Only spawn request threads if necessary

Cargo.lock

@@ -494,7 +494,7 @@ checksum = "8c3c1a368f70d6cf7302d78f8f7093da241fb8e8807c05cc9e51a125895a6d5b"
 
 [[package]]
 name = "benchmarks"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "anyhow",
  "bytes",
@@ -639,7 +639,7 @@ dependencies = [
 
 [[package]]
 name = "build-info"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "anyhow",
  "time",
@@ -1539,7 +1539,7 @@ dependencies = [
 
 [[package]]
 name = "dump"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "anyhow",
  "big_s",
@@ -1787,7 +1787,7 @@ dependencies = [
 
 [[package]]
 name = "file-store"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "faux",
  "tempfile",
@@ -1810,7 +1810,7 @@ dependencies = [
 
 [[package]]
 name = "filter-parser"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "insta",
  "nom",
@@ -1830,7 +1830,7 @@ dependencies = [
 
 [[package]]
 name = "flatten-serde-json"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "criterion",
  "serde_json",
@@ -1948,7 +1948,7 @@ dependencies = [
 
 [[package]]
 name = "fuzzers"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "arbitrary",
  "clap",
@@ -2442,7 +2442,7 @@ checksum = "206ca75c9c03ba3d4ace2460e57b189f39f43de612c2f85836e65c929701bb2d"
 
 [[package]]
 name = "index-scheduler"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "anyhow",
  "big_s",
@@ -2638,7 +2638,7 @@ dependencies = [
 
 [[package]]
 name = "json-depth-checker"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "criterion",
  "serde_json",
@@ -3275,7 +3275,7 @@ checksum = "490cc448043f947bae3cbee9c203358d62dbee0db12107a74be5c30ccfd09771"
 
 [[package]]
 name = "meili-snap"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "insta",
  "md5",
@@ -3284,7 +3284,7 @@ dependencies = [
 
 [[package]]
 name = "meilisearch"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "actix-cors",
  "actix-http",
@@ -3377,7 +3377,7 @@ dependencies = [
 
 [[package]]
 name = "meilisearch-auth"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "base64 0.21.7",
  "enum-iterator",
@@ -3396,7 +3396,7 @@ dependencies = [
 
 [[package]]
 name = "meilisearch-types"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "actix-web",
  "anyhow",
@@ -3426,7 +3426,7 @@ dependencies = [
 
 [[package]]
 name = "meilitool"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "anyhow",
  "clap",
@@ -3465,7 +3465,7 @@ dependencies = [
 
 [[package]]
 name = "milli"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "arroy",
  "big_s",
@@ -3906,7 +3906,7 @@ checksum = "e3148f5046208a5d56bcfc03053e3ca6334e51da8dfb19b6cdc8b306fae3283e"
 
 [[package]]
 name = "permissive-json-pointer"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "big_s",
  "serde_json",
@@ -6074,7 +6074,7 @@ dependencies = [
 
 [[package]]
 name = "xtask"
-version = "1.8.1"
+version = "1.8.0"
 dependencies = [
  "anyhow",
  "build-info",

Cargo.toml

@@ -22,7 +22,7 @@ members = [
 ]
 
 [workspace.package]
-version = "1.8.1"
+version = "1.8.0"
 authors = [
     "Quentin de Quelen <quentin@dequelen.me>",
     "Clément Renault <clement@meilisearch.com>",

index-scheduler/src/lib.rs

@@ -567,16 +567,16 @@ impl IndexScheduler {
 
         tracing::debug!("index budget: {budget}B");
         let mut index_count = budget / base_map_size;
-        if index_count < 3 {
+        if index_count < 2 {
             // take a bit less than half than the budget to make sure we can always afford to open an index
             let map_size = (budget * 2) / 5;
             // single index of max budget
             tracing::debug!("1 index of {map_size}B can be opened simultaneously.");
             return IndexBudget { map_size, index_count: 1, task_db_size };
         }
-        // give us some space for additional indexes when the cache is already full
-        // decrement is OK because index_count >= 3.
-        index_count -= 2;
+        // give us some space for an additional index when the cache is already full
+        // decrement is OK because index_count >= 2.
+        index_count -= 1;
         if index_count > max_index_count {
             index_count = max_index_count;
         }
@@ -1834,7 +1834,7 @@ mod tests {
                 task_db_size: 1000 * 1000, // 1 MB, we don't use MiB on purpose.
                 index_base_map_size: 1000 * 1000, // 1 MB, we don't use MiB on purpose.
                 enable_mdb_writemap: false,
-                index_growth_amount: 1000 * 1000 * 1000 * 1000, // 1 TB
+                index_growth_amount: 1000 * 1000, // 1 MB
                 index_count: 5,
                 indexer_config,
                 autobatching_enabled: true,

meilisearch/tests/search/geo.rs

@@ -117,69 +117,3 @@ async fn geo_bounding_box_with_string_and_number() {
         )
         .await;
 }
-
-#[actix_rt::test]
-async fn bug_4640() {
-    // https://github.com/meilisearch/meilisearch/issues/4640
-    let server = Server::new().await;
-    let index = server.index("test");
-
-    let documents = DOCUMENTS.clone();
-    index.add_documents(documents, None).await;
-    index.update_settings_filterable_attributes(json!(["_geo"])).await;
-    let (ret, _code) = index.update_settings_sortable_attributes(json!(["_geo"])).await;
-    index.wait_task(ret.uid()).await;
-
-    // Sort the document with the second one first
-    index
-        .search(
-            json!({
-                "sort": ["_geoPoint(45.4777599, 9.1967508):asc"],
-            }),
-            |response, code| {
-                assert_eq!(code, 200, "{}", response);
-                snapshot!(json_string!(response, { ".processingTimeMs" => "[time]" }), @r###"
-                {
-                  "hits": [
-                    {
-                      "id": 2,
-                      "name": "La Bella Italia",
-                      "address": "456 Elm Street, Townsville",
-                      "type": "Italian",
-                      "rating": 9,
-                      "_geo": {
-                        "lat": "45.4777599",
-                        "lng": "9.1967508"
-                      }
-                    },
-                    {
-                      "id": 1,
-                      "name": "Taco Truck",
-                      "address": "444 Salsa Street, Burritoville",
-                      "type": "Mexican",
-                      "rating": 9,
-                      "_geo": {
-                        "lat": 34.0522,
-                        "lng": -118.2437
-                      },
-                      "_geoDistance": 9714063
-                    },
-                    {
-                      "id": 3,
-                      "name": "Crêpe Truck",
-                      "address": "2 Billig Avenue, Rouenville",
-                      "type": "French",
-                      "rating": 10
-                    }
-                  ],
-                  "query": "",
-                  "processingTimeMs": "[time]",
-                  "limit": 20,
-                  "offset": 0,
-                  "estimatedTotalHits": 3
-                }
-                "###);
-            },
-        )
-        .await;
-}

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

@@ -45,6 +45,7 @@ pub fn extract_fid_docid_facet_values<R: io::Read + io::Seek>(
     obkv_documents: grenad::Reader<R>,
     indexer: GrenadParameters,
     settings_diff: &InnerIndexSettingsDiff,
+    geo_fields_ids: Option<(FieldId, FieldId)>,
 ) -> Result<ExtractedFacetValues> {
     puffin::profile_function!();
 
@@ -126,18 +127,12 @@ pub fn extract_fid_docid_facet_values<R: io::Read + io::Seek>(
                     add_exists.insert(document);
                 }
 
-                let del_geo_support = settings_diff
-                    .old
-                    .geo_fields_ids
-                    .map_or(false, |(lat, lng)| field_id == lat || field_id == lng);
-                let add_geo_support = settings_diff
-                    .new
-                    .geo_fields_ids
-                    .map_or(false, |(lat, lng)| field_id == lat || field_id == lng);
+                let geo_support =
+                    geo_fields_ids.map_or(false, |(lat, lng)| field_id == lat || field_id == lng);
                 let del_filterable_values =
-                    del_value.map(|value| extract_facet_values(&value, del_geo_support));
+                    del_value.map(|value| extract_facet_values(&value, geo_support));
                 let add_filterable_values =
-                    add_value.map(|value| extract_facet_values(&value, add_geo_support));
+                    add_value.map(|value| extract_facet_values(&value, geo_support));
 
                 // Those closures are just here to simplify things a bit.
                 let mut insert_numbers_diff = |del_numbers, add_numbers| {

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

@@ -8,7 +8,6 @@ use super::helpers::{create_writer, writer_into_reader, GrenadParameters};
 use crate::error::GeoError;
 use crate::update::del_add::{DelAdd, KvReaderDelAdd, KvWriterDelAdd};
 use crate::update::index_documents::extract_finite_float_from_value;
-use crate::update::settings::{InnerIndexSettings, InnerIndexSettingsDiff};
 use crate::{FieldId, InternalError, Result};
 
 /// Extracts the geographical coordinates contained in each document under the `_geo` field.
@@ -19,7 +18,7 @@ pub fn extract_geo_points<R: io::Read + io::Seek>(
     obkv_documents: grenad::Reader<R>,
     indexer: GrenadParameters,
     primary_key_id: FieldId,
-    settings_diff: &InnerIndexSettingsDiff,
+    (lat_fid, lng_fid): (FieldId, FieldId),
 ) -> Result<grenad::Reader<BufReader<File>>> {
     puffin::profile_function!();
 
@@ -41,27 +40,47 @@ pub fn extract_geo_points<R: io::Read + io::Seek>(
             serde_json::from_slice(document_id).unwrap()
         };
 
-        // extract old version
-        let del_lat_lng =
-            extract_lat_lng(&obkv, &settings_diff.old, DelAdd::Deletion, document_id)?;
-        // extract new version
-        let add_lat_lng =
-            extract_lat_lng(&obkv, &settings_diff.new, DelAdd::Addition, document_id)?;
+        // first we get the two fields
+        match (obkv.get(lat_fid), obkv.get(lng_fid)) {
+            (Some(lat), Some(lng)) => {
+                let deladd_lat_obkv = KvReaderDelAdd::new(lat);
+                let deladd_lng_obkv = KvReaderDelAdd::new(lng);
 
-        if del_lat_lng != add_lat_lng {
-            let mut obkv = KvWriterDelAdd::memory();
-            if let Some([lat, lng]) = del_lat_lng {
-                #[allow(clippy::drop_non_drop)]
-                let bytes: [u8; 16] = concat_arrays![lat.to_ne_bytes(), lng.to_ne_bytes()];
-                obkv.insert(DelAdd::Deletion, bytes)?;
+                // then we extract the values
+                let del_lat_lng = deladd_lat_obkv
+                    .get(DelAdd::Deletion)
+                    .zip(deladd_lng_obkv.get(DelAdd::Deletion))
+                    .map(|(lat, lng)| extract_lat_lng(lat, lng, document_id))
+                    .transpose()?;
+                let add_lat_lng = deladd_lat_obkv
+                    .get(DelAdd::Addition)
+                    .zip(deladd_lng_obkv.get(DelAdd::Addition))
+                    .map(|(lat, lng)| extract_lat_lng(lat, lng, document_id))
+                    .transpose()?;
+
+                if del_lat_lng != add_lat_lng {
+                    let mut obkv = KvWriterDelAdd::memory();
+                    if let Some([lat, lng]) = del_lat_lng {
+                        #[allow(clippy::drop_non_drop)]
+                        let bytes: [u8; 16] = concat_arrays![lat.to_ne_bytes(), lng.to_ne_bytes()];
+                        obkv.insert(DelAdd::Deletion, bytes)?;
+                    }
+                    if let Some([lat, lng]) = add_lat_lng {
+                        #[allow(clippy::drop_non_drop)]
+                        let bytes: [u8; 16] = concat_arrays![lat.to_ne_bytes(), lng.to_ne_bytes()];
+                        obkv.insert(DelAdd::Addition, bytes)?;
+                    }
+                    let bytes = obkv.into_inner()?;
+                    writer.insert(docid_bytes, bytes)?;
+                }
             }
-            if let Some([lat, lng]) = add_lat_lng {
-                #[allow(clippy::drop_non_drop)]
-                let bytes: [u8; 16] = concat_arrays![lat.to_ne_bytes(), lng.to_ne_bytes()];
-                obkv.insert(DelAdd::Addition, bytes)?;
+            (None, Some(_)) => {
+                return Err(GeoError::MissingLatitude { document_id: document_id() }.into())
             }
-            let bytes = obkv.into_inner()?;
-            writer.insert(docid_bytes, bytes)?;
+            (Some(_), None) => {
+                return Err(GeoError::MissingLongitude { document_id: document_id() }.into())
+            }
+            (None, None) => (),
         }
     }
 
@@ -69,37 +88,16 @@ pub fn extract_geo_points<R: io::Read + io::Seek>(
 }
 
 /// Extract the finite floats lat and lng from two bytes slices.
-fn extract_lat_lng(
-    document: &obkv::KvReader<FieldId>,
-    settings: &InnerIndexSettings,
-    deladd: DelAdd,
-    document_id: impl Fn() -> Value,
-) -> Result<Option<[f64; 2]>> {
-    match settings.geo_fields_ids {
-        Some((lat_fid, lng_fid)) => {
-            let lat = document.get(lat_fid).map(KvReaderDelAdd::new).and_then(|r| r.get(deladd));
-            let lng = document.get(lng_fid).map(KvReaderDelAdd::new).and_then(|r| r.get(deladd));
-            let (lat, lng) = match (lat, lng) {
-                (Some(lat), Some(lng)) => (lat, lng),
-                (Some(_), None) => {
-                    return Err(GeoError::MissingLatitude { document_id: document_id() }.into())
-                }
-                (None, Some(_)) => {
-                    return Err(GeoError::MissingLongitude { document_id: document_id() }.into())
-                }
-                (None, None) => return Ok(None),
-            };
-            let lat = extract_finite_float_from_value(
-                serde_json::from_slice(lat).map_err(InternalError::SerdeJson)?,
-            )
-            .map_err(|lat| GeoError::BadLatitude { document_id: document_id(), value: lat })?;
+fn extract_lat_lng(lat: &[u8], lng: &[u8], document_id: impl Fn() -> Value) -> Result<[f64; 2]> {
+    let lat = extract_finite_float_from_value(
+        serde_json::from_slice(lat).map_err(InternalError::SerdeJson)?,
+    )
+    .map_err(|lat| GeoError::BadLatitude { document_id: document_id(), value: lat })?;
 
-            let lng = extract_finite_float_from_value(
-                serde_json::from_slice(lng).map_err(InternalError::SerdeJson)?,
-            )
-            .map_err(|lng| GeoError::BadLongitude { document_id: document_id(), value: lng })?;
-            Ok(Some([lat, lng]))
-        }
-        None => Ok(None),
-    }
+    let lng = extract_finite_float_from_value(
+        serde_json::from_slice(lng).map_err(InternalError::SerdeJson)?,
+    )
+    .map_err(|lng| GeoError::BadLongitude { document_id: document_id(), value: lng })?;
+
+    Ok([lat, lng])
 }

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

@@ -43,6 +43,7 @@ pub(crate) fn data_from_obkv_documents(
     indexer: GrenadParameters,
     lmdb_writer_sx: Sender<Result<TypedChunk>>,
     primary_key_id: FieldId,
+    geo_fields_ids: Option<(FieldId, FieldId)>,
     settings_diff: Arc<InnerIndexSettingsDiff>,
     max_positions_per_attributes: Option<u32>,
 ) -> Result<()> {
@@ -71,6 +72,7 @@ pub(crate) fn data_from_obkv_documents(
                         indexer,
                         lmdb_writer_sx.clone(),
                         primary_key_id,
+                        geo_fields_ids,
                         settings_diff.clone(),
                         max_positions_per_attributes,
                     )
@@ -227,12 +229,15 @@ fn send_original_documents_data(
     let documents_chunk_cloned = original_documents_chunk.clone();
     let lmdb_writer_sx_cloned = lmdb_writer_sx.clone();
 
-    let request_threads = ThreadPoolNoAbortBuilder::new()
-        .num_threads(crate::vector::REQUEST_PARALLELISM)
-        .thread_name(|index| format!("embedding-request-{index}"))
-        .build()?;
+    let new_embedding_configs = settings_diff.new.embedding_configs.clone();
 
-    if settings_diff.reindex_vectors() || !settings_diff.settings_update_only() {
+    if (settings_diff.reindex_vectors() || !settings_diff.settings_update_only())
+        && new_embedding_configs.get_default().is_some()
+    {
+        let request_threads = ThreadPoolNoAbortBuilder::new()
+            .num_threads(crate::vector::REQUEST_PARALLELISM)
+            .thread_name(|index| format!("embedding-request-{index}"))
+            .build()?;
         let settings_diff = settings_diff.clone();
         rayon::spawn(move || {
             for (name, (embedder, prompt)) in settings_diff.new.embedding_configs.clone() {
@@ -298,6 +303,7 @@ fn send_and_extract_flattened_documents_data(
     indexer: GrenadParameters,
     lmdb_writer_sx: Sender<Result<TypedChunk>>,
     primary_key_id: FieldId,
+    geo_fields_ids: Option<(FieldId, FieldId)>,
     settings_diff: Arc<InnerIndexSettingsDiff>,
     max_positions_per_attributes: Option<u32>,
 ) -> Result<(
@@ -307,13 +313,12 @@ fn send_and_extract_flattened_documents_data(
     let flattened_documents_chunk =
         flattened_documents_chunk.and_then(|c| unsafe { as_cloneable_grenad(&c) })?;
 
-    if settings_diff.run_geo_indexing() {
+    if let Some(geo_fields_ids) = geo_fields_ids {
         let documents_chunk_cloned = flattened_documents_chunk.clone();
         let lmdb_writer_sx_cloned = lmdb_writer_sx.clone();
-        let settings_diff = settings_diff.clone();
         rayon::spawn(move || {
             let result =
-                extract_geo_points(documents_chunk_cloned, indexer, primary_key_id, &settings_diff);
+                extract_geo_points(documents_chunk_cloned, indexer, primary_key_id, geo_fields_ids);
             let _ = match result {
                 Ok(geo_points) => lmdb_writer_sx_cloned.send(Ok(TypedChunk::GeoPoints(geo_points))),
                 Err(error) => lmdb_writer_sx_cloned.send(Err(error)),
@@ -352,6 +357,7 @@ fn send_and_extract_flattened_documents_data(
                     flattened_documents_chunk.clone(),
                     indexer,
                     &settings_diff,
+                    geo_fields_ids,
                 )?;
 
                 // send fid_docid_facet_numbers_chunk to DB writer

milli/src/update/index_documents/mod.rs

@@ -324,6 +324,28 @@ where
         // get the primary key field id
         let primary_key_id = settings_diff.new.fields_ids_map.id(&primary_key).unwrap();
 
+        // get the fid of the `_geo.lat` and `_geo.lng` fields.
+        let mut field_id_map = self.index.fields_ids_map(self.wtxn)?;
+
+        // self.index.fields_ids_map($a)? ==>> field_id_map
+        let geo_fields_ids = match field_id_map.id("_geo") {
+            Some(gfid) => {
+                let is_sortable = self.index.sortable_fields_ids(self.wtxn)?.contains(&gfid);
+                let is_filterable = self.index.filterable_fields_ids(self.wtxn)?.contains(&gfid);
+                // if `_geo` is faceted then we get the `lat` and `lng`
+                if is_sortable || is_filterable {
+                    let field_ids = field_id_map
+                        .insert("_geo.lat")
+                        .zip(field_id_map.insert("_geo.lng"))
+                        .ok_or(UserError::AttributeLimitReached)?;
+                    Some(field_ids)
+                } else {
+                    None
+                }
+            }
+            None => None,
+        };
+
         let pool_params = GrenadParameters {
             chunk_compression_type: self.indexer_config.chunk_compression_type,
             chunk_compression_level: self.indexer_config.chunk_compression_level,
@@ -390,6 +412,7 @@ where
                         pool_params,
                         lmdb_writer_sx.clone(),
                         primary_key_id,
+                        geo_fields_ids,
                         settings_diff.clone(),
                         max_positions_per_attributes,
                     )

milli/src/update/settings.rs

@@ -1161,11 +1161,6 @@ impl InnerIndexSettingsDiff {
     pub fn settings_update_only(&self) -> bool {
         self.settings_update_only
     }
-
-    pub fn run_geo_indexing(&self) -> bool {
-        self.old.geo_fields_ids != self.new.geo_fields_ids
-            || (!self.settings_update_only && self.new.geo_fields_ids.is_some())
-    }
 }
 
 #[derive(Clone)]
@@ -1182,7 +1177,6 @@ pub(crate) struct InnerIndexSettings {
     pub proximity_precision: ProximityPrecision,
     pub embedding_configs: EmbeddingConfigs,
     pub existing_fields: HashSet<String>,
-    pub geo_fields_ids: Option<(FieldId, FieldId)>,
 }
 
 impl InnerIndexSettings {
@@ -1191,7 +1185,7 @@ impl InnerIndexSettings {
         let stop_words = stop_words.map(|sw| sw.map_data(Vec::from).unwrap());
         let allowed_separators = index.allowed_separators(rtxn)?;
         let dictionary = index.dictionary(rtxn)?;
-        let mut fields_ids_map = index.fields_ids_map(rtxn)?;
+        let fields_ids_map = index.fields_ids_map(rtxn)?;
         let user_defined_searchable_fields = index.user_defined_searchable_fields(rtxn)?;
         let user_defined_searchable_fields =
             user_defined_searchable_fields.map(|sf| sf.into_iter().map(String::from).collect());
@@ -1206,24 +1200,6 @@ impl InnerIndexSettings {
             .into_iter()
             .filter_map(|(field, count)| (count != 0).then_some(field))
             .collect();
-        // index.fields_ids_map($a)? ==>> fields_ids_map
-        let geo_fields_ids = match fields_ids_map.id("_geo") {
-            Some(gfid) => {
-                let is_sortable = index.sortable_fields_ids(rtxn)?.contains(&gfid);
-                let is_filterable = index.filterable_fields_ids(rtxn)?.contains(&gfid);
-                // if `_geo` is faceted then we get the `lat` and `lng`
-                if is_sortable || is_filterable {
-                    let field_ids = fields_ids_map
-                        .insert("_geo.lat")
-                        .zip(fields_ids_map.insert("_geo.lng"))
-                        .ok_or(UserError::AttributeLimitReached)?;
-                    Some(field_ids)
-                } else {
-                    None
-                }
-            }
-            None => None,
-        };
 
         Ok(Self {
             stop_words,
@@ -1238,7 +1214,6 @@ impl InnerIndexSettings {
             proximity_precision,
             embedding_configs,
             existing_fields,
-            geo_fields_ids,
         })
     }