Merge pull request #5494 from meilisearch/deactivate-numbers-in-typos

Deactivate numbers in typos

crates/milli/src/disabled_typos_terms.rs

@@ -0,0 +1,50 @@
+use heed::{
+    types::{SerdeJson, Str},
+    RoTxn, RwTxn,
+};
+use serde::{Deserialize, Serialize};
+
+use crate::{index::main_key, Index};
+
+#[derive(Debug, Clone, Copy, Serialize, Deserialize, PartialEq, Eq, Default)]
+#[serde(rename_all = "camelCase")]
+pub struct DisabledTyposTerms {
+    pub disable_on_numbers: bool,
+}
+
+impl Index {
+    pub fn disabled_typos_terms(&self, txn: &RoTxn<'_>) -> heed::Result<DisabledTyposTerms> {
+        self.main
+            .remap_types::<Str, SerdeJson<DisabledTyposTerms>>()
+            .get(txn, main_key::DISABLED_TYPOS_TERMS)
+            .map(|option| option.unwrap_or_default())
+    }
+
+    pub(crate) fn put_disabled_typos_terms(
+        &self,
+        txn: &mut RwTxn<'_>,
+        disabled_typos_terms: &DisabledTyposTerms,
+    ) -> heed::Result<()> {
+        self.main.remap_types::<Str, SerdeJson<DisabledTyposTerms>>().put(
+            txn,
+            main_key::DISABLED_TYPOS_TERMS,
+            disabled_typos_terms,
+        )?;
+
+        Ok(())
+    }
+
+    pub(crate) fn delete_disabled_typos_terms(&self, txn: &mut RwTxn<'_>) -> heed::Result<()> {
+        self.main
+            .remap_types::<Str, SerdeJson<DisabledTyposTerms>>()
+            .delete(txn, main_key::DISABLED_TYPOS_TERMS)?;
+        Ok(())
+    }
+}
+
+impl DisabledTyposTerms {
+    pub fn is_exact(&self, word: &str) -> bool {
+        // If disable_on_numbers is true, we disable the word if it contains only numbers or punctuation
+        self.disable_on_numbers && word.chars().all(|c| c.is_numeric() || c.is_ascii_punctuation())
+    }
+}

crates/milli/src/index.rs

@@ -78,6 +78,7 @@ pub mod main_key {
     pub const FACET_SEARCH: &str = "facet_search";
     pub const PREFIX_SEARCH: &str = "prefix_search";
     pub const DOCUMENTS_STATS: &str = "documents_stats";
+    pub const DISABLED_TYPOS_TERMS: &str = "disabled_typos_terms";
 }
 
 pub mod db_name {

crates/milli/src/lib.rs

@@ -12,6 +12,7 @@ mod asc_desc;
 mod attribute_patterns;
 mod criterion;
 pub mod database_stats;
+pub mod disabled_typos_terms;
 mod error;
 mod external_documents_ids;
 pub mod facet;

crates/milli/src/update/index_documents/extract/extract_word_docids.rs

@@ -127,7 +127,8 @@ pub fn extract_word_docids<R: io::Read + io::Seek>(
         // merge all deletions
         let obkv = KvReaderDelAdd::from_slice(value);
         if let Some(value) = obkv.get(DelAdd::Deletion) {
-            let delete_from_exact = settings_diff.old.exact_attributes.contains(&fid);
+            let delete_from_exact = settings_diff.old.exact_attributes.contains(&fid)
+                || settings_diff.old.disabled_typos_terms.is_exact(w);
             buffer.clear();
             let mut obkv = KvWriterDelAdd::new(&mut buffer);
             obkv.insert(DelAdd::Deletion, value)?;
@@ -139,7 +140,8 @@ pub fn extract_word_docids<R: io::Read + io::Seek>(
         }
         // merge all additions
         if let Some(value) = obkv.get(DelAdd::Addition) {
-            let add_in_exact = settings_diff.new.exact_attributes.contains(&fid);
+            let add_in_exact = settings_diff.new.exact_attributes.contains(&fid)
+                || settings_diff.new.disabled_typos_terms.is_exact(w);
             buffer.clear();
             let mut obkv = KvWriterDelAdd::new(&mut buffer);
             obkv.insert(DelAdd::Addition, value)?;

crates/milli/src/update/index_documents/typed_chunk.rs

@@ -273,14 +273,11 @@ pub(crate) fn write_typed_chunk_into_index(
                     unreachable!();
                 };
                 let clonable_word_docids = unsafe { as_cloneable_grenad(&word_docids_reader) }?;
-                let clonable_exact_word_docids =
-                    unsafe { as_cloneable_grenad(&exact_word_docids_reader) }?;
 
                 word_docids_builder.push(word_docids_reader.into_cursor()?);
                 exact_word_docids_builder.push(exact_word_docids_reader.into_cursor()?);
                 word_fid_docids_builder.push(word_fid_docids_reader.into_cursor()?);
                 fst_merger_builder.push(clonable_word_docids.into_cursor()?);
-                fst_merger_builder.push(clonable_exact_word_docids.into_cursor()?);
             }
 
             let word_docids_merger = word_docids_builder.build();

crates/milli/src/update/new/extract/searchable/extract_word_docids.rs

@@ -319,8 +319,11 @@ impl WordDocidsExtractors {
         let doc_alloc = &context.doc_alloc;
 
         let exact_attributes = index.exact_attributes(rtxn)?;
-        let is_exact_attribute =
-            |fname: &str| exact_attributes.iter().any(|attr| contained_in(fname, attr));
+        let disabled_typos_terms = index.disabled_typos_terms(rtxn)?;
+        let is_exact = |fname: &str, word: &str| {
+            exact_attributes.iter().any(|attr| contained_in(fname, attr))
+                || disabled_typos_terms.is_exact(word)
+        };
         match document_change {
             DocumentChange::Deletion(inner) => {
                 let mut token_fn = |fname: &str, fid, pos, word: &str| {
@@ -328,7 +331,7 @@ impl WordDocidsExtractors {
                         fid,
                         pos,
                         word,
-                        is_exact_attribute(fname),
+                        is_exact(fname, word),
                         inner.docid(),
                         doc_alloc,
                     )
@@ -356,7 +359,7 @@ impl WordDocidsExtractors {
                         fid,
                         pos,
                         word,
-                        is_exact_attribute(fname),
+                        is_exact(fname, word),
                         inner.docid(),
                         doc_alloc,
                     )
@@ -372,7 +375,7 @@ impl WordDocidsExtractors {
                         fid,
                         pos,
                         word,
-                        is_exact_attribute(fname),
+                        is_exact(fname, word),
                         inner.docid(),
                         doc_alloc,
                     )
@@ -389,7 +392,7 @@ impl WordDocidsExtractors {
                         fid,
                         pos,
                         word,
-                        is_exact_attribute(fname),
+                        is_exact(fname, word),
                         inner.docid(),
                         doc_alloc,
                     )

crates/milli/src/update/new/indexer/mod.rs

@@ -9,6 +9,7 @@ pub use document_operation::{DocumentOperation, PayloadStats};
 use hashbrown::HashMap;
 use heed::RwTxn;
 pub use partial_dump::PartialDump;
+pub use post_processing::recompute_word_fst_from_word_docids_database;
 pub use update_by_function::UpdateByFunction;
 pub use write::ChannelCongestion;
 use write::{build_vectors, update_index, write_to_db};

crates/milli/src/update/new/indexer/post_processing.rs

@@ -131,6 +131,20 @@ fn compute_word_fst(
     }
 }
 
+pub fn recompute_word_fst_from_word_docids_database(index: &Index, wtxn: &mut RwTxn) -> Result<()> {
+    let fst = fst::Set::default().map_data(std::borrow::Cow::Owned)?;
+    let mut word_fst_builder = WordFstBuilder::new(&fst)?;
+    let words = index.word_docids.iter(wtxn)?.remap_data_type::<DecodeIgnore>();
+    for res in words {
+        let (word, _) = res?;
+        word_fst_builder.register_word(DelAdd::Addition, word.as_ref())?;
+    }
+    let (word_fst_mmap, _) = word_fst_builder.build(index, wtxn)?;
+    index.main.remap_types::<Str, Bytes>().put(wtxn, WORDS_FST_KEY, &word_fst_mmap)?;
+
+    Ok(())
+}
+
 #[tracing::instrument(level = "trace", skip_all, target = "indexing::facet_search")]
 fn compute_facet_search_database(
     index: &Index,

crates/milli/src/update/settings.rs

@@ -17,6 +17,7 @@ use super::IndexerConfig;
 use crate::attribute_patterns::PatternMatch;
 use crate::constants::RESERVED_GEO_FIELD_NAME;
 use crate::criterion::Criterion;
+use crate::disabled_typos_terms::DisabledTyposTerms;
 use crate::error::UserError;
 use crate::fields_ids_map::metadata::{FieldIdMapWithMetadata, MetadataBuilder};
 use crate::filterable_attributes_rules::match_faceted_field;
@@ -169,6 +170,7 @@ pub struct Settings<'a, 't, 'i> {
     synonyms: Setting<BTreeMap<String, Vec<String>>>,
     primary_key: Setting<String>,
     authorize_typos: Setting<bool>,
+    disable_on_numbers: Setting<bool>,
     min_word_len_two_typos: Setting<u8>,
     min_word_len_one_typo: Setting<u8>,
     exact_words: Setting<BTreeSet<String>>,
@@ -207,6 +209,7 @@ impl<'a, 't, 'i> Settings<'a, 't, 'i> {
             synonyms: Setting::NotSet,
             primary_key: Setting::NotSet,
             authorize_typos: Setting::NotSet,
+            disable_on_numbers: Setting::NotSet,
             exact_words: Setting::NotSet,
             min_word_len_two_typos: Setting::NotSet,
             min_word_len_one_typo: Setting::NotSet,
@@ -354,6 +357,14 @@ impl<'a, 't, 'i> Settings<'a, 't, 'i> {
         self.min_word_len_one_typo = Setting::Reset;
     }
 
+    pub fn set_disable_on_numbers(&mut self, disable_on_numbers: bool) {
+        self.disable_on_numbers = Setting::Set(disable_on_numbers);
+    }
+
+    pub fn reset_disable_on_numbers(&mut self) {
+        self.disable_on_numbers = Setting::Reset;
+    }
+
     pub fn set_exact_words(&mut self, words: BTreeSet<String>) {
         self.exact_words = Setting::Set(words);
     }
@@ -866,6 +877,24 @@ impl<'a, 't, 'i> Settings<'a, 't, 'i> {
         Ok(())
     }
 
+    fn update_disabled_typos_terms(&mut self) -> Result<()> {
+        let mut disabled_typos_terms = self.index.disabled_typos_terms(self.wtxn)?;
+        match self.disable_on_numbers {
+            Setting::Set(disable_on_numbers) => {
+                disabled_typos_terms.disable_on_numbers = disable_on_numbers;
+            }
+            Setting::Reset => {
+                self.index.delete_disabled_typos_terms(self.wtxn)?;
+                disabled_typos_terms.disable_on_numbers =
+                    DisabledTyposTerms::default().disable_on_numbers;
+            }
+            Setting::NotSet => (),
+        }
+
+        self.index.put_disabled_typos_terms(self.wtxn, &disabled_typos_terms)?;
+        Ok(())
+    }
+
     fn update_exact_words(&mut self) -> Result<()> {
         match self.exact_words {
             Setting::Set(ref mut words) => {
@@ -1246,6 +1275,7 @@ impl<'a, 't, 'i> Settings<'a, 't, 'i> {
         self.update_prefix_search()?;
         self.update_facet_search()?;
         self.update_localized_attributes_rules()?;
+        self.update_disabled_typos_terms()?;
 
         let embedding_config_updates = self.update_embedding_configs()?;
 
@@ -1327,6 +1357,7 @@ impl InnerIndexSettingsDiff {
                 || old_settings.prefix_search != new_settings.prefix_search
                 || old_settings.localized_attributes_rules
                     != new_settings.localized_attributes_rules
+                || old_settings.disabled_typos_terms != new_settings.disabled_typos_terms
         };
 
         let cache_exact_attributes = old_settings.exact_attributes != new_settings.exact_attributes;
@@ -1526,6 +1557,7 @@ pub(crate) struct InnerIndexSettings {
     pub user_defined_searchable_attributes: Option<Vec<String>>,
     pub sortable_fields: HashSet<String>,
     pub exact_attributes: HashSet<FieldId>,
+    pub disabled_typos_terms: DisabledTyposTerms,
     pub proximity_precision: ProximityPrecision,
     pub embedding_configs: EmbeddingConfigs,
     pub geo_fields_ids: Option<(FieldId, FieldId)>,
@@ -1574,7 +1606,7 @@ impl InnerIndexSettings {
             .map(|fields| fields.into_iter().map(|f| f.to_string()).collect());
         let builder = MetadataBuilder::from_index(index, rtxn)?;
         let fields_ids_map = FieldIdMapWithMetadata::new(fields_ids_map, builder);
-
+        let disabled_typos_terms = index.disabled_typos_terms(rtxn)?;
         Ok(Self {
             stop_words,
             allowed_separators,
@@ -1592,6 +1624,7 @@ impl InnerIndexSettings {
             geo_fields_ids,
             prefix_search,
             facet_search,
+            disabled_typos_terms,
         })
     }
 

crates/milli/src/update/test_settings.rs

@@ -896,6 +896,7 @@ fn test_correct_settings_init() {
                 localized_attributes_rules,
                 prefix_search,
                 facet_search,
+                disable_on_numbers,
             } = settings;
             assert!(matches!(searchable_fields, Setting::NotSet));
             assert!(matches!(displayed_fields, Setting::NotSet));
@@ -923,6 +924,7 @@ fn test_correct_settings_init() {
             assert!(matches!(localized_attributes_rules, Setting::NotSet));
             assert!(matches!(prefix_search, Setting::NotSet));
             assert!(matches!(facet_search, Setting::NotSet));
+            assert!(matches!(disable_on_numbers, Setting::NotSet));
         })
         .unwrap();
 }

crates/milli/src/update/upgrade/mod.rs

@@ -1,11 +1,12 @@
 mod v1_12;
 mod v1_13;
 mod v1_14;
-
+mod v1_15;
 use heed::RwTxn;
 use v1_12::{V1_12_3_To_V1_13_0, V1_12_To_V1_12_3};
 use v1_13::{V1_13_0_To_V1_13_1, V1_13_1_To_Latest_V1_13};
 use v1_14::Latest_V1_13_To_Latest_V1_14;
+use v1_15::Latest_V1_14_To_Latest_V1_15;
 
 use crate::progress::{Progress, VariableNameStep};
 use crate::{Index, InternalError, Result};
@@ -36,6 +37,7 @@ pub fn upgrade(
         &V1_13_0_To_V1_13_1 {},
         &V1_13_1_To_Latest_V1_13 {},
         &Latest_V1_13_To_Latest_V1_14 {},
+        &Latest_V1_14_To_Latest_V1_15 {},
     ];
 
     let start = match from {
@@ -43,8 +45,9 @@ pub fn upgrade(
         (1, 12, 3..) => 1,
         (1, 13, 0) => 2,
         (1, 13, _) => 4,
+        (1, 14, _) => 5,
         // We must handle the current version in the match because in case of a failure some index may have been upgraded but not other.
-        (1, 14, _) => 4,
+        (1, 15, _) => 5,
         (major, minor, patch) => {
             return Err(InternalError::CannotUpgradeToVersion(major, minor, patch).into())
         }

crates/milli/src/update/upgrade/v1_15.rs

@@ -0,0 +1,35 @@
+use heed::RwTxn;
+
+use super::UpgradeIndex;
+use crate::progress::Progress;
+use crate::update::new::indexer::recompute_word_fst_from_word_docids_database;
+use crate::{make_enum_progress, Index, Result};
+
+#[allow(non_camel_case_types)]
+pub(super) struct Latest_V1_14_To_Latest_V1_15();
+
+impl UpgradeIndex for Latest_V1_14_To_Latest_V1_15 {
+    fn upgrade(
+        &self,
+        wtxn: &mut RwTxn,
+        index: &Index,
+        _original: (u32, u32, u32),
+        progress: Progress,
+    ) -> Result<bool> {
+        // Recompute the word FST from the word docids database.
+        make_enum_progress! {
+            enum TypoTolerance {
+                RecomputeWordFst,
+            }
+        };
+
+        progress.update_progress(TypoTolerance::RecomputeWordFst);
+        recompute_word_fst_from_word_docids_database(index, wtxn)?;
+
+        Ok(false)
+    }
+
+    fn target_version(&self) -> (u32, u32, u32) {
+        (1, 15, 0)
+    }
+}