Only insert facets in bulks no more incrementally

4318: Hide embedders r=ManyTheFish a=dureuill

Hides `embedders` when it is an empty dictionary.

Manual tests:

- getting settings with empty embedders: not displayed
- getting settings with non-empty embedders: displayed like before
- dump with empty embedders: can be imported
- dump with non-empty embedders: can be imported

Co-authored-by: Louis Dureuil <louis@meilisearch.com>

meilisearch-types/src/settings.rs

@@ -600,11 +600,12 @@ pub fn settings(
         ),
     };
 
-    let embedders = index
+    let embedders: BTreeMap<_, _> = index
         .embedding_configs(rtxn)?
         .into_iter()
         .map(|(name, config)| (name, Setting::Set(config.into())))
         .collect();
+    let embedders = if embedders.is_empty() { Setting::NotSet } else { Setting::Set(embedders) };
 
     Ok(Settings {
         displayed_attributes: match displayed_attributes {
@@ -631,7 +632,7 @@ pub fn settings(
         typo_tolerance: Setting::Set(typo_tolerance),
         faceting: Setting::Set(faceting),
         pagination: Setting::Set(pagination),
-        embedders: Setting::Set(embedders),
+        embedders,
         _kind: PhantomData,
     })
 }

meilisearch/src/routes/indexes/settings.rs

@@ -458,7 +458,7 @@ make_setting_route!(
             json!({
                 "proximity_precision": {
                     "set": precision.is_some(),
-                    "value": precision,
+                    "value": precision.unwrap_or_default(),
                 }
             }),
             Some(req),
@@ -690,7 +690,8 @@ pub async fn update_all(
                 "set": new_settings.distinct_attribute.as_ref().set().is_some()
             },
             "proximity_precision": {
-                "set": new_settings.proximity_precision.as_ref().set().is_some()
+                "set": new_settings.proximity_precision.as_ref().set().is_some(),
+                "value": new_settings.proximity_precision.as_ref().set().copied().unwrap_or_default()
             },
             "typo_tolerance": {
                 "enabled": new_settings.typo_tolerance

meilisearch/src/search.rs

@@ -735,6 +735,9 @@ pub fn perform_facet_search(
     if let Some(facet_query) = &facet_query {
         facet_search.query(facet_query);
     }
+    if let Some(max_facets) = index.max_values_per_facet(&rtxn)? {
+        facet_search.max_values(max_facets as usize);
+    }
 
     Ok(FacetSearchResult {
         facet_hits: facet_search.execute()?,
@@ -897,6 +900,14 @@ fn format_fields<'a>(
     let mut matches_position = compute_matches.then(BTreeMap::new);
     let mut document = document.clone();
 
+    // reduce the formatted option list to the attributes that should be formatted,
+    // instead of all the attributes to display.
+    let formatting_fields_options: Vec<_> = formatted_options
+        .iter()
+        .filter(|(_, option)| option.should_format())
+        .map(|(fid, option)| (field_ids_map.name(*fid).unwrap(), option))
+        .collect();
+
     // select the attributes to retrieve
     let displayable_names =
         displayable_ids.iter().map(|&fid| field_ids_map.name(fid).expect("Missing field name"));
@@ -905,13 +916,15 @@ fn format_fields<'a>(
         // to the value and merge them together. eg. If a user said he wanted to highlight `doggo`
         // and crop `doggo.name`. `doggo.name` needs to be highlighted + cropped while `doggo.age` is only
         // highlighted.
-        let format = formatted_options
+        // Warn: The time to compute the format list scales with the number of fields to format;
+        // cumulated with map_leaf_values that iterates over all the nested fields, it gives a quadratic complexity:
+        // d*f where d is the total number of fields to display and f is the total number of fields to format.
+        let format = formatting_fields_options
             .iter()
-            .filter(|(field, _option)| {
-                let name = field_ids_map.name(**field).unwrap();
+            .filter(|(name, _option)| {
                 milli::is_faceted_by(name, key) || milli::is_faceted_by(key, name)
             })
-            .map(|(_, option)| *option)
+            .map(|(_, option)| **option)
             .reduce(|acc, option| acc.merge(option));
         let mut infos = Vec::new();
 
@@ -1008,7 +1021,7 @@ fn format_value<'a>(
                     let value = matcher.format(format_options);
                     Value::String(value.into_owned())
                 }
-                None => Value::Number(number),
+                None => Value::String(s),
             }
         }
         value => value,

meilisearch/tests/dumps/mod.rs

@@ -77,8 +77,7 @@ async fn import_dump_v1_movie_raw() {
       },
       "pagination": {
         "maxTotalHits": 1000
-      },
-      "embedders": {}
+      }
     }
     "###
     );
@@ -239,8 +238,7 @@ async fn import_dump_v1_movie_with_settings() {
       },
       "pagination": {
         "maxTotalHits": 1000
-      },
-      "embedders": {}
+      }
     }
     "###
     );
@@ -387,8 +385,7 @@ async fn import_dump_v1_rubygems_with_settings() {
       },
       "pagination": {
         "maxTotalHits": 1000
-      },
-      "embedders": {}
+      }
     }
     "###
     );
@@ -521,8 +518,7 @@ async fn import_dump_v2_movie_raw() {
       },
       "pagination": {
         "maxTotalHits": 1000
-      },
-      "embedders": {}
+      }
     }
     "###
     );
@@ -667,8 +663,7 @@ async fn import_dump_v2_movie_with_settings() {
       },
       "pagination": {
         "maxTotalHits": 1000
-      },
-      "embedders": {}
+      }
     }
     "###
     );
@@ -812,8 +807,7 @@ async fn import_dump_v2_rubygems_with_settings() {
       },
       "pagination": {
         "maxTotalHits": 1000
-      },
-      "embedders": {}
+      }
     }
     "###
     );
@@ -946,8 +940,7 @@ async fn import_dump_v3_movie_raw() {
       },
       "pagination": {
         "maxTotalHits": 1000
-      },
-      "embedders": {}
+      }
     }
     "###
     );
@@ -1092,8 +1085,7 @@ async fn import_dump_v3_movie_with_settings() {
       },
       "pagination": {
         "maxTotalHits": 1000
-      },
-      "embedders": {}
+      }
     }
     "###
     );
@@ -1237,8 +1229,7 @@ async fn import_dump_v3_rubygems_with_settings() {
       },
       "pagination": {
         "maxTotalHits": 1000
-      },
-      "embedders": {}
+      }
     }
     "###
     );
@@ -1371,8 +1362,7 @@ async fn import_dump_v4_movie_raw() {
       },
       "pagination": {
         "maxTotalHits": 1000
-      },
-      "embedders": {}
+      }
     }
     "###
     );
@@ -1517,8 +1507,7 @@ async fn import_dump_v4_movie_with_settings() {
       },
       "pagination": {
         "maxTotalHits": 1000
-      },
-      "embedders": {}
+      }
     }
     "###
     );
@@ -1662,8 +1651,7 @@ async fn import_dump_v4_rubygems_with_settings() {
       },
       "pagination": {
         "maxTotalHits": 1000
-      },
-      "embedders": {}
+      }
     }
     "###
     );
@@ -1907,8 +1895,7 @@ async fn import_dump_v6_containing_experimental_features() {
       },
       "pagination": {
         "maxTotalHits": 1000
-      },
-      "embedders": {}
+      }
     }
     "###);
 

meilisearch/tests/search/facet_search.rs

@@ -105,6 +105,24 @@ async fn more_advanced_facet_search() {
     snapshot!(response["facetHits"].as_array().unwrap().len(), @"1");
 }
 
+#[actix_rt::test]
+async fn simple_facet_search_with_max_values() {
+    let server = Server::new().await;
+    let index = server.index("test");
+
+    let documents = DOCUMENTS.clone();
+    index.update_settings_faceting(json!({ "maxValuesPerFacet": 1 })).await;
+    index.update_settings_filterable_attributes(json!(["genres"])).await;
+    index.add_documents(documents, None).await;
+    index.wait_task(2).await;
+
+    let (response, code) =
+        index.facet_search(json!({"facetName": "genres", "facetQuery": "a"})).await;
+
+    assert_eq!(code, 200, "{}", response);
+    assert_eq!(dbg!(response)["facetHits"].as_array().unwrap().len(), 1);
+}
+
 #[actix_rt::test]
 async fn non_filterable_facet_search_error() {
     let server = Server::new().await;

meilisearch/tests/settings/get_settings.rs

@@ -54,7 +54,7 @@ async fn get_settings() {
     let (response, code) = index.settings().await;
     assert_eq!(code, 200);
     let settings = response.as_object().unwrap();
-    assert_eq!(settings.keys().len(), 16);
+    assert_eq!(settings.keys().len(), 15);
     assert_eq!(settings["displayedAttributes"], json!(["*"]));
     assert_eq!(settings["searchableAttributes"], json!(["*"]));
     assert_eq!(settings["filterableAttributes"], json!([]));
@@ -83,7 +83,6 @@ async fn get_settings() {
             "maxTotalHits": 1000,
         })
     );
-    assert_eq!(settings["embedders"], json!({}));
     assert_eq!(settings["proximityPrecision"], json!("byWord"));
 }
 

milli/src/search/mod.rs

@@ -27,8 +27,8 @@ static LEVDIST0: Lazy<LevBuilder> = Lazy::new(|| LevBuilder::new(0, true));
 static LEVDIST1: Lazy<LevBuilder> = Lazy::new(|| LevBuilder::new(1, true));
 static LEVDIST2: Lazy<LevBuilder> = Lazy::new(|| LevBuilder::new(2, true));
 
-/// The maximum number of facets returned by the facet search route.
-const MAX_NUMBER_OF_FACETS: usize = 100;
+/// The maximum number of values per facet returned by the facet search route.
+const DEFAULT_MAX_NUMBER_OF_VALUES_PER_FACET: usize = 100;
 
 pub mod facet;
 mod fst_utils;
@@ -306,6 +306,7 @@ pub struct SearchForFacetValues<'a> {
     query: Option<String>,
     facet: String,
     search_query: Search<'a>,
+    max_values: usize,
     is_hybrid: bool,
 }
 
@@ -315,7 +316,13 @@ impl<'a> SearchForFacetValues<'a> {
         search_query: Search<'a>,
         is_hybrid: bool,
     ) -> SearchForFacetValues<'a> {
-        SearchForFacetValues { query: None, facet, search_query, is_hybrid }
+        SearchForFacetValues {
+            query: None,
+            facet,
+            search_query,
+            max_values: DEFAULT_MAX_NUMBER_OF_VALUES_PER_FACET,
+            is_hybrid,
+        }
     }
 
     pub fn query(&mut self, query: impl Into<String>) -> &mut Self {
@@ -323,6 +330,11 @@ impl<'a> SearchForFacetValues<'a> {
         self
     }
 
+    pub fn max_values(&mut self, max: usize) -> &mut Self {
+        self.max_values = max;
+        self
+    }
+
     fn one_original_value_of(
         &self,
         field_id: FieldId,
@@ -462,7 +474,7 @@ impl<'a> SearchForFacetValues<'a> {
                             .unwrap_or_else(|| left_bound.to_string());
                         results.push(FacetValueHit { value, count });
                     }
-                    if results.len() >= MAX_NUMBER_OF_FACETS {
+                    if results.len() >= self.max_values {
                         break;
                     }
                 }
@@ -507,7 +519,7 @@ impl<'a> SearchForFacetValues<'a> {
                     .unwrap_or_else(|| query.to_string());
                 results.push(FacetValueHit { value, count });
             }
-            if results.len() >= MAX_NUMBER_OF_FACETS {
+            if results.len() >= self.max_values {
                 return Ok(ControlFlow::Break(()));
             }
         }

milli/src/search/new/bucket_sort.rs

@@ -15,6 +15,7 @@ pub struct BucketSortOutput {
 
 // TODO: would probably be good to regroup some of these inside of a struct?
 #[allow(clippy::too_many_arguments)]
+#[logging_timer::time]
 pub fn bucket_sort<'ctx, Q: RankingRuleQueryTrait>(
     ctx: &mut SearchContext<'ctx>,
     mut ranking_rules: Vec<BoxRankingRule<'ctx, Q>>,

milli/src/search/new/matches/mod.rs

@@ -72,7 +72,7 @@ impl<'m> MatcherBuilder<'m> {
     }
 }
 
-#[derive(Copy, Clone, Default)]
+#[derive(Copy, Clone, Default, Debug)]
 pub struct FormatOptions {
     pub highlight: bool,
     pub crop: Option<usize>,
@@ -82,6 +82,10 @@ impl FormatOptions {
     pub fn merge(self, other: Self) -> Self {
         Self { highlight: self.highlight || other.highlight, crop: self.crop.or(other.crop) }
     }
+
+    pub fn should_format(&self) -> bool {
+        self.highlight || self.crop.is_some()
+    }
 }
 
 #[derive(Clone, Debug)]

milli/src/search/new/mod.rs

@@ -191,6 +191,7 @@ fn resolve_maximally_reduced_query_graph(
     Ok(docids)
 }
 
+#[logging_timer::time]
 fn resolve_universe(
     ctx: &mut SearchContext,
     initial_universe: &RoaringBitmap,
@@ -556,6 +557,7 @@ pub fn execute_vector_search(
 }
 
 #[allow(clippy::too_many_arguments)]
+#[logging_timer::time]
 pub fn execute_search(
     ctx: &mut SearchContext,
     query: Option<&str>,

milli/src/search/new/query_term/parse_query.rs

@@ -5,6 +5,7 @@ use super::*;
 use crate::{Result, SearchContext, MAX_WORD_LENGTH};
 
 /// Convert the tokenised search query into a list of located query terms.
+#[logging_timer::time]
 pub fn located_query_terms_from_tokens(
     ctx: &mut SearchContext,
     query: NormalizedTokenIter,

milli/src/update/facet/incremental.rs

@@ -61,6 +61,7 @@ impl FacetsUpdateIncremental {
         }
     }
 
+    #[logging_timer::time("FacetsUpdateIncremental::{}")]
     pub fn execute(self, wtxn: &mut RwTxn) -> crate::Result<()> {
         let mut cursor = self.delta_data.into_cursor()?;
         while let Some((key, value)) = cursor.move_on_next()? {

milli/src/update/facet/mod.rs

@@ -76,26 +76,18 @@ pub const FACET_MAX_GROUP_SIZE: u8 = 8;
 pub const FACET_GROUP_SIZE: u8 = 4;
 pub const FACET_MIN_LEVEL_SIZE: u8 = 5;
 
-use std::collections::BTreeSet;
 use std::fs::File;
 use std::io::BufReader;
-use std::iter::FromIterator;
 
-use charabia::normalizer::{Normalize, NormalizerOption};
-use grenad::{CompressionType, SortAlgorithm};
-use heed::types::{Bytes, DecodeIgnore, SerdeJson};
-use heed::BytesEncode;
 use log::debug;
 use time::OffsetDateTime;
 
 use self::incremental::FacetsUpdateIncremental;
 use super::FacetsUpdateBulk;
 use crate::facet::FacetType;
-use crate::heed_codec::facet::{FacetGroupKey, FacetGroupKeyCodec, FacetGroupValueCodec};
+use crate::heed_codec::facet::{FacetGroupKeyCodec, FacetGroupValueCodec};
 use crate::heed_codec::BytesRefCodec;
-use crate::update::index_documents::create_sorter;
-use crate::update::merge_btreeset_string;
-use crate::{BEU16StrCodec, Index, Result, MAX_FACET_VALUE_LENGTH};
+use crate::{Index, Result};
 
 pub mod bulk;
 pub mod incremental;
@@ -146,115 +138,114 @@ impl<'i> FacetsUpdate<'i> {
         self.index.set_updated_at(wtxn, &OffsetDateTime::now_utc())?;
 
         // See self::comparison_bench::benchmark_facet_indexing
-        if self.delta_data.len() >= (self.database.len(wtxn)? / 50) {
-            let field_ids =
-                self.index.faceted_fields_ids(wtxn)?.iter().copied().collect::<Vec<_>>();
-            let bulk_update = FacetsUpdateBulk::new(
-                self.index,
-                field_ids,
-                self.facet_type,
-                self.delta_data,
-                self.group_size,
-                self.min_level_size,
-            );
-            bulk_update.execute(wtxn)?;
-        } else {
-            let incremental_update = FacetsUpdateIncremental::new(
-                self.index,
-                self.facet_type,
-                self.delta_data,
-                self.group_size,
-                self.min_level_size,
-                self.max_group_size,
-            );
-            incremental_update.execute(wtxn)?;
-        }
-
-        // We clear the list of normalized-for-search facets
-        // and the previous FSTs to compute everything from scratch
-        self.index.facet_id_normalized_string_strings.clear(wtxn)?;
-        self.index.facet_id_string_fst.clear(wtxn)?;
-
-        // As we can't use the same write transaction to read and write in two different databases
-        // we must create a temporary sorter that we will write into LMDB afterward.
-        // As multiple unnormalized facet values can become the same normalized facet value
-        // we must merge them together.
-        let mut sorter = create_sorter(
-            SortAlgorithm::Unstable,
-            merge_btreeset_string,
-            CompressionType::None,
-            None,
-            None,
-            None,
+        // if self.delta_data.len() >= (self.database.len(wtxn)? / 50) {
+        let field_ids = self.index.faceted_fields_ids(wtxn)?.iter().copied().collect::<Vec<_>>();
+        let bulk_update = FacetsUpdateBulk::new(
+            self.index,
+            field_ids,
+            self.facet_type,
+            self.delta_data,
+            self.group_size,
+            self.min_level_size,
         );
+        bulk_update.execute(wtxn)?;
+        // } else {
+        //     let incremental_update = FacetsUpdateIncremental::new(
+        //         self.index,
+        //         self.facet_type,
+        //         self.delta_data,
+        //         self.group_size,
+        //         self.min_level_size,
+        //         self.max_group_size,
+        //     );
+        //     incremental_update.execute(wtxn)?;
+        // }
 
-        // We iterate on the list of original, semi-normalized, facet values
-        // and normalize them for search, inserting them in LMDB in any given order.
-        let options = NormalizerOption { lossy: true, ..Default::default() };
-        let database = self.index.facet_id_string_docids.remap_data_type::<DecodeIgnore>();
-        for result in database.iter(wtxn)? {
-            let (facet_group_key, ()) = result?;
-            if let FacetGroupKey { field_id, level: 0, left_bound } = facet_group_key {
-                let mut normalized_facet = left_bound.normalize(&options);
-                let normalized_truncated_facet: String;
-                if normalized_facet.len() > MAX_FACET_VALUE_LENGTH {
-                    normalized_truncated_facet = normalized_facet
-                        .char_indices()
-                        .take_while(|(idx, _)| *idx < MAX_FACET_VALUE_LENGTH)
-                        .map(|(_, c)| c)
-                        .collect();
-                    normalized_facet = normalized_truncated_facet.into();
-                }
-                let set = BTreeSet::from_iter(std::iter::once(left_bound));
-                let key = (field_id, normalized_facet.as_ref());
-                let key = BEU16StrCodec::bytes_encode(&key).map_err(heed::Error::Encoding)?;
-                let val = SerdeJson::bytes_encode(&set).map_err(heed::Error::Encoding)?;
-                sorter.insert(key, val)?;
-            }
-        }
+        // // We clear the list of normalized-for-search facets
+        // // and the previous FSTs to compute everything from scratch
+        // self.index.facet_id_normalized_string_strings.clear(wtxn)?;
+        // self.index.facet_id_string_fst.clear(wtxn)?;
 
-        // In this loop we don't need to take care of merging bitmaps
-        // as the grenad sorter already merged them for us.
-        let mut merger_iter = sorter.into_stream_merger_iter()?;
-        while let Some((key_bytes, btreeset_bytes)) = merger_iter.next()? {
-            self.index.facet_id_normalized_string_strings.remap_types::<Bytes, Bytes>().put(
-                wtxn,
-                key_bytes,
-                btreeset_bytes,
-            )?;
-        }
+        // // As we can't use the same write transaction to read and write in two different databases
+        // // we must create a temporary sorter that we will write into LMDB afterward.
+        // // As multiple unnormalized facet values can become the same normalized facet value
+        // // we must merge them together.
+        // let mut sorter = create_sorter(
+        //     SortAlgorithm::Unstable,
+        //     merge_btreeset_string,
+        //     CompressionType::None,
+        //     None,
+        //     None,
+        //     None,
+        // );
 
-        // We compute one FST by string facet
-        let mut text_fsts = vec![];
-        let mut current_fst: Option<(u16, fst::SetBuilder<Vec<u8>>)> = None;
-        let database =
-            self.index.facet_id_normalized_string_strings.remap_data_type::<DecodeIgnore>();
-        for result in database.iter(wtxn)? {
-            let ((field_id, normalized_facet), _) = result?;
-            current_fst = match current_fst.take() {
-                Some((fid, fst_builder)) if fid != field_id => {
-                    let fst = fst_builder.into_set();
-                    text_fsts.push((fid, fst));
-                    Some((field_id, fst::SetBuilder::memory()))
-                }
-                Some((field_id, fst_builder)) => Some((field_id, fst_builder)),
-                None => Some((field_id, fst::SetBuilder::memory())),
-            };
+        // // We iterate on the list of original, semi-normalized, facet values
+        // // and normalize them for search, inserting them in LMDB in any given order.
+        // let options = NormalizerOption { lossy: true, ..Default::default() };
+        // let database = self.index.facet_id_string_docids.remap_data_type::<DecodeIgnore>();
+        // for result in database.iter(wtxn)? {
+        //     let (facet_group_key, ()) = result?;
+        //     if let FacetGroupKey { field_id, level: 0, left_bound } = facet_group_key {
+        //         let mut normalized_facet = left_bound.normalize(&options);
+        //         let normalized_truncated_facet: String;
+        //         if normalized_facet.len() > MAX_FACET_VALUE_LENGTH {
+        //             normalized_truncated_facet = normalized_facet
+        //                 .char_indices()
+        //                 .take_while(|(idx, _)| *idx < MAX_FACET_VALUE_LENGTH)
+        //                 .map(|(_, c)| c)
+        //                 .collect();
+        //             normalized_facet = normalized_truncated_facet.into();
+        //         }
+        //         let set = BTreeSet::from_iter(std::iter::once(left_bound));
+        //         let key = (field_id, normalized_facet.as_ref());
+        //         let key = BEU16StrCodec::bytes_encode(&key).map_err(heed::Error::Encoding)?;
+        //         let val = SerdeJson::bytes_encode(&set).map_err(heed::Error::Encoding)?;
+        //         sorter.insert(key, val)?;
+        //     }
+        // }
 
-            if let Some((_, fst_builder)) = current_fst.as_mut() {
-                fst_builder.insert(normalized_facet)?;
-            }
-        }
+        // // In this loop we don't need to take care of merging bitmaps
+        // // as the grenad sorter already merged them for us.
+        // let mut merger_iter = sorter.into_stream_merger_iter()?;
+        // while let Some((key_bytes, btreeset_bytes)) = merger_iter.next()? {
+        //     self.index.facet_id_normalized_string_strings.remap_types::<Bytes, Bytes>().put(
+        //         wtxn,
+        //         key_bytes,
+        //         btreeset_bytes,
+        //     )?;
+        // }
 
-        if let Some((field_id, fst_builder)) = current_fst {
-            let fst = fst_builder.into_set();
-            text_fsts.push((field_id, fst));
-        }
+        // // We compute one FST by string facet
+        // let mut text_fsts = vec![];
+        // let mut current_fst: Option<(u16, fst::SetBuilder<Vec<u8>>)> = None;
+        // let database =
+        //     self.index.facet_id_normalized_string_strings.remap_data_type::<DecodeIgnore>();
+        // for result in database.iter(wtxn)? {
+        //     let ((field_id, normalized_facet), _) = result?;
+        //     current_fst = match current_fst.take() {
+        //         Some((fid, fst_builder)) if fid != field_id => {
+        //             let fst = fst_builder.into_set();
+        //             text_fsts.push((fid, fst));
+        //             Some((field_id, fst::SetBuilder::memory()))
+        //         }
+        //         Some((field_id, fst_builder)) => Some((field_id, fst_builder)),
+        //         None => Some((field_id, fst::SetBuilder::memory())),
+        //     };
 
-        // We write those FSTs in LMDB now
-        for (field_id, fst) in text_fsts {
-            self.index.facet_id_string_fst.put(wtxn, &field_id, &fst)?;
-        }
+        //     if let Some((_, fst_builder)) = current_fst.as_mut() {
+        //         fst_builder.insert(normalized_facet)?;
+        //     }
+        // }
+
+        // if let Some((field_id, fst_builder)) = current_fst {
+        //     let fst = fst_builder.into_set();
+        //     text_fsts.push((field_id, fst));
+        // }
+
+        // // We write those FSTs in LMDB now
+        // for (field_id, fst) in text_fsts {
+        //     self.index.facet_id_string_fst.put(wtxn, &field_id, &fst)?;
+        // }
 
         Ok(())
     }

milli/src/update/index_documents/typed_chunk.rs

@@ -123,6 +123,8 @@ pub(crate) fn write_typed_chunk_into_index(
 ) -> Result<(RoaringBitmap, bool)> {
     puffin::profile_function!(typed_chunk.to_debug_string());
 
+    log::debug!("Received a chunk to process: {}", typed_chunk.to_debug_string());
+
     let mut is_merged_database = false;
     match typed_chunk {
         TypedChunk::Documents(obkv_documents_iter) => {