2020-10-25 21:49:04 +08:00
|
|
|
use std::borrow::Cow;
|
2021-04-28 23:58:16 +08:00
|
|
|
use std::collections::{HashMap, HashSet};
|
2021-07-29 21:49:23 +08:00
|
|
|
use std::mem::size_of;
|
2020-10-30 17:56:35 +08:00
|
|
|
use std::path::Path;
|
2020-10-25 21:49:04 +08:00
|
|
|
|
2021-04-07 16:53:57 +08:00
|
|
|
use chrono::{DateTime, Utc};
|
2021-07-05 22:36:13 +08:00
|
|
|
use heed::flags::Flags;
|
2020-10-21 21:55:48 +08:00
|
|
|
use heed::types::*;
|
2021-06-17 00:33:33 +08:00
|
|
|
use heed::{Database, PolyDatabase, RoTxn, RwTxn};
|
2020-10-21 21:55:48 +08:00
|
|
|
use roaring::RoaringBitmap;
|
2021-08-23 22:32:11 +08:00
|
|
|
use rstar::RTree;
|
2020-10-21 21:55:48 +08:00
|
|
|
|
2021-07-22 23:11:17 +08:00
|
|
|
use crate::error::{InternalError, UserError};
|
2021-06-17 00:33:33 +08:00
|
|
|
use crate::fields_ids_map::FieldsIdsMap;
|
2021-04-21 17:49:26 +08:00
|
|
|
use crate::heed_codec::facet::{
|
2021-07-17 18:50:01 +08:00
|
|
|
FacetLevelValueF64Codec, FacetStringLevelZeroCodec, FacetStringLevelZeroValueCodec,
|
|
|
|
FieldDocIdFacetF64Codec, FieldDocIdFacetStringCodec,
|
2021-06-17 00:33:33 +08:00
|
|
|
};
|
|
|
|
use crate::{
|
|
|
|
default_criteria, BEU32StrCodec, BoRoaringBitmapCodec, CboRoaringBitmapCodec, Criterion,
|
2021-06-22 21:36:22 +08:00
|
|
|
DocumentId, ExternalDocumentsIds, FacetDistribution, FieldDistribution, FieldId,
|
2021-08-23 22:32:11 +08:00
|
|
|
FieldIdWordCountCodec, GeoPoint, ObkvCodec, Result, RoaringBitmapCodec, RoaringBitmapLenCodec,
|
|
|
|
Search, StrLevelPositionCodec, StrStrU8Codec, BEU32,
|
2021-04-21 17:49:26 +08:00
|
|
|
};
|
2020-10-21 21:55:48 +08:00
|
|
|
|
2021-06-15 17:06:42 +08:00
|
|
|
pub mod main_key {
|
|
|
|
pub const CRITERIA_KEY: &str = "criteria";
|
|
|
|
pub const DISPLAYED_FIELDS_KEY: &str = "displayed-fields";
|
|
|
|
pub const DISTINCT_FIELD_KEY: &str = "distinct-field-key";
|
|
|
|
pub const DOCUMENTS_IDS_KEY: &str = "documents-ids";
|
|
|
|
pub const FILTERABLE_FIELDS_KEY: &str = "filterable-fields";
|
2021-08-23 17:37:18 +08:00
|
|
|
pub const SORTABLE_FIELDS_KEY: &str = "sortable-fields";
|
2021-06-17 21:16:20 +08:00
|
|
|
pub const FIELD_DISTRIBUTION_KEY: &str = "fields-distribution";
|
2021-06-15 17:06:42 +08:00
|
|
|
pub const FIELDS_IDS_MAP_KEY: &str = "fields-ids-map";
|
2021-08-23 22:32:11 +08:00
|
|
|
pub const GEO_RTREE_KEY: &str = "geo";
|
2021-06-15 17:06:42 +08:00
|
|
|
pub const HARD_EXTERNAL_DOCUMENTS_IDS_KEY: &str = "hard-external-documents-ids";
|
|
|
|
pub const NUMBER_FACETED_DOCUMENTS_IDS_PREFIX: &str = "number-faceted-documents-ids";
|
|
|
|
pub const PRIMARY_KEY_KEY: &str = "primary-key";
|
|
|
|
pub const SEARCHABLE_FIELDS_KEY: &str = "searchable-fields";
|
|
|
|
pub const SOFT_EXTERNAL_DOCUMENTS_IDS_KEY: &str = "soft-external-documents-ids";
|
|
|
|
pub const STOP_WORDS_KEY: &str = "stop-words";
|
|
|
|
pub const STRING_FACETED_DOCUMENTS_IDS_PREFIX: &str = "string-faceted-documents-ids";
|
|
|
|
pub const SYNONYMS_KEY: &str = "synonyms";
|
|
|
|
pub const WORDS_FST_KEY: &str = "words-fst";
|
|
|
|
pub const WORDS_PREFIXES_FST_KEY: &str = "words-prefixes-fst";
|
|
|
|
pub const CREATED_AT_KEY: &str = "created-at";
|
|
|
|
pub const UPDATED_AT_KEY: &str = "updated-at";
|
|
|
|
}
|
|
|
|
|
|
|
|
pub mod db_name {
|
|
|
|
pub const MAIN: &str = "main";
|
|
|
|
pub const WORD_DOCIDS: &str = "word-docids";
|
|
|
|
pub const WORD_PREFIX_DOCIDS: &str = "word-prefix-docids";
|
|
|
|
pub const DOCID_WORD_POSITIONS: &str = "docid-word-positions";
|
|
|
|
pub const WORD_PAIR_PROXIMITY_DOCIDS: &str = "word-pair-proximity-docids";
|
|
|
|
pub const WORD_PREFIX_PAIR_PROXIMITY_DOCIDS: &str = "word-prefix-pair-proximity-docids";
|
|
|
|
pub const WORD_LEVEL_POSITION_DOCIDS: &str = "word-level-position-docids";
|
|
|
|
pub const WORD_PREFIX_LEVEL_POSITION_DOCIDS: &str = "word-prefix-level-position-docids";
|
|
|
|
pub const FIELD_ID_WORD_COUNT_DOCIDS: &str = "field-id-word-count-docids";
|
|
|
|
pub const FACET_ID_F64_DOCIDS: &str = "facet-id-f64-docids";
|
|
|
|
pub const FACET_ID_STRING_DOCIDS: &str = "facet-id-string-docids";
|
|
|
|
pub const FIELD_ID_DOCID_FACET_F64S: &str = "field-id-docid-facet-f64s";
|
|
|
|
pub const FIELD_ID_DOCID_FACET_STRINGS: &str = "field-id-docid-facet-strings";
|
|
|
|
pub const DOCUMENTS: &str = "documents";
|
|
|
|
}
|
2020-10-21 21:55:48 +08:00
|
|
|
|
|
|
|
#[derive(Clone)]
|
|
|
|
pub struct Index {
|
2020-10-30 17:56:35 +08:00
|
|
|
/// The LMDB environment which this index is associated with.
|
|
|
|
pub env: heed::Env,
|
2021-04-21 17:49:26 +08:00
|
|
|
|
2020-10-22 20:23:33 +08:00
|
|
|
/// Contains many different types (e.g. the fields ids map).
|
2020-10-21 21:55:48 +08:00
|
|
|
pub main: PolyDatabase,
|
2021-04-21 17:49:26 +08:00
|
|
|
|
2020-10-21 21:55:48 +08:00
|
|
|
/// A word and all the documents ids containing the word.
|
|
|
|
pub word_docids: Database<Str, RoaringBitmapCodec>,
|
2021-02-03 17:30:33 +08:00
|
|
|
/// A prefix of word and all the documents ids containing this prefix.
|
|
|
|
pub word_prefix_docids: Database<Str, RoaringBitmapCodec>,
|
2021-04-21 17:49:26 +08:00
|
|
|
|
2020-10-21 21:55:48 +08:00
|
|
|
/// Maps a word and a document id (u32) to all the positions where the given word appears.
|
|
|
|
pub docid_word_positions: Database<BEU32StrCodec, BoRoaringBitmapCodec>,
|
2021-04-21 17:49:26 +08:00
|
|
|
|
2020-10-21 21:55:48 +08:00
|
|
|
/// Maps the proximity between a pair of words with all the docids where this relation appears.
|
|
|
|
pub word_pair_proximity_docids: Database<StrStrU8Codec, CboRoaringBitmapCodec>,
|
2021-02-10 17:28:15 +08:00
|
|
|
/// Maps the proximity between a pair of word and prefix with all the docids where this relation appears.
|
|
|
|
pub word_prefix_pair_proximity_docids: Database<StrStrU8Codec, CboRoaringBitmapCodec>,
|
2021-05-27 21:27:41 +08:00
|
|
|
|
2021-03-12 00:24:35 +08:00
|
|
|
/// Maps the word, level and position range with the docids that corresponds to it.
|
|
|
|
pub word_level_position_docids: Database<StrLevelPositionCodec, CboRoaringBitmapCodec>,
|
2021-05-27 21:27:41 +08:00
|
|
|
/// Maps the field id and the word count with the docids that corresponds to it.
|
|
|
|
pub field_id_word_count_docids: Database<FieldIdWordCountCodec, CboRoaringBitmapCodec>,
|
2021-03-25 18:10:12 +08:00
|
|
|
/// Maps the level positions of a word prefix with all the docids where this prefix appears.
|
|
|
|
pub word_prefix_level_position_docids: Database<StrLevelPositionCodec, CboRoaringBitmapCodec>,
|
2021-04-21 17:49:26 +08:00
|
|
|
|
|
|
|
/// Maps the facet field id, level and the number with the docids that corresponds to it.
|
|
|
|
pub facet_id_f64_docids: Database<FacetLevelValueF64Codec, CboRoaringBitmapCodec>,
|
2021-07-17 18:50:01 +08:00
|
|
|
/// Maps the facet field id and the string with the original string and docids that corresponds to it.
|
2021-08-16 19:36:30 +08:00
|
|
|
pub facet_id_string_docids: Database<FacetStringLevelZeroCodec, FacetStringLevelZeroValueCodec>,
|
2021-04-21 17:49:26 +08:00
|
|
|
|
|
|
|
/// Maps the document id, the facet field id and the numbers.
|
|
|
|
pub field_id_docid_facet_f64s: Database<FieldDocIdFacetF64Codec, Unit>,
|
|
|
|
/// Maps the document id, the facet field id and the strings.
|
2021-07-15 16:19:35 +08:00
|
|
|
pub field_id_docid_facet_strings: Database<FieldDocIdFacetStringCodec, Str>,
|
2021-04-21 17:49:26 +08:00
|
|
|
|
2020-10-22 20:23:33 +08:00
|
|
|
/// Maps the document id to the document as an obkv store.
|
2020-10-22 00:26:29 +08:00
|
|
|
pub documents: Database<OwnedType<BEU32>, ObkvCodec>,
|
2020-10-21 21:55:48 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
impl Index {
|
2021-06-14 22:46:19 +08:00
|
|
|
pub fn new<P: AsRef<Path>>(mut options: heed::EnvOpenOptions, path: P) -> Result<Index> {
|
2021-06-15 17:06:42 +08:00
|
|
|
use db_name::*;
|
|
|
|
|
2021-05-27 21:27:41 +08:00
|
|
|
options.max_dbs(14);
|
2021-07-05 22:36:13 +08:00
|
|
|
unsafe { options.flag(Flags::MdbAlwaysFreePages) };
|
2020-10-30 17:56:35 +08:00
|
|
|
|
|
|
|
let env = options.open(path)?;
|
2021-06-15 17:06:42 +08:00
|
|
|
let main = env.create_poly_database(Some(MAIN))?;
|
|
|
|
let word_docids = env.create_database(Some(WORD_DOCIDS))?;
|
|
|
|
let word_prefix_docids = env.create_database(Some(WORD_PREFIX_DOCIDS))?;
|
|
|
|
let docid_word_positions = env.create_database(Some(DOCID_WORD_POSITIONS))?;
|
|
|
|
let word_pair_proximity_docids = env.create_database(Some(WORD_PAIR_PROXIMITY_DOCIDS))?;
|
2021-06-17 00:33:33 +08:00
|
|
|
let word_prefix_pair_proximity_docids =
|
|
|
|
env.create_database(Some(WORD_PREFIX_PAIR_PROXIMITY_DOCIDS))?;
|
2021-06-15 17:06:42 +08:00
|
|
|
let word_level_position_docids = env.create_database(Some(WORD_LEVEL_POSITION_DOCIDS))?;
|
|
|
|
let field_id_word_count_docids = env.create_database(Some(FIELD_ID_WORD_COUNT_DOCIDS))?;
|
2021-06-17 00:33:33 +08:00
|
|
|
let word_prefix_level_position_docids =
|
|
|
|
env.create_database(Some(WORD_PREFIX_LEVEL_POSITION_DOCIDS))?;
|
2021-06-15 17:06:42 +08:00
|
|
|
let facet_id_f64_docids = env.create_database(Some(FACET_ID_F64_DOCIDS))?;
|
|
|
|
let facet_id_string_docids = env.create_database(Some(FACET_ID_STRING_DOCIDS))?;
|
|
|
|
let field_id_docid_facet_f64s = env.create_database(Some(FIELD_ID_DOCID_FACET_F64S))?;
|
2021-06-17 00:33:33 +08:00
|
|
|
let field_id_docid_facet_strings =
|
|
|
|
env.create_database(Some(FIELD_ID_DOCID_FACET_STRINGS))?;
|
2021-06-15 17:06:42 +08:00
|
|
|
let documents = env.create_database(Some(DOCUMENTS))?;
|
2020-10-30 17:56:35 +08:00
|
|
|
|
2021-04-21 17:48:23 +08:00
|
|
|
Index::initialize_creation_dates(&env, main)?;
|
2021-03-12 01:32:04 +08:00
|
|
|
|
2020-11-11 23:04:04 +08:00
|
|
|
Ok(Index {
|
|
|
|
env,
|
|
|
|
main,
|
|
|
|
word_docids,
|
2021-02-03 17:30:33 +08:00
|
|
|
word_prefix_docids,
|
2020-11-11 23:04:04 +08:00
|
|
|
docid_word_positions,
|
|
|
|
word_pair_proximity_docids,
|
2021-02-10 17:28:15 +08:00
|
|
|
word_prefix_pair_proximity_docids,
|
2021-03-12 00:24:35 +08:00
|
|
|
word_level_position_docids,
|
2021-03-25 18:10:12 +08:00
|
|
|
word_prefix_level_position_docids,
|
2021-05-27 21:27:41 +08:00
|
|
|
field_id_word_count_docids,
|
2021-04-21 17:49:26 +08:00
|
|
|
facet_id_f64_docids,
|
|
|
|
facet_id_string_docids,
|
|
|
|
field_id_docid_facet_f64s,
|
|
|
|
field_id_docid_facet_strings,
|
2020-11-11 23:04:04 +08:00
|
|
|
documents,
|
|
|
|
})
|
2020-10-30 17:56:35 +08:00
|
|
|
}
|
|
|
|
|
2021-04-21 17:49:26 +08:00
|
|
|
fn initialize_creation_dates(env: &heed::Env, main: PolyDatabase) -> heed::Result<()> {
|
|
|
|
let mut txn = env.write_txn()?;
|
|
|
|
// The db was just created, we update its metadata with the relevant information.
|
2021-06-15 17:06:42 +08:00
|
|
|
if main.get::<_, Str, SerdeJson<DateTime<Utc>>>(&txn, main_key::CREATED_AT_KEY)?.is_none() {
|
2021-04-21 17:49:26 +08:00
|
|
|
let now = Utc::now();
|
2021-06-15 17:06:42 +08:00
|
|
|
main.put::<_, Str, SerdeJson<DateTime<Utc>>>(&mut txn, main_key::UPDATED_AT_KEY, &now)?;
|
|
|
|
main.put::<_, Str, SerdeJson<DateTime<Utc>>>(&mut txn, main_key::CREATED_AT_KEY, &now)?;
|
2021-04-21 17:49:26 +08:00
|
|
|
txn.commit()?;
|
|
|
|
}
|
|
|
|
Ok(())
|
|
|
|
}
|
|
|
|
|
2020-10-30 17:56:35 +08:00
|
|
|
/// Create a write transaction to be able to write into the index.
|
2020-11-02 18:48:33 +08:00
|
|
|
pub fn write_txn(&self) -> heed::Result<RwTxn> {
|
2020-10-30 17:56:35 +08:00
|
|
|
self.env.write_txn()
|
|
|
|
}
|
|
|
|
|
|
|
|
/// Create a read transaction to be able to read the index.
|
2020-11-02 18:48:33 +08:00
|
|
|
pub fn read_txn(&self) -> heed::Result<RoTxn> {
|
2020-10-30 17:56:35 +08:00
|
|
|
self.env.read_txn()
|
2020-10-21 21:55:48 +08:00
|
|
|
}
|
|
|
|
|
2020-10-30 18:46:00 +08:00
|
|
|
/// Returns the canonicalized path where the heed `Env` of this `Index` lives.
|
|
|
|
pub fn path(&self) -> &Path {
|
|
|
|
self.env.path()
|
|
|
|
}
|
|
|
|
|
2020-10-30 18:46:14 +08:00
|
|
|
/// Returns an `EnvClosingEvent` that can be used to wait for the closing event,
|
|
|
|
/// multiple threads can wait on this event.
|
|
|
|
///
|
|
|
|
/// Make sure that you drop all the copies of `Index`es you have, env closing are triggered
|
|
|
|
/// when all references are dropped, the last one will eventually close the environment.
|
|
|
|
pub fn prepare_for_closing(self) -> heed::EnvClosingEvent {
|
|
|
|
self.env.prepare_for_closing()
|
|
|
|
}
|
|
|
|
|
2020-11-02 18:48:33 +08:00
|
|
|
/* documents ids */
|
|
|
|
|
2020-10-22 21:33:09 +08:00
|
|
|
/// Writes the documents ids that corresponds to the user-ids-documents-ids FST.
|
2021-06-17 00:33:33 +08:00
|
|
|
pub(crate) fn put_documents_ids(
|
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
docids: &RoaringBitmap,
|
|
|
|
) -> heed::Result<()> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.put::<_, Str, RoaringBitmapCodec>(wtxn, main_key::DOCUMENTS_IDS_KEY, docids)
|
2020-10-21 21:55:48 +08:00
|
|
|
}
|
|
|
|
|
2020-10-22 21:33:09 +08:00
|
|
|
/// Returns the internal documents ids.
|
2020-11-02 18:48:33 +08:00
|
|
|
pub fn documents_ids(&self, rtxn: &RoTxn) -> heed::Result<RoaringBitmap> {
|
2021-06-17 00:33:33 +08:00
|
|
|
Ok(self
|
|
|
|
.main
|
|
|
|
.get::<_, Str, RoaringBitmapCodec>(rtxn, main_key::DOCUMENTS_IDS_KEY)?
|
|
|
|
.unwrap_or_default())
|
2020-10-21 21:55:48 +08:00
|
|
|
}
|
|
|
|
|
2021-03-09 17:24:27 +08:00
|
|
|
/// Returns the number of documents indexed in the database.
|
2021-06-14 22:46:19 +08:00
|
|
|
pub fn number_of_documents(&self, rtxn: &RoTxn) -> Result<u64> {
|
2021-06-17 00:33:33 +08:00
|
|
|
let count =
|
|
|
|
self.main.get::<_, Str, RoaringBitmapLenCodec>(rtxn, main_key::DOCUMENTS_IDS_KEY)?;
|
2021-03-09 17:24:27 +08:00
|
|
|
Ok(count.unwrap_or_default())
|
|
|
|
}
|
|
|
|
|
2020-11-02 18:48:33 +08:00
|
|
|
/* primary key */
|
|
|
|
|
2020-10-31 18:28:48 +08:00
|
|
|
/// Writes the documents primary key, this is the field name that is used to store the id.
|
2021-06-15 19:45:20 +08:00
|
|
|
pub(crate) fn put_primary_key(&self, wtxn: &mut RwTxn, primary_key: &str) -> heed::Result<()> {
|
2021-03-12 21:43:17 +08:00
|
|
|
self.set_updated_at(wtxn, &Utc::now())?;
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.put::<_, Str, Str>(wtxn, main_key::PRIMARY_KEY_KEY, &primary_key)
|
2020-10-31 18:28:48 +08:00
|
|
|
}
|
|
|
|
|
2020-11-02 00:52:04 +08:00
|
|
|
/// Deletes the primary key of the documents, this can be done to reset indexes settings.
|
2021-06-30 16:07:31 +08:00
|
|
|
pub(crate) fn delete_primary_key(&self, wtxn: &mut RwTxn) -> heed::Result<bool> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.delete::<_, Str>(wtxn, main_key::PRIMARY_KEY_KEY)
|
2020-10-31 18:28:48 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/// Returns the documents primary key, `None` if it hasn't been defined.
|
2021-01-21 00:27:43 +08:00
|
|
|
pub fn primary_key<'t>(&self, rtxn: &'t RoTxn) -> heed::Result<Option<&'t str>> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.get::<_, Str, Str>(rtxn, main_key::PRIMARY_KEY_KEY)
|
2020-10-31 18:28:48 +08:00
|
|
|
}
|
|
|
|
|
2020-11-22 18:54:04 +08:00
|
|
|
/* external documents ids */
|
2020-11-02 18:48:33 +08:00
|
|
|
|
2020-11-23 00:28:41 +08:00
|
|
|
/// Writes the external documents ids and internal ids (i.e. `u32`).
|
2021-06-15 19:45:20 +08:00
|
|
|
pub(crate) fn put_external_documents_ids<'a>(
|
2020-11-23 00:28:41 +08:00
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
external_documents_ids: &ExternalDocumentsIds<'a>,
|
2021-06-17 00:33:33 +08:00
|
|
|
) -> heed::Result<()> {
|
2020-11-23 00:28:41 +08:00
|
|
|
let ExternalDocumentsIds { hard, soft } = external_documents_ids;
|
|
|
|
let hard = hard.as_fst().as_bytes();
|
|
|
|
let soft = soft.as_fst().as_bytes();
|
2021-06-17 00:33:33 +08:00
|
|
|
self.main.put::<_, Str, ByteSlice>(
|
|
|
|
wtxn,
|
|
|
|
main_key::HARD_EXTERNAL_DOCUMENTS_IDS_KEY,
|
|
|
|
hard,
|
|
|
|
)?;
|
|
|
|
self.main.put::<_, Str, ByteSlice>(
|
|
|
|
wtxn,
|
|
|
|
main_key::SOFT_EXTERNAL_DOCUMENTS_IDS_KEY,
|
|
|
|
soft,
|
|
|
|
)?;
|
2020-11-23 00:28:41 +08:00
|
|
|
Ok(())
|
2020-10-21 21:55:48 +08:00
|
|
|
}
|
|
|
|
|
2020-11-23 00:28:41 +08:00
|
|
|
/// Returns the external documents ids map which associate the external ids
|
2020-10-22 21:33:09 +08:00
|
|
|
/// with the internal ids (i.e. `u32`).
|
2021-06-14 22:46:19 +08:00
|
|
|
pub fn external_documents_ids<'t>(&self, rtxn: &'t RoTxn) -> Result<ExternalDocumentsIds<'t>> {
|
2021-06-17 00:33:33 +08:00
|
|
|
let hard =
|
|
|
|
self.main.get::<_, Str, ByteSlice>(rtxn, main_key::HARD_EXTERNAL_DOCUMENTS_IDS_KEY)?;
|
|
|
|
let soft =
|
|
|
|
self.main.get::<_, Str, ByteSlice>(rtxn, main_key::SOFT_EXTERNAL_DOCUMENTS_IDS_KEY)?;
|
2020-11-23 00:53:33 +08:00
|
|
|
let hard = match hard {
|
|
|
|
Some(hard) => fst::Map::new(hard)?.map_data(Cow::Borrowed)?,
|
|
|
|
None => fst::Map::default().map_data(Cow::Owned)?,
|
|
|
|
};
|
|
|
|
let soft = match soft {
|
|
|
|
Some(soft) => fst::Map::new(soft)?.map_data(Cow::Borrowed)?,
|
|
|
|
None => fst::Map::default().map_data(Cow::Owned)?,
|
|
|
|
};
|
|
|
|
Ok(ExternalDocumentsIds::new(hard, soft))
|
2020-10-21 21:55:48 +08:00
|
|
|
}
|
|
|
|
|
2020-11-02 18:48:33 +08:00
|
|
|
/* fields ids map */
|
|
|
|
|
2020-10-22 21:33:09 +08:00
|
|
|
/// Writes the fields ids map which associate the documents keys with an internal field id
|
|
|
|
/// (i.e. `u8`), this field id is used to identify fields in the obkv documents.
|
2021-06-17 00:33:33 +08:00
|
|
|
pub(crate) fn put_fields_ids_map(
|
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
map: &FieldsIdsMap,
|
|
|
|
) -> heed::Result<()> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.put::<_, Str, SerdeJson<FieldsIdsMap>>(wtxn, main_key::FIELDS_IDS_MAP_KEY, map)
|
2020-10-22 21:33:09 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/// Returns the fields ids map which associate the documents keys with an internal field id
|
|
|
|
/// (i.e. `u8`), this field id is used to identify fields in the obkv documents.
|
2020-11-02 18:48:33 +08:00
|
|
|
pub fn fields_ids_map(&self, rtxn: &RoTxn) -> heed::Result<FieldsIdsMap> {
|
2021-06-17 00:33:33 +08:00
|
|
|
Ok(self
|
|
|
|
.main
|
|
|
|
.get::<_, Str, SerdeJson<FieldsIdsMap>>(rtxn, main_key::FIELDS_IDS_MAP_KEY)?
|
|
|
|
.unwrap_or_default())
|
2020-10-22 21:33:09 +08:00
|
|
|
}
|
|
|
|
|
2021-08-23 22:32:11 +08:00
|
|
|
/* geo rtree */
|
|
|
|
|
|
|
|
pub(crate) fn put_geo_rtree<A: AsRef<[u8]>>(
|
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
rtree: &RTree<GeoPoint>,
|
|
|
|
) -> heed::Result<()> {
|
|
|
|
self.main.put::<_, Str, SerdeBincode<RTree<GeoPoint>>>(wtxn, main_key::GEO_RTREE_KEY, rtree)
|
|
|
|
}
|
|
|
|
|
|
|
|
pub(crate) fn delete_geo_rtree(&self, wtxn: &mut RwTxn) -> heed::Result<bool> {
|
|
|
|
self.main.delete::<_, Str>(wtxn, main_key::GEO_RTREE_KEY)
|
|
|
|
}
|
|
|
|
|
|
|
|
pub fn geo_rtree<'t>(&self, rtxn: &'t RoTxn) -> Result<Option<RTree<GeoPoint>>> {
|
|
|
|
match self
|
|
|
|
.main
|
|
|
|
.get::<_, Str, SerdeBincode<RTree<GeoPoint>>>(rtxn, main_key::GEO_RTREE_KEY)?
|
|
|
|
{
|
|
|
|
Some(rtree) => Ok(Some(rtree)),
|
|
|
|
None => Ok(None),
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-06-17 21:16:20 +08:00
|
|
|
/* field distribution */
|
2021-03-31 23:14:23 +08:00
|
|
|
|
2021-06-17 21:16:20 +08:00
|
|
|
/// Writes the field distribution which associates every field name with
|
2021-04-01 15:07:16 +08:00
|
|
|
/// the number of times it occurs in the documents.
|
2021-06-17 21:16:20 +08:00
|
|
|
pub(crate) fn put_field_distribution(
|
2021-06-17 00:33:33 +08:00
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
2021-06-21 21:57:41 +08:00
|
|
|
distribution: &FieldDistribution,
|
2021-06-17 00:33:33 +08:00
|
|
|
) -> heed::Result<()> {
|
2021-06-21 21:57:41 +08:00
|
|
|
self.main.put::<_, Str, SerdeJson<FieldDistribution>>(
|
2021-06-17 00:33:33 +08:00
|
|
|
wtxn,
|
2021-06-17 21:16:20 +08:00
|
|
|
main_key::FIELD_DISTRIBUTION_KEY,
|
2021-06-17 00:33:33 +08:00
|
|
|
distribution,
|
|
|
|
)
|
2021-03-31 23:14:23 +08:00
|
|
|
}
|
2021-03-31 23:14:23 +08:00
|
|
|
|
2021-06-17 21:16:20 +08:00
|
|
|
/// Returns the field distribution which associates every field name with
|
2021-04-01 15:07:16 +08:00
|
|
|
/// the number of times it occurs in the documents.
|
2021-06-21 21:57:41 +08:00
|
|
|
pub fn field_distribution(&self, rtxn: &RoTxn) -> heed::Result<FieldDistribution> {
|
2021-06-17 00:33:33 +08:00
|
|
|
Ok(self
|
|
|
|
.main
|
2021-06-21 21:57:41 +08:00
|
|
|
.get::<_, Str, SerdeJson<FieldDistribution>>(rtxn, main_key::FIELD_DISTRIBUTION_KEY)?
|
2021-06-17 00:33:33 +08:00
|
|
|
.unwrap_or_default())
|
2021-03-31 23:14:23 +08:00
|
|
|
}
|
|
|
|
|
2020-11-02 18:48:33 +08:00
|
|
|
/* displayed fields */
|
|
|
|
|
2021-01-21 00:27:43 +08:00
|
|
|
/// Writes the fields that must be displayed in the defined order.
|
2020-11-02 18:45:16 +08:00
|
|
|
/// There must be not be any duplicate field id.
|
2021-06-17 00:33:33 +08:00
|
|
|
pub(crate) fn put_displayed_fields(
|
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
fields: &[&str],
|
|
|
|
) -> heed::Result<()> {
|
|
|
|
self.main.put::<_, Str, SerdeBincode<&[&str]>>(
|
|
|
|
wtxn,
|
|
|
|
main_key::DISPLAYED_FIELDS_KEY,
|
|
|
|
&fields,
|
|
|
|
)
|
2020-11-02 18:45:16 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/// Deletes the displayed fields ids, this will make the engine to display
|
|
|
|
/// all the documents attributes in the order of the `FieldsIdsMap`.
|
2021-06-30 16:07:31 +08:00
|
|
|
pub(crate) fn delete_displayed_fields(&self, wtxn: &mut RwTxn) -> heed::Result<bool> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.delete::<_, Str>(wtxn, main_key::DISPLAYED_FIELDS_KEY)
|
2020-11-02 18:45:16 +08:00
|
|
|
}
|
|
|
|
|
2021-01-21 00:27:43 +08:00
|
|
|
/// Returns the displayed fields in the order they were set by the user. If it returns
|
|
|
|
/// `None` it means that all the attributes are set as displayed in the order of the `FieldsIdsMap`.
|
|
|
|
pub fn displayed_fields<'t>(&self, rtxn: &'t RoTxn) -> heed::Result<Option<Vec<&'t str>>> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.get::<_, Str, SerdeBincode<Vec<&'t str>>>(rtxn, main_key::DISPLAYED_FIELDS_KEY)
|
2021-01-21 00:27:43 +08:00
|
|
|
}
|
|
|
|
|
2021-06-15 17:51:32 +08:00
|
|
|
/// Identical to `displayed_fields`, but returns the ids instead.
|
|
|
|
pub fn displayed_fields_ids(&self, rtxn: &RoTxn) -> Result<Option<Vec<FieldId>>> {
|
|
|
|
match self.displayed_fields(rtxn)? {
|
|
|
|
Some(fields) => {
|
|
|
|
let fields_ids_map = self.fields_ids_map(rtxn)?;
|
|
|
|
let mut fields_ids = Vec::new();
|
|
|
|
for name in fields.into_iter() {
|
2021-07-22 23:11:17 +08:00
|
|
|
if let Some(field_id) = fields_ids_map.id(name) {
|
|
|
|
fields_ids.push(field_id);
|
2021-06-15 17:51:32 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
Ok(Some(fields_ids))
|
2021-06-17 00:33:33 +08:00
|
|
|
}
|
2021-06-15 17:51:32 +08:00
|
|
|
None => Ok(None),
|
|
|
|
}
|
2020-11-02 18:45:16 +08:00
|
|
|
}
|
|
|
|
|
2020-11-02 00:52:04 +08:00
|
|
|
/* searchable fields */
|
|
|
|
|
|
|
|
/// Writes the searchable fields, when this list is specified, only these are indexed.
|
2021-06-17 00:33:33 +08:00
|
|
|
pub(crate) fn put_searchable_fields(
|
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
fields: &[&str],
|
|
|
|
) -> heed::Result<()> {
|
|
|
|
self.main.put::<_, Str, SerdeBincode<&[&str]>>(
|
|
|
|
wtxn,
|
|
|
|
main_key::SEARCHABLE_FIELDS_KEY,
|
|
|
|
&fields,
|
|
|
|
)
|
2020-11-02 00:52:04 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/// Deletes the searchable fields, when no fields are specified, all fields are indexed.
|
2021-06-30 16:07:31 +08:00
|
|
|
pub(crate) fn delete_searchable_fields(&self, wtxn: &mut RwTxn) -> heed::Result<bool> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.delete::<_, Str>(wtxn, main_key::SEARCHABLE_FIELDS_KEY)
|
2020-11-02 00:52:04 +08:00
|
|
|
}
|
|
|
|
|
2021-01-21 00:27:43 +08:00
|
|
|
/// Returns the searchable fields, those are the fields that are indexed,
|
2020-11-02 00:52:04 +08:00
|
|
|
/// if the searchable fields aren't there it means that **all** the fields are indexed.
|
2021-01-21 00:27:43 +08:00
|
|
|
pub fn searchable_fields<'t>(&self, rtxn: &'t RoTxn) -> heed::Result<Option<Vec<&'t str>>> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.get::<_, Str, SerdeBincode<Vec<&'t str>>>(rtxn, main_key::SEARCHABLE_FIELDS_KEY)
|
2021-01-21 00:27:43 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/// Identical to `searchable_fields`, but returns the ids instead.
|
2021-06-15 17:51:32 +08:00
|
|
|
pub fn searchable_fields_ids(&self, rtxn: &RoTxn) -> Result<Option<Vec<FieldId>>> {
|
2021-01-21 00:27:43 +08:00
|
|
|
match self.searchable_fields(rtxn)? {
|
2021-06-15 17:51:32 +08:00
|
|
|
Some(fields) => {
|
|
|
|
let fields_ids_map = self.fields_ids_map(rtxn)?;
|
|
|
|
let mut fields_ids = Vec::new();
|
|
|
|
for name in fields {
|
2021-07-22 23:11:17 +08:00
|
|
|
if let Some(field_id) = fields_ids_map.id(name) {
|
|
|
|
fields_ids.push(field_id);
|
2021-06-15 17:51:32 +08:00
|
|
|
}
|
2021-01-21 00:27:43 +08:00
|
|
|
}
|
2021-06-15 17:51:32 +08:00
|
|
|
Ok(Some(fields_ids))
|
2021-06-17 00:33:33 +08:00
|
|
|
}
|
2021-01-21 00:27:43 +08:00
|
|
|
None => Ok(None),
|
|
|
|
}
|
2020-11-02 00:52:04 +08:00
|
|
|
}
|
|
|
|
|
2021-06-01 18:19:55 +08:00
|
|
|
/* filterable fields */
|
2020-11-11 23:17:37 +08:00
|
|
|
|
2021-06-01 18:19:55 +08:00
|
|
|
/// Writes the filterable fields names in the database.
|
2021-06-17 00:33:33 +08:00
|
|
|
pub(crate) fn put_filterable_fields(
|
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
fields: &HashSet<String>,
|
|
|
|
) -> heed::Result<()> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.put::<_, Str, SerdeJson<_>>(wtxn, main_key::FILTERABLE_FIELDS_KEY, fields)
|
2020-11-11 23:17:37 +08:00
|
|
|
}
|
|
|
|
|
2021-06-01 18:19:55 +08:00
|
|
|
/// Deletes the filterable fields ids in the database.
|
2021-06-30 16:07:31 +08:00
|
|
|
pub(crate) fn delete_filterable_fields(&self, wtxn: &mut RwTxn) -> heed::Result<bool> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.delete::<_, Str>(wtxn, main_key::FILTERABLE_FIELDS_KEY)
|
2020-11-11 23:17:37 +08:00
|
|
|
}
|
|
|
|
|
2021-06-01 18:19:55 +08:00
|
|
|
/// Returns the filterable fields names.
|
|
|
|
pub fn filterable_fields(&self, rtxn: &RoTxn) -> heed::Result<HashSet<String>> {
|
2021-06-17 00:33:33 +08:00
|
|
|
Ok(self
|
|
|
|
.main
|
|
|
|
.get::<_, Str, SerdeJson<_>>(rtxn, main_key::FILTERABLE_FIELDS_KEY)?
|
|
|
|
.unwrap_or_default())
|
2021-06-01 18:19:55 +08:00
|
|
|
}
|
|
|
|
|
2021-06-15 17:51:32 +08:00
|
|
|
/// Identical to `filterable_fields`, but returns ids instead.
|
|
|
|
pub fn filterable_fields_ids(&self, rtxn: &RoTxn) -> Result<HashSet<FieldId>> {
|
|
|
|
let fields = self.filterable_fields(rtxn)?;
|
2021-06-01 18:19:55 +08:00
|
|
|
let fields_ids_map = self.fields_ids_map(rtxn)?;
|
|
|
|
|
2021-06-15 17:51:32 +08:00
|
|
|
let mut fields_ids = HashSet::new();
|
|
|
|
for name in fields {
|
2021-07-22 23:11:17 +08:00
|
|
|
if let Some(field_id) = fields_ids_map.id(&name) {
|
|
|
|
fields_ids.insert(field_id);
|
2021-06-15 17:51:32 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
Ok(fields_ids)
|
2021-06-01 18:19:55 +08:00
|
|
|
}
|
2021-06-01 18:20:29 +08:00
|
|
|
|
2021-08-23 17:37:18 +08:00
|
|
|
/* sortable fields */
|
|
|
|
|
|
|
|
/// Writes the sortable fields names in the database.
|
|
|
|
pub(crate) fn put_sortable_fields(
|
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
fields: &HashSet<String>,
|
|
|
|
) -> heed::Result<()> {
|
|
|
|
self.main.put::<_, Str, SerdeJson<_>>(wtxn, main_key::SORTABLE_FIELDS_KEY, fields)
|
|
|
|
}
|
|
|
|
|
|
|
|
/// Deletes the sortable fields ids in the database.
|
|
|
|
pub(crate) fn delete_sortable_fields(&self, wtxn: &mut RwTxn) -> heed::Result<bool> {
|
|
|
|
self.main.delete::<_, Str>(wtxn, main_key::SORTABLE_FIELDS_KEY)
|
|
|
|
}
|
|
|
|
|
|
|
|
/// Returns the sortable fields names.
|
|
|
|
pub fn sortable_fields(&self, rtxn: &RoTxn) -> heed::Result<HashSet<String>> {
|
|
|
|
Ok(self
|
|
|
|
.main
|
|
|
|
.get::<_, Str, SerdeJson<_>>(rtxn, main_key::SORTABLE_FIELDS_KEY)?
|
|
|
|
.unwrap_or_default())
|
|
|
|
}
|
|
|
|
|
|
|
|
/// Identical to `sortable_fields`, but returns ids instead.
|
|
|
|
pub fn sortable_fields_ids(&self, rtxn: &RoTxn) -> Result<HashSet<FieldId>> {
|
|
|
|
let fields = self.sortable_fields(rtxn)?;
|
|
|
|
let fields_ids_map = self.fields_ids_map(rtxn)?;
|
|
|
|
Ok(fields.into_iter().filter_map(|name| fields_ids_map.id(&name)).collect())
|
|
|
|
}
|
|
|
|
|
2021-06-01 18:20:29 +08:00
|
|
|
/* faceted documents ids */
|
|
|
|
|
|
|
|
/// Returns the faceted fields names.
|
|
|
|
///
|
2021-08-23 17:37:18 +08:00
|
|
|
/// Faceted fields are the union of all the filterable, sortable, distinct, and Asc/Desc fields.
|
2021-06-15 17:51:32 +08:00
|
|
|
pub fn faceted_fields(&self, rtxn: &RoTxn) -> Result<HashSet<String>> {
|
2021-06-01 18:20:29 +08:00
|
|
|
let filterable_fields = self.filterable_fields(rtxn)?;
|
2021-08-23 17:37:18 +08:00
|
|
|
let sortable_fields = self.sortable_fields(rtxn)?;
|
2021-06-01 22:29:14 +08:00
|
|
|
let distinct_field = self.distinct_field(rtxn)?;
|
2021-06-17 00:33:33 +08:00
|
|
|
let asc_desc_fields =
|
|
|
|
self.criteria(rtxn)?.into_iter().filter_map(|criterion| match criterion {
|
2021-06-01 18:20:29 +08:00
|
|
|
Criterion::Asc(field) | Criterion::Desc(field) => Some(field),
|
|
|
|
_otherwise => None,
|
|
|
|
});
|
|
|
|
|
|
|
|
let mut faceted_fields = filterable_fields;
|
2021-08-23 17:37:18 +08:00
|
|
|
faceted_fields.extend(sortable_fields);
|
2021-06-01 18:20:29 +08:00
|
|
|
faceted_fields.extend(asc_desc_fields);
|
|
|
|
if let Some(field) = distinct_field {
|
|
|
|
faceted_fields.insert(field.to_owned());
|
|
|
|
}
|
|
|
|
|
|
|
|
Ok(faceted_fields)
|
2021-01-21 00:27:43 +08:00
|
|
|
}
|
|
|
|
|
2021-06-15 17:51:32 +08:00
|
|
|
/// Identical to `faceted_fields`, but returns ids instead.
|
|
|
|
pub fn faceted_fields_ids(&self, rtxn: &RoTxn) -> Result<HashSet<FieldId>> {
|
|
|
|
let fields = self.faceted_fields(rtxn)?;
|
2021-01-21 00:27:43 +08:00
|
|
|
let fields_ids_map = self.fields_ids_map(rtxn)?;
|
2021-04-28 23:58:16 +08:00
|
|
|
|
2021-06-15 17:51:32 +08:00
|
|
|
let mut fields_ids = HashSet::new();
|
|
|
|
for name in fields.into_iter() {
|
2021-07-22 23:11:17 +08:00
|
|
|
if let Some(field_id) = fields_ids_map.id(&name) {
|
|
|
|
fields_ids.insert(field_id);
|
2021-06-15 17:51:32 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
Ok(fields_ids)
|
2020-11-11 23:17:37 +08:00
|
|
|
}
|
|
|
|
|
2020-11-23 20:08:57 +08:00
|
|
|
/* faceted documents ids */
|
|
|
|
|
2021-04-28 23:58:16 +08:00
|
|
|
/// Writes the documents ids that are faceted with numbers under this field id.
|
2021-06-15 19:45:20 +08:00
|
|
|
pub(crate) fn put_number_faceted_documents_ids(
|
2021-04-28 23:58:16 +08:00
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
field_id: FieldId,
|
|
|
|
docids: &RoaringBitmap,
|
2021-06-17 00:33:33 +08:00
|
|
|
) -> heed::Result<()> {
|
2021-07-29 21:49:23 +08:00
|
|
|
let mut buffer =
|
|
|
|
[0u8; main_key::NUMBER_FACETED_DOCUMENTS_IDS_PREFIX.len() + size_of::<FieldId>()];
|
2021-07-23 01:18:22 +08:00
|
|
|
buffer[..main_key::NUMBER_FACETED_DOCUMENTS_IDS_PREFIX.len()]
|
|
|
|
.copy_from_slice(main_key::NUMBER_FACETED_DOCUMENTS_IDS_PREFIX.as_bytes());
|
|
|
|
buffer[main_key::NUMBER_FACETED_DOCUMENTS_IDS_PREFIX.len()..]
|
2021-07-06 17:31:24 +08:00
|
|
|
.copy_from_slice(&field_id.to_be_bytes());
|
2021-04-28 23:58:16 +08:00
|
|
|
self.main.put::<_, ByteSlice, RoaringBitmapCodec>(wtxn, &buffer, docids)
|
|
|
|
}
|
|
|
|
|
|
|
|
/// Retrieve all the documents ids that faceted with numbers under this field id.
|
|
|
|
pub fn number_faceted_documents_ids(
|
|
|
|
&self,
|
|
|
|
rtxn: &RoTxn,
|
|
|
|
field_id: FieldId,
|
2021-06-17 00:33:33 +08:00
|
|
|
) -> heed::Result<RoaringBitmap> {
|
2021-07-29 21:49:23 +08:00
|
|
|
let mut buffer =
|
|
|
|
[0u8; main_key::NUMBER_FACETED_DOCUMENTS_IDS_PREFIX.len() + size_of::<FieldId>()];
|
2021-07-23 01:18:22 +08:00
|
|
|
buffer[..main_key::NUMBER_FACETED_DOCUMENTS_IDS_PREFIX.len()]
|
|
|
|
.copy_from_slice(main_key::NUMBER_FACETED_DOCUMENTS_IDS_PREFIX.as_bytes());
|
|
|
|
buffer[main_key::NUMBER_FACETED_DOCUMENTS_IDS_PREFIX.len()..]
|
2021-07-06 17:31:24 +08:00
|
|
|
.copy_from_slice(&field_id.to_be_bytes());
|
2021-04-28 23:58:16 +08:00
|
|
|
match self.main.get::<_, ByteSlice, RoaringBitmapCodec>(rtxn, &buffer)? {
|
|
|
|
Some(docids) => Ok(docids),
|
|
|
|
None => Ok(RoaringBitmap::new()),
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/// Writes the documents ids that are faceted with strings under this field id.
|
2021-06-15 19:45:20 +08:00
|
|
|
pub(crate) fn put_string_faceted_documents_ids(
|
2021-04-28 23:58:16 +08:00
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
field_id: FieldId,
|
|
|
|
docids: &RoaringBitmap,
|
2021-06-17 00:33:33 +08:00
|
|
|
) -> heed::Result<()> {
|
2021-07-29 21:49:23 +08:00
|
|
|
let mut buffer =
|
|
|
|
[0u8; main_key::STRING_FACETED_DOCUMENTS_IDS_PREFIX.len() + size_of::<FieldId>()];
|
2021-07-23 01:18:22 +08:00
|
|
|
buffer[..main_key::STRING_FACETED_DOCUMENTS_IDS_PREFIX.len()]
|
|
|
|
.copy_from_slice(main_key::STRING_FACETED_DOCUMENTS_IDS_PREFIX.as_bytes());
|
|
|
|
buffer[main_key::STRING_FACETED_DOCUMENTS_IDS_PREFIX.len()..]
|
2021-07-06 17:31:24 +08:00
|
|
|
.copy_from_slice(&field_id.to_be_bytes());
|
2020-11-23 20:08:57 +08:00
|
|
|
self.main.put::<_, ByteSlice, RoaringBitmapCodec>(wtxn, &buffer, docids)
|
|
|
|
}
|
|
|
|
|
2021-04-28 23:58:16 +08:00
|
|
|
/// Retrieve all the documents ids that faceted with strings under this field id.
|
|
|
|
pub fn string_faceted_documents_ids(
|
|
|
|
&self,
|
|
|
|
rtxn: &RoTxn,
|
|
|
|
field_id: FieldId,
|
2021-06-17 00:33:33 +08:00
|
|
|
) -> heed::Result<RoaringBitmap> {
|
2021-07-29 21:49:23 +08:00
|
|
|
let mut buffer =
|
|
|
|
[0u8; main_key::STRING_FACETED_DOCUMENTS_IDS_PREFIX.len() + size_of::<FieldId>()];
|
2021-07-23 01:18:22 +08:00
|
|
|
buffer[..main_key::STRING_FACETED_DOCUMENTS_IDS_PREFIX.len()]
|
|
|
|
.copy_from_slice(main_key::STRING_FACETED_DOCUMENTS_IDS_PREFIX.as_bytes());
|
|
|
|
buffer[main_key::STRING_FACETED_DOCUMENTS_IDS_PREFIX.len()..]
|
2021-07-06 17:31:24 +08:00
|
|
|
.copy_from_slice(&field_id.to_be_bytes());
|
2020-11-23 20:08:57 +08:00
|
|
|
match self.main.get::<_, ByteSlice, RoaringBitmapCodec>(rtxn, &buffer)? {
|
|
|
|
Some(docids) => Ok(docids),
|
|
|
|
None => Ok(RoaringBitmap::new()),
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-06-01 22:29:14 +08:00
|
|
|
/* distinct field */
|
2021-04-14 18:00:45 +08:00
|
|
|
|
2021-06-17 00:33:33 +08:00
|
|
|
pub(crate) fn put_distinct_field(
|
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
distinct_field: &str,
|
|
|
|
) -> heed::Result<()> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.put::<_, Str, Str>(wtxn, main_key::DISTINCT_FIELD_KEY, distinct_field)
|
2021-04-14 18:00:45 +08:00
|
|
|
}
|
|
|
|
|
2021-06-01 22:29:14 +08:00
|
|
|
pub fn distinct_field<'a>(&self, rtxn: &'a RoTxn) -> heed::Result<Option<&'a str>> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.get::<_, Str, Str>(rtxn, main_key::DISTINCT_FIELD_KEY)
|
2021-04-14 18:00:45 +08:00
|
|
|
}
|
|
|
|
|
2021-06-01 22:29:14 +08:00
|
|
|
pub(crate) fn delete_distinct_field(&self, wtxn: &mut RwTxn) -> heed::Result<bool> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.delete::<_, Str>(wtxn, main_key::DISTINCT_FIELD_KEY)
|
2021-04-14 18:00:45 +08:00
|
|
|
}
|
|
|
|
|
2020-11-27 19:14:56 +08:00
|
|
|
/* criteria */
|
|
|
|
|
2021-06-17 00:33:33 +08:00
|
|
|
pub(crate) fn put_criteria(
|
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
criteria: &[Criterion],
|
|
|
|
) -> heed::Result<()> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.put::<_, Str, SerdeJson<&[Criterion]>>(wtxn, main_key::CRITERIA_KEY, &criteria)
|
2020-11-27 19:14:56 +08:00
|
|
|
}
|
|
|
|
|
2021-06-30 16:07:31 +08:00
|
|
|
pub(crate) fn delete_criteria(&self, wtxn: &mut RwTxn) -> heed::Result<bool> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.delete::<_, Str>(wtxn, main_key::CRITERIA_KEY)
|
2020-11-27 19:14:56 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
pub fn criteria(&self, rtxn: &RoTxn) -> heed::Result<Vec<Criterion>> {
|
2021-06-15 17:06:42 +08:00
|
|
|
match self.main.get::<_, Str, SerdeJson<Vec<Criterion>>>(rtxn, main_key::CRITERIA_KEY)? {
|
2020-11-27 19:14:56 +08:00
|
|
|
Some(criteria) => Ok(criteria),
|
|
|
|
None => Ok(default_criteria()),
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2020-11-02 18:48:33 +08:00
|
|
|
/* words fst */
|
|
|
|
|
2021-04-07 16:53:57 +08:00
|
|
|
/// Writes the FST which is the words dictionary of the engine.
|
2021-06-17 00:33:33 +08:00
|
|
|
pub(crate) fn put_words_fst<A: AsRef<[u8]>>(
|
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
fst: &fst::Set<A>,
|
|
|
|
) -> heed::Result<()> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.put::<_, Str, ByteSlice>(wtxn, main_key::WORDS_FST_KEY, fst.as_fst().as_bytes())
|
2020-10-21 21:55:48 +08:00
|
|
|
}
|
|
|
|
|
2021-04-07 16:53:57 +08:00
|
|
|
/// Returns the FST which is the words dictionary of the engine.
|
2021-06-14 22:46:19 +08:00
|
|
|
pub fn words_fst<'t>(&self, rtxn: &'t RoTxn) -> Result<fst::Set<Cow<'t, [u8]>>> {
|
2021-06-15 17:06:42 +08:00
|
|
|
match self.main.get::<_, Str, ByteSlice>(rtxn, main_key::WORDS_FST_KEY)? {
|
2020-10-25 21:49:04 +08:00
|
|
|
Some(bytes) => Ok(fst::Set::new(bytes)?.map_data(Cow::Borrowed)?),
|
|
|
|
None => Ok(fst::Set::default().map_data(Cow::Owned)?),
|
2020-10-21 21:55:48 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-03-30 01:15:47 +08:00
|
|
|
/* stop words */
|
|
|
|
|
2021-06-17 00:33:33 +08:00
|
|
|
pub(crate) fn put_stop_words<A: AsRef<[u8]>>(
|
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
fst: &fst::Set<A>,
|
|
|
|
) -> heed::Result<()> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.put::<_, Str, ByteSlice>(wtxn, main_key::STOP_WORDS_KEY, fst.as_fst().as_bytes())
|
2021-03-30 01:15:47 +08:00
|
|
|
}
|
|
|
|
|
2021-06-30 16:07:31 +08:00
|
|
|
pub(crate) fn delete_stop_words(&self, wtxn: &mut RwTxn) -> heed::Result<bool> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.delete::<_, Str>(wtxn, main_key::STOP_WORDS_KEY)
|
2021-03-30 01:15:47 +08:00
|
|
|
}
|
2021-04-07 16:53:57 +08:00
|
|
|
|
2021-06-14 22:46:19 +08:00
|
|
|
pub fn stop_words<'t>(&self, rtxn: &'t RoTxn) -> Result<Option<fst::Set<&'t [u8]>>> {
|
2021-06-15 17:06:42 +08:00
|
|
|
match self.main.get::<_, Str, ByteSlice>(rtxn, main_key::STOP_WORDS_KEY)? {
|
2021-03-30 01:15:47 +08:00
|
|
|
Some(bytes) => Ok(Some(fst::Set::new(bytes)?)),
|
|
|
|
None => Ok(None),
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-04-07 16:53:57 +08:00
|
|
|
/* synonyms */
|
|
|
|
|
2021-06-15 19:45:20 +08:00
|
|
|
pub(crate) fn put_synonyms(
|
2021-06-15 17:06:42 +08:00
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
synonyms: &HashMap<Vec<String>, Vec<Vec<String>>>,
|
2021-06-17 00:33:33 +08:00
|
|
|
) -> heed::Result<()> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.put::<_, Str, SerdeBincode<_>>(wtxn, main_key::SYNONYMS_KEY, synonyms)
|
2021-04-07 16:53:57 +08:00
|
|
|
}
|
|
|
|
|
2021-06-30 16:07:31 +08:00
|
|
|
pub(crate) fn delete_synonyms(&self, wtxn: &mut RwTxn) -> heed::Result<bool> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.delete::<_, Str>(wtxn, main_key::SYNONYMS_KEY)
|
2021-04-07 16:53:57 +08:00
|
|
|
}
|
|
|
|
|
2021-04-10 03:56:20 +08:00
|
|
|
pub fn synonyms(&self, rtxn: &RoTxn) -> heed::Result<HashMap<Vec<String>, Vec<Vec<String>>>> {
|
2021-06-17 00:33:33 +08:00
|
|
|
Ok(self
|
|
|
|
.main
|
|
|
|
.get::<_, Str, SerdeBincode<_>>(rtxn, main_key::SYNONYMS_KEY)?
|
|
|
|
.unwrap_or_default())
|
2021-04-07 16:53:57 +08:00
|
|
|
}
|
|
|
|
|
2021-06-15 17:06:42 +08:00
|
|
|
pub fn words_synonyms<S: AsRef<str>>(
|
|
|
|
&self,
|
|
|
|
rtxn: &RoTxn,
|
|
|
|
words: &[S],
|
2021-06-17 00:33:33 +08:00
|
|
|
) -> heed::Result<Option<Vec<Vec<String>>>> {
|
2021-04-10 03:56:20 +08:00
|
|
|
let words: Vec<_> = words.iter().map(|s| s.as_ref().to_owned()).collect();
|
|
|
|
Ok(self.synonyms(rtxn)?.remove(&words))
|
2021-04-07 16:53:57 +08:00
|
|
|
}
|
|
|
|
|
2021-02-03 17:30:33 +08:00
|
|
|
/* words prefixes fst */
|
|
|
|
|
|
|
|
/// Writes the FST which is the words prefixes dictionnary of the engine.
|
2021-06-17 00:33:33 +08:00
|
|
|
pub(crate) fn put_words_prefixes_fst<A: AsRef<[u8]>>(
|
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
fst: &fst::Set<A>,
|
|
|
|
) -> heed::Result<()> {
|
|
|
|
self.main.put::<_, Str, ByteSlice>(
|
|
|
|
wtxn,
|
|
|
|
main_key::WORDS_PREFIXES_FST_KEY,
|
|
|
|
fst.as_fst().as_bytes(),
|
|
|
|
)
|
2021-02-03 17:30:33 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/// Returns the FST which is the words prefixes dictionnary of the engine.
|
2021-06-14 22:46:19 +08:00
|
|
|
pub fn words_prefixes_fst<'t>(&self, rtxn: &'t RoTxn) -> Result<fst::Set<Cow<'t, [u8]>>> {
|
2021-06-15 17:06:42 +08:00
|
|
|
match self.main.get::<_, Str, ByteSlice>(rtxn, main_key::WORDS_PREFIXES_FST_KEY)? {
|
2021-02-03 17:30:33 +08:00
|
|
|
Some(bytes) => Ok(fst::Set::new(bytes)?.map_data(Cow::Borrowed)?),
|
|
|
|
None => Ok(fst::Set::default().map_data(Cow::Owned)?),
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-02-18 21:35:14 +08:00
|
|
|
/* word documents count */
|
|
|
|
|
|
|
|
/// Returns the number of documents ids associated with the given word,
|
|
|
|
/// it is much faster than deserializing the bitmap and getting the length of it.
|
2021-02-18 21:59:37 +08:00
|
|
|
pub fn word_documents_count(&self, rtxn: &RoTxn, word: &str) -> heed::Result<Option<u64>> {
|
|
|
|
self.word_docids.remap_data_type::<RoaringBitmapLenCodec>().get(rtxn, word)
|
2021-02-18 21:35:14 +08:00
|
|
|
}
|
|
|
|
|
2021-02-03 17:30:33 +08:00
|
|
|
/* documents */
|
|
|
|
|
2020-10-21 21:55:48 +08:00
|
|
|
/// Returns a [`Vec`] of the requested documents. Returns an error if a document is missing.
|
|
|
|
pub fn documents<'t>(
|
|
|
|
&self,
|
2020-11-02 18:48:33 +08:00
|
|
|
rtxn: &'t RoTxn,
|
2021-06-17 00:33:33 +08:00
|
|
|
ids: impl IntoIterator<Item = DocumentId>,
|
2021-07-06 17:31:24 +08:00
|
|
|
) -> Result<Vec<(DocumentId, obkv::KvReaderU16<'t>)>> {
|
2020-10-22 00:26:29 +08:00
|
|
|
let mut documents = Vec::new();
|
2020-10-21 21:55:48 +08:00
|
|
|
|
2020-10-22 00:26:29 +08:00
|
|
|
for id in ids {
|
2021-06-17 00:33:33 +08:00
|
|
|
let kv = self
|
|
|
|
.documents
|
|
|
|
.get(rtxn, &BEU32::new(id))?
|
2021-06-14 22:46:19 +08:00
|
|
|
.ok_or_else(|| UserError::UnknownInternalDocumentId { document_id: id })?;
|
2020-10-22 00:26:29 +08:00
|
|
|
documents.push((id, kv));
|
2020-10-21 21:55:48 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
Ok(documents)
|
|
|
|
}
|
|
|
|
|
2021-05-04 17:23:51 +08:00
|
|
|
/// Returns an iterator over all the documents in the index.
|
|
|
|
pub fn all_documents<'t>(
|
|
|
|
&self,
|
|
|
|
rtxn: &'t RoTxn,
|
2021-07-06 17:31:24 +08:00
|
|
|
) -> Result<impl Iterator<Item = heed::Result<(DocumentId, obkv::KvReaderU16<'t>)>>> {
|
2021-05-04 17:23:51 +08:00
|
|
|
Ok(self
|
|
|
|
.documents
|
|
|
|
.iter(rtxn)?
|
|
|
|
// we cast the BEU32 to a DocumentId
|
|
|
|
.map(|document| document.map(|(id, obkv)| (id.get(), obkv))))
|
|
|
|
}
|
|
|
|
|
2021-01-27 21:15:33 +08:00
|
|
|
pub fn facets_distribution<'a>(&'a self, rtxn: &'a RoTxn) -> FacetDistribution<'a> {
|
2020-12-29 02:08:53 +08:00
|
|
|
FacetDistribution::new(rtxn, self)
|
|
|
|
}
|
|
|
|
|
2020-11-02 18:48:33 +08:00
|
|
|
pub fn search<'a>(&'a self, rtxn: &'a RoTxn) -> Search<'a> {
|
2020-10-21 21:55:48 +08:00
|
|
|
Search::new(rtxn, self)
|
|
|
|
}
|
2021-03-12 01:32:04 +08:00
|
|
|
|
|
|
|
/// Returns the index creation time.
|
2021-06-15 17:51:32 +08:00
|
|
|
pub fn created_at(&self, rtxn: &RoTxn) -> Result<DateTime<Utc>> {
|
2021-06-17 00:33:33 +08:00
|
|
|
Ok(self
|
|
|
|
.main
|
2021-06-15 17:06:42 +08:00
|
|
|
.get::<_, Str, SerdeJson<DateTime<Utc>>>(rtxn, main_key::CREATED_AT_KEY)?
|
2021-06-15 17:51:32 +08:00
|
|
|
.ok_or(InternalError::DatabaseMissingEntry {
|
|
|
|
db_name: db_name::MAIN,
|
|
|
|
key: Some(main_key::CREATED_AT_KEY),
|
|
|
|
})?)
|
2021-03-12 01:32:04 +08:00
|
|
|
}
|
|
|
|
|
2021-03-12 01:42:21 +08:00
|
|
|
/// Returns the index last updated time.
|
2021-06-15 17:51:32 +08:00
|
|
|
pub fn updated_at(&self, rtxn: &RoTxn) -> Result<DateTime<Utc>> {
|
2021-06-17 00:33:33 +08:00
|
|
|
Ok(self
|
|
|
|
.main
|
2021-06-15 17:06:42 +08:00
|
|
|
.get::<_, Str, SerdeJson<DateTime<Utc>>>(rtxn, main_key::UPDATED_AT_KEY)?
|
2021-06-15 17:51:32 +08:00
|
|
|
.ok_or(InternalError::DatabaseMissingEntry {
|
|
|
|
db_name: db_name::MAIN,
|
|
|
|
key: Some(main_key::UPDATED_AT_KEY),
|
|
|
|
})?)
|
2021-03-12 01:32:04 +08:00
|
|
|
}
|
|
|
|
|
2021-06-17 00:33:33 +08:00
|
|
|
pub(crate) fn set_updated_at(
|
|
|
|
&self,
|
|
|
|
wtxn: &mut RwTxn,
|
|
|
|
time: &DateTime<Utc>,
|
|
|
|
) -> heed::Result<()> {
|
2021-06-15 17:06:42 +08:00
|
|
|
self.main.put::<_, Str, SerdeJson<DateTime<Utc>>>(wtxn, main_key::UPDATED_AT_KEY, &time)
|
2021-03-12 01:32:04 +08:00
|
|
|
}
|
2020-10-21 21:55:48 +08:00
|
|
|
}
|
2021-03-31 23:14:23 +08:00
|
|
|
|
|
|
|
#[cfg(test)]
|
2021-04-15 21:29:37 +08:00
|
|
|
pub(crate) mod tests {
|
|
|
|
use std::ops::Deref;
|
|
|
|
|
2021-03-31 23:14:23 +08:00
|
|
|
use heed::EnvOpenOptions;
|
2021-06-17 23:05:34 +08:00
|
|
|
use maplit::btreemap;
|
2021-04-15 21:29:37 +08:00
|
|
|
use tempfile::TempDir;
|
2021-03-31 23:14:23 +08:00
|
|
|
|
|
|
|
use crate::update::{IndexDocuments, UpdateFormat};
|
2021-06-17 00:33:33 +08:00
|
|
|
use crate::Index;
|
2021-03-31 23:14:23 +08:00
|
|
|
|
2021-04-15 21:29:37 +08:00
|
|
|
pub(crate) struct TempIndex {
|
|
|
|
inner: Index,
|
|
|
|
_tempdir: TempDir,
|
|
|
|
}
|
|
|
|
|
|
|
|
impl Deref for TempIndex {
|
|
|
|
type Target = Index;
|
|
|
|
|
|
|
|
fn deref(&self) -> &Self::Target {
|
|
|
|
&self.inner
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
impl TempIndex {
|
|
|
|
/// Creates a temporary index, with a default `4096 * 100` size. This should be enough for
|
|
|
|
/// most tests.
|
|
|
|
pub fn new() -> Self {
|
|
|
|
let mut options = EnvOpenOptions::new();
|
|
|
|
options.map_size(100 * 4096);
|
|
|
|
let _tempdir = TempDir::new_in(".").unwrap();
|
|
|
|
let inner = Index::new(options, _tempdir.path()).unwrap();
|
2021-06-17 00:33:33 +08:00
|
|
|
Self { inner, _tempdir }
|
2021-04-15 21:29:37 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-04-01 15:07:16 +08:00
|
|
|
#[test]
|
2021-06-17 21:16:20 +08:00
|
|
|
fn initial_field_distribution() {
|
2021-03-31 23:14:23 +08:00
|
|
|
let path = tempfile::tempdir().unwrap();
|
|
|
|
let mut options = EnvOpenOptions::new();
|
|
|
|
options.map_size(10 * 1024 * 1024); // 10 MB
|
|
|
|
let index = Index::new(options, &path).unwrap();
|
|
|
|
|
|
|
|
let mut wtxn = index.write_txn().unwrap();
|
2021-04-01 15:07:16 +08:00
|
|
|
let content = &br#"[
|
2021-05-05 03:01:11 +08:00
|
|
|
{ "id": 1, "name": "kevin" },
|
|
|
|
{ "id": 2, "name": "bob", "age": 20 },
|
|
|
|
{ "id": 2, "name": "bob", "age": 20 }
|
2021-04-01 15:07:16 +08:00
|
|
|
]"#[..];
|
2021-03-31 23:14:23 +08:00
|
|
|
let mut builder = IndexDocuments::new(&mut wtxn, &index, 0);
|
2021-04-01 15:07:16 +08:00
|
|
|
builder.update_format(UpdateFormat::Json);
|
2021-03-31 23:14:23 +08:00
|
|
|
builder.execute(content, |_, _| ()).unwrap();
|
|
|
|
wtxn.commit().unwrap();
|
|
|
|
|
|
|
|
let rtxn = index.read_txn().unwrap();
|
|
|
|
|
2021-06-17 21:16:20 +08:00
|
|
|
let field_distribution = index.field_distribution(&rtxn).unwrap();
|
2021-06-17 00:33:33 +08:00
|
|
|
assert_eq!(
|
2021-06-17 21:16:20 +08:00
|
|
|
field_distribution,
|
2021-06-17 23:05:34 +08:00
|
|
|
btreemap! {
|
2021-06-17 00:33:33 +08:00
|
|
|
"id".to_string() => 2,
|
|
|
|
"name".to_string() => 2,
|
|
|
|
"age".to_string() => 1,
|
|
|
|
}
|
|
|
|
);
|
2021-06-17 23:05:34 +08:00
|
|
|
|
|
|
|
// we add all the documents a second time. we are supposed to get the same
|
|
|
|
// field_distribution in the end
|
|
|
|
let mut wtxn = index.write_txn().unwrap();
|
|
|
|
let mut builder = IndexDocuments::new(&mut wtxn, &index, 0);
|
|
|
|
builder.update_format(UpdateFormat::Json);
|
|
|
|
builder.execute(content, |_, _| ()).unwrap();
|
|
|
|
wtxn.commit().unwrap();
|
|
|
|
|
|
|
|
let rtxn = index.read_txn().unwrap();
|
|
|
|
|
|
|
|
let field_distribution = index.field_distribution(&rtxn).unwrap();
|
|
|
|
assert_eq!(
|
|
|
|
field_distribution,
|
|
|
|
btreemap! {
|
|
|
|
"id".to_string() => 2,
|
|
|
|
"name".to_string() => 2,
|
|
|
|
"age".to_string() => 1,
|
|
|
|
}
|
|
|
|
);
|
|
|
|
|
|
|
|
// then we update a document by removing one field and another by adding one field
|
|
|
|
let content = &br#"[
|
|
|
|
{ "id": 1, "name": "kevin", "has_dog": true },
|
|
|
|
{ "id": 2, "name": "bob" }
|
|
|
|
]"#[..];
|
|
|
|
let mut wtxn = index.write_txn().unwrap();
|
|
|
|
let mut builder = IndexDocuments::new(&mut wtxn, &index, 0);
|
|
|
|
builder.update_format(UpdateFormat::Json);
|
|
|
|
builder.execute(content, |_, _| ()).unwrap();
|
|
|
|
wtxn.commit().unwrap();
|
|
|
|
|
|
|
|
let rtxn = index.read_txn().unwrap();
|
|
|
|
|
|
|
|
let field_distribution = index.field_distribution(&rtxn).unwrap();
|
|
|
|
assert_eq!(
|
|
|
|
field_distribution,
|
|
|
|
btreemap! {
|
|
|
|
"id".to_string() => 2,
|
|
|
|
"name".to_string() => 2,
|
|
|
|
"has_dog".to_string() => 1,
|
|
|
|
}
|
|
|
|
);
|
2021-03-31 23:14:23 +08:00
|
|
|
}
|
|
|
|
}
|