Remove the FacetValueStringCodec

milli/src/heed_codec/facet/facet_level_value_u32_codec.rs

@@ -2,7 +2,7 @@ use std::borrow::Cow;
 use std::convert::TryInto;
 use std::num::NonZeroU8;
 
-use crate::FieldId;
+use crate::{try_split_array_at, FieldId};
 
 /// A codec that stores the field id, level 1 and higher and the groups ids.
 ///
@@ -13,12 +13,13 @@ impl<'a> heed::BytesDecode<'a> for FacetLevelValueU32Codec {
     type DItem = (FieldId, NonZeroU8, u32, u32);
 
     fn bytes_decode(bytes: &'a [u8]) -> Option<Self::DItem> {
-        let (field_id, bytes) = bytes.split_first()?;
+        let (field_id_bytes, bytes) = try_split_array_at(bytes)?;
+        let field_id = u16::from_be_bytes(field_id_bytes);
         let (level, bytes) = bytes.split_first()?;
         let level = NonZeroU8::new(*level)?;
-        let left = bytes[16..20].try_into().ok().map(u32::from_be_bytes)?;
-        let right = bytes[20..].try_into().ok().map(u32::from_be_bytes)?;
-        Some((*field_id, level, left, right))
+        let left = bytes[8..12].try_into().ok().map(u32::from_be_bytes)?;
+        let right = bytes[12..].try_into().ok().map(u32::from_be_bytes)?;
+        Some((field_id, level, left, right))
     }
 }
 
@@ -42,8 +43,8 @@ impl heed::BytesEncode<'_> for FacetLevelValueU32Codec {
         let bytes = right.to_be_bytes();
         buffer[12..].copy_from_slice(&bytes[..]);
 
-        let mut bytes = Vec::with_capacity(buffer.len() + 2);
-        bytes.push(*field_id);
+        let mut bytes = Vec::with_capacity(buffer.len() + 2 + 1);
+        bytes.extend_from_slice(&field_id.to_be_bytes());
         bytes.push(level.get());
         bytes.extend_from_slice(&buffer);
 

milli/src/heed_codec/facet/facet_string_level_zero_codec.rs

@@ -1,7 +1,7 @@
 use std::borrow::Cow;
 use std::str;
 
-use crate::FieldId;
+use crate::{try_split_array_at, FieldId};
 
 /// A codec that stores the field id, level 0, and facet string.
 ///
@@ -16,7 +16,7 @@ pub struct FacetStringLevelZeroCodec;
 impl FacetStringLevelZeroCodec {
     pub fn serialize_into(field_id: FieldId, value: &str, out: &mut Vec<u8>) {
         out.reserve(value.len() + 2);
-        out.push(field_id);
+        out.extend_from_slice(&field_id.to_be_bytes());
         out.push(0); // the level zero (for LMDB ordering only)
         out.extend_from_slice(value.as_bytes());
     }
@@ -26,7 +26,8 @@ impl<'a> heed::BytesDecode<'a> for FacetStringLevelZeroCodec {
     type DItem = (FieldId, &'a str);
 
     fn bytes_decode(bytes: &'a [u8]) -> Option<Self::DItem> {
-        let (field_id, bytes) = bytes.split_first()?;
+        let (field_id_bytes, bytes) = try_split_array_at(bytes)?;
+        let field_id = u16::from_be_bytes(field_id_bytes);
         let (level, bytes) = bytes.split_first()?;
 
         if *level != 0 {
@@ -34,7 +35,7 @@ impl<'a> heed::BytesDecode<'a> for FacetStringLevelZeroCodec {
         }
 
         let value = str::from_utf8(bytes).ok()?;
-        Some((*field_id, value))
+        Some((field_id, value))
     }
 }
 

milli/src/heed_codec/facet/mod.rs

@@ -2,7 +2,6 @@ mod facet_level_value_f64_codec;
 mod facet_level_value_u32_codec;
 mod facet_string_level_zero_codec;
 mod facet_string_zero_bounds_value_codec;
-mod facet_value_string_codec;
 mod field_doc_id_facet_f64_codec;
 mod field_doc_id_facet_string_codec;
 
@@ -10,6 +9,5 @@ pub use self::facet_level_value_f64_codec::FacetLevelValueF64Codec;
 pub use self::facet_level_value_u32_codec::FacetLevelValueU32Codec;
 pub use self::facet_string_level_zero_codec::FacetStringLevelZeroCodec;
 pub use self::facet_string_zero_bounds_value_codec::FacetStringZeroBoundsValueCodec;
-pub use self::facet_value_string_codec::FacetValueStringCodec;
 pub use self::field_doc_id_facet_f64_codec::FieldDocIdFacetF64Codec;
 pub use self::field_doc_id_facet_string_codec::FieldDocIdFacetStringCodec;

milli/src/index.rs

@@ -11,7 +11,7 @@ use roaring::RoaringBitmap;
 use crate::error::{FieldIdMapMissingEntry, InternalError, UserError};
 use crate::fields_ids_map::FieldsIdsMap;
 use crate::heed_codec::facet::{
-    FacetLevelValueF64Codec, FacetValueStringCodec, FieldDocIdFacetF64Codec,
+    FacetLevelValueF64Codec, FacetStringLevelZeroCodec, FieldDocIdFacetF64Codec,
     FieldDocIdFacetStringCodec,
 };
 use crate::{
@@ -91,7 +91,7 @@ pub struct Index {
     /// Maps the facet field id, level and the number with the docids that corresponds to it.
     pub facet_id_f64_docids: Database<FacetLevelValueF64Codec, CboRoaringBitmapCodec>,
     /// Maps the facet field id and the string with the docids that corresponds to it.
-    pub facet_id_string_docids: Database<FacetValueStringCodec, CboRoaringBitmapCodec>,
+    pub facet_id_string_docids: Database<FacetStringLevelZeroCodec, CboRoaringBitmapCodec>,
 
     /// Maps the document id, the facet field id and the numbers.
     pub field_id_docid_facet_f64s: Database<FieldDocIdFacetF64Codec, Unit>,

milli/src/search/facet/facet_distribution.rs

@@ -1,6 +1,6 @@
 use std::collections::{BTreeMap, HashSet};
 use std::ops::Bound::Unbounded;
-use std::{cmp, fmt};
+use std::{cmp, fmt, mem};
 
 use heed::types::{ByteSlice, Unit};
 use heed::{BytesDecode, Database};
@@ -8,7 +8,7 @@ use roaring::RoaringBitmap;
 
 use crate::error::{FieldIdMapMissingEntry, UserError};
 use crate::facet::FacetType;
-use crate::heed_codec::facet::FacetValueStringCodec;
+use crate::heed_codec::facet::FacetStringLevelZeroCodec;
 use crate::search::facet::{FacetNumberIter, FacetNumberRange};
 use crate::{DocumentId, FieldId, Index, Result};
 
@@ -81,7 +81,7 @@ impl<'a> FacetDistribution<'a> {
             let mut key_buffer: Vec<_> = field_id.to_be_bytes().iter().copied().collect();
 
             for docid in candidates.into_iter().take(CANDIDATES_THRESHOLD as usize) {
-                key_buffer.truncate(1);
+                key_buffer.truncate(mem::size_of::<FieldId>());
                 key_buffer.extend_from_slice(&docid.to_be_bytes());
                 let iter = db
                     .remap_key_type::<ByteSlice>()
@@ -158,7 +158,7 @@ impl<'a> FacetDistribution<'a> {
             .facet_id_string_docids
             .remap_key_type::<ByteSlice>()
             .prefix_iter(self.rtxn, &field_id.to_be_bytes())?
-            .remap_key_type::<FacetValueStringCodec>();
+            .remap_key_type::<FacetStringLevelZeroCodec>();
 
         for result in iter {
             let ((_, value), docids) = result?;

milli/src/search/facet/facet_string.rs

@@ -130,7 +130,7 @@ use std::ops::Bound;
 use std::ops::Bound::{Excluded, Included, Unbounded};
 
 use either::{Either, Left, Right};
-use heed::types::{ByteSlice, DecodeIgnore, Str};
+use heed::types::{ByteSlice, DecodeIgnore};
 use heed::{Database, LazyDecode, RoRange};
 use roaring::RoaringBitmap;
 
@@ -298,10 +298,10 @@ impl<'t> FacetStringIter<'t> {
     ) -> heed::Result<Option<u8>> {
         Ok(db
             .remap_types::<ByteSlice, DecodeIgnore>()
-            .prefix_iter(rtxn, &[fid][..])? // the field id is the first bit
+            .prefix_iter(rtxn, &fid.to_be_bytes())? // the field id is the first two bits
             .last()
             .transpose()?
-            .map(|(key_bytes, _)| key_bytes[1])) // the level is the second bit
+            .map(|(key_bytes, _)| key_bytes[2])) // the level is the third bit
     }
 }
 

milli/src/search/facet/filter_condition.rs

@@ -17,7 +17,7 @@ use self::Operator::*;
 use super::parser::{FilterParser, Rule, PREC_CLIMBER};
 use super::FacetNumberRange;
 use crate::error::UserError;
-use crate::heed_codec::facet::{FacetLevelValueF64Codec, FacetValueStringCodec};
+use crate::heed_codec::facet::{FacetLevelValueF64Codec, FacetStringLevelZeroCodec};
 use crate::{CboRoaringBitmapCodec, FieldId, FieldsIdsMap, Index, Result};
 
 #[derive(Debug, Clone, PartialEq)]
@@ -363,7 +363,7 @@ impl FilterCondition {
         rtxn: &heed::RoTxn,
         index: &Index,
         numbers_db: heed::Database<FacetLevelValueF64Codec, CboRoaringBitmapCodec>,
-        strings_db: heed::Database<FacetValueStringCodec, CboRoaringBitmapCodec>,
+        strings_db: heed::Database<FacetStringLevelZeroCodec, CboRoaringBitmapCodec>,
         field_id: FieldId,
         operator: &Operator,
     ) -> Result<RoaringBitmap> {

milli/src/update/index_documents/store.rs

@@ -26,7 +26,7 @@ use super::merge_function::{
 use super::{create_sorter, create_writer, writer_into_reader, MergeFn};
 use crate::error::{Error, InternalError, SerializationError};
 use crate::heed_codec::facet::{
-    FacetLevelValueF64Codec, FacetValueStringCodec, FieldDocIdFacetF64Codec,
+    FacetLevelValueF64Codec, FacetStringLevelZeroCodec, FieldDocIdFacetF64Codec,
     FieldDocIdFacetStringCodec,
 };
 use crate::heed_codec::{BoRoaringBitmapCodec, CboRoaringBitmapCodec};
@@ -522,7 +522,7 @@ impl<'s, A: AsRef<[u8]>> Store<'s, A> {
             key_buffer.clear();
             data_buffer.clear();
 
-            FacetValueStringCodec::serialize_into(field_id, &value, &mut key_buffer);
+            FacetStringLevelZeroCodec::serialize_into(field_id, &value, &mut key_buffer);
             CboRoaringBitmapCodec::serialize_into(&docids, &mut data_buffer);
 
             if lmdb_key_valid_size(&key_buffer) {