mirror of
https://github.com/meilisearch/meilisearch.git
synced 2025-08-01 03:10:04 +00:00
Reintroduce filter range search and facet extractors
This commit is contained in:
committed by
Loïc Lecrenier
parent
22d80eeaf9
commit
39a4a0a362
@ -2,22 +2,20 @@ use std::collections::btree_map::Entry;
|
||||
|
||||
use fst::IntoStreamer;
|
||||
use heed::types::{ByteSlice, Str};
|
||||
use heed::{BytesDecode, BytesEncode, Database};
|
||||
use obkv::Key;
|
||||
use heed::Database;
|
||||
use roaring::RoaringBitmap;
|
||||
use serde::{Deserialize, Serialize};
|
||||
use serde_json::Value;
|
||||
use time::OffsetDateTime;
|
||||
|
||||
use super::{ClearDocuments, Facets};
|
||||
use crate::error::{InternalError, SerializationError, UserError};
|
||||
// use crate::heed_codec::facet::FacetStringZeroBoundsValueCodec;
|
||||
use crate::error::{InternalError, UserError};
|
||||
use crate::heed_codec::facet::new::{FacetGroupValueCodec, FacetKeyCodec, MyByteSlice};
|
||||
use crate::heed_codec::CboRoaringBitmapCodec;
|
||||
use crate::index::{db_name, main_key};
|
||||
use crate::{
|
||||
fields_ids_map, DocumentId, ExternalDocumentsIds, FieldId, FieldIdMapMissingEntry,
|
||||
FieldsIdsMap, Index, Result, RoaringBitmapCodec, SmallString32, BEU32,
|
||||
DocumentId, ExternalDocumentsIds, FieldId, FieldIdMapMissingEntry, FieldsIdsMap, Index, Result,
|
||||
RoaringBitmapCodec, SmallString32, BEU32,
|
||||
};
|
||||
|
||||
pub struct DeleteDocuments<'t, 'u, 'i> {
|
||||
|
@ -6,6 +6,8 @@ use heed::{BytesDecode, BytesEncode};
|
||||
use super::helpers::{
|
||||
create_sorter, merge_cbo_roaring_bitmaps, sorter_into_reader, GrenadParameters,
|
||||
};
|
||||
use crate::heed_codec::facet::new::ordered_f64_codec::OrderedF64Codec;
|
||||
use crate::heed_codec::facet::new::{FacetKey, FacetKeyCodec};
|
||||
use crate::heed_codec::facet::FieldDocIdFacetF64Codec;
|
||||
use crate::Result;
|
||||
|
||||
@ -31,14 +33,13 @@ pub fn extract_facet_number_docids<R: io::Read + io::Seek>(
|
||||
|
||||
let mut cursor = docid_fid_facet_number.into_cursor()?;
|
||||
while let Some((key_bytes, _)) = cursor.move_on_next()? {
|
||||
todo!()
|
||||
// let (field_id, document_id, number) =
|
||||
// FieldDocIdFacetF64Codec::bytes_decode(key_bytes).unwrap();
|
||||
let (field_id, document_id, number) =
|
||||
FieldDocIdFacetF64Codec::bytes_decode(key_bytes).unwrap();
|
||||
|
||||
// let key = (field_id, 0, number, number);
|
||||
// // let key_bytes = FacetLevelValueF64Codec::bytes_encode(&key).unwrap();
|
||||
let key = FacetKey { field_id, level: 0, left_bound: number };
|
||||
let key_bytes = FacetKeyCodec::<OrderedF64Codec>::bytes_encode(&key).unwrap();
|
||||
|
||||
// facet_number_docids_sorter.insert(key_bytes, document_id.to_ne_bytes())?;
|
||||
facet_number_docids_sorter.insert(key_bytes, document_id.to_ne_bytes())?;
|
||||
}
|
||||
|
||||
sorter_into_reader(facet_number_docids_sorter, indexer)
|
||||
|
@ -1,13 +1,11 @@
|
||||
use std::fs::File;
|
||||
use std::iter::FromIterator;
|
||||
use std::{io, str};
|
||||
|
||||
use roaring::RoaringBitmap;
|
||||
|
||||
use super::helpers::{create_sorter, sorter_into_reader, try_split_array_at, GrenadParameters};
|
||||
use crate::heed_codec::facet::new::str_ref::StrRefCodec;
|
||||
use crate::heed_codec::facet::new::{FacetKey, FacetKeyCodec};
|
||||
use crate::update::index_documents::merge_cbo_roaring_bitmaps;
|
||||
// use crate::heed_codec::facet::{encode_prefix_string, FacetStringLevelZeroCodec};
|
||||
use crate::{FieldId, Result};
|
||||
use heed::BytesEncode;
|
||||
use std::fs::File;
|
||||
use std::io;
|
||||
|
||||
/// Extracts the facet string and the documents ids where this facet string appear.
|
||||
///
|
||||
@ -22,38 +20,26 @@ pub fn extract_facet_string_docids<R: io::Read + io::Seek>(
|
||||
|
||||
let mut facet_string_docids_sorter = create_sorter(
|
||||
grenad::SortAlgorithm::Stable,
|
||||
merge_cbo_roaring_bitmaps, // TODO: check
|
||||
merge_cbo_roaring_bitmaps, // TODO: check that it is correct
|
||||
indexer.chunk_compression_type,
|
||||
indexer.chunk_compression_level,
|
||||
indexer.max_nb_chunks,
|
||||
max_memory,
|
||||
);
|
||||
|
||||
let mut key_buffer = Vec::new();
|
||||
let mut value_buffer = Vec::new();
|
||||
let mut cursor = docid_fid_facet_string.into_cursor()?;
|
||||
while let Some((key, original_value_bytes)) = cursor.move_on_next()? {
|
||||
while let Some((key, _original_value_bytes)) = cursor.move_on_next()? {
|
||||
let (field_id_bytes, bytes) = try_split_array_at(key).unwrap();
|
||||
let field_id = FieldId::from_be_bytes(field_id_bytes);
|
||||
let (document_id_bytes, normalized_value_bytes) = try_split_array_at(bytes).unwrap();
|
||||
let document_id = u32::from_be_bytes(document_id_bytes);
|
||||
let original_value = str::from_utf8(original_value_bytes)?;
|
||||
|
||||
key_buffer.clear();
|
||||
// TODO
|
||||
// FacetStringLevelZeroCodec::serialize_into(
|
||||
// field_id,
|
||||
// str::from_utf8(normalized_value_bytes)?,
|
||||
// &mut key_buffer,
|
||||
// );
|
||||
let (document_id_bytes, normalized_value_bytes) =
|
||||
try_split_array_at::<_, 4>(bytes).unwrap();
|
||||
|
||||
value_buffer.clear();
|
||||
// TODO
|
||||
// encode_prefix_string(original_value, &mut value_buffer)?;
|
||||
let bitmap = RoaringBitmap::from_iter(Some(document_id));
|
||||
bitmap.serialize_into(&mut value_buffer)?;
|
||||
let normalised_value = std::str::from_utf8(normalized_value_bytes)?;
|
||||
let key = FacetKey { field_id, level: 0, left_bound: normalised_value };
|
||||
let key_bytes = FacetKeyCodec::<StrRefCodec>::bytes_encode(&key).unwrap();
|
||||
|
||||
facet_string_docids_sorter.insert(&key_buffer, &value_buffer)?;
|
||||
facet_string_docids_sorter.insert(&key_bytes, &document_id_bytes)?;
|
||||
}
|
||||
|
||||
sorter_into_reader(facet_string_docids_sorter, indexer)
|
||||
|
Reference in New Issue
Block a user