std::collections::HashSet;
+use std::collections::{HashMap, HashSet};
 use std::convert::TryInto;
 use std::fs::File;
 use std::{io, mem, str};
 
-use charabia::{SeparatorKind, Token, TokenKind, TokenizerBuilder};
+use charabia::{Language, Script, SeparatorKind, Token, TokenKind, TokenizerBuilder};
 use roaring::RoaringBitmap;
 use serde_json::Value;
 
@@ -25,12 +25,13 @@ pub fn extract_docid_word_positions<R: io::Read + io::Seek>(
     searchable_fields: &Option<HashSet<FieldId>>,
     stop_words: Option<&fst::Set<&[u8]>>,
     max_positions_per_attributes: Option<u32>,
-) -> Result<(RoaringBitmap, grenad::Reader<File>)> {
+) -> Result<(RoaringBitmap, grenad::Reader<File>, HashMap<(Script, Language), RoaringBitmap>)> {
     let max_positions_per_attributes = max_positions_per_attributes
         .map_or(MAX_POSITION_PER_ATTRIBUTE, |max| max.min(MAX_POSITION_PER_ATTRIBUTE));
     let max_memory = indexer.max_memory_by_thread();
 
     let mut documents_ids = RoaringBitmap::new();
+    let mut script_language_pair = HashMap::new();
     let mut docid_word_positions_sorter = create_sorter(
         grenad::SortAlgorithm::Stable,
         concat_u32s_array,
@@ -70,6 +71,12 @@ pub fn extract_docid_word_positions<R: io::Read + io::Seek>(
                         .take_while(|(p, _)| (*p as u32) < max_positions_per_attributes);
 
                     for (index, token) in tokens {
+                        let script = token.script;
+                        let language = token.language.unwrap_or_default();
+                        let entry = script_language_pair
+                            .entry((script, language))
+                            .or_insert_with(RoaringBitmap::new);
+                        entry.push(document_id);
                         let token = token.lemma().trim();
                         if !token.is_empty() && token.len() <= MAX_WORD_LENGTH {
                             key_buffer.truncate(mem::size_of::<u32>());
@@ -88,7 +95,7 @@ pub fn extract_docid_word_positions<R: io::Read + io::Seek>(
         }
     }
 
-    sorter_into_reader(docid_word_positions_sorter, indexer).map(|reader| (documents_ids, reader))
+    sorter_into_reader(docid_word_positions_sorter, indexer).map(|reader| (documents_ids, reader, script_language_pair))
 }
 
 /// Transform a JSON value into a string that can be indexed.

milli/src/update/index_documents/extract/mod.rs

@@ -257,7 +257,7 @@ fn send_and_extract_flattened_documents_data(
     let (docid_word_positions_chunk, docid_fid_facet_values_chunks): (Result<_>, Result<_>) =
         rayon::join(
             || {
-                let (documents_ids, docid_word_positions_chunk) = extract_docid_word_positions(
+                let (documents_ids, docid_word_positions_chunk, script_language_pair) = extract_docid_word_positions(
                     flattened_documents_chunk.clone(),
                     indexer,
                     searchable_fields,
@@ -274,6 +274,8 @@ fn send_and_extract_flattened_documents_data(
                 let _ = lmdb_writer_sx
                     .send(Ok(TypedChunk::DocidWordPositions(docid_word_positions_chunk.clone())));
 
+                let _ = lmdb_writer_sx.send(Ok(TypedChunk::ScriptLanguageDocids(script_language_pair)));
+
                 Ok(docid_word_positions_chunk)
             },
             || {

milli/src/update/index_documents/typed_chunk.rs

@@ -1,8 +1,10 @@
 use std::borrow::Cow;
+use std::collections::HashMap;
 use std::convert::TryInto;
 use std::fs::File;
 use std::io;
 
+use charabia::{Language, Script};
 use grenad::MergerBuilder;
 use heed::types::ByteSlice;
 use heed::{BytesDecode, RwTxn};
@@ -16,10 +18,7 @@ use super::{ClonableMmap, MergeFn};
 use crate::facet::FacetType;
 use crate::update::facet::FacetsUpdate;
 use crate::update::index_documents::helpers::as_cloneable_grenad;
-use crate::{
+use crate::{BoRoaringBitmapCodec, CboRoaringBitmapCodec, DocumentId, GeoPoint, Index, Result, lat_lng_to_xyz};
-    lat_lng_to_xyz, BoRoaringBitmapCodec, CboRoaringBitmapCodec, DocumentId, GeoPoint, Index,
-    Result,
-};
 
 pub(crate) enum TypedChunk {
     DocidWordPositions(grenad::Reader<CursorClonableMmap>),
@@ -38,6 +37,7 @@ pub(crate) enum TypedChunk {
     FieldIdFacetNumberDocids(grenad::Reader<File>),
     FieldIdFacetExistsDocids(grenad::Reader<File>),
     GeoPoints(grenad::Reader<File>),
+    ScriptLanguageDocids(HashMap<(Script, Language), RoaringBitmap>)
 }
 
 /// Write typed chunk in the corresponding LMDB database of the provided index.
@@ -210,6 +210,25 @@ pub(crate) fn write_typed_chunk_into_index(
             index.put_geo_rtree(wtxn, &rtree)?;
             index.put_geo_faceted_documents_ids(wtxn, &geo_faceted_docids)?;
         }
+        TypedChunk::ScriptLanguageDocids(hash_pair) => {
+            let mut buffer = Vec::new();
+            for (key, value) in hash_pair {
+                buffer.clear();
+                let final_value = match index.script_language_docids.get(wtxn, &key)? {
+                    Some(db_values) => {
+                        let mut db_value_buffer = Vec::new();
+                        serialize_roaring_bitmap(&db_values, &mut db_value_buffer)?;
+                        let mut new_value_buffer = Vec::new();
+                        serialize_roaring_bitmap(&value, &mut new_value_buffer)?;
+                        merge_roaring_bitmaps(&new_value_buffer, &db_value_buffer, &mut buffer)?;
+                        let merged_db_values = RoaringBitmap::deserialize_from(&buffer[..])?;
+                        merged_db_values
+                    }
+                    None => value
+                };
+                index.script_language_docids.put(wtxn, &key, &final_value)?;
+            }
+        } 
     }
 
     Ok((RoaringBitmap::new(), is_merged_database))