2021-04-28 23:58:16 +08:00
use std ::collections ::{ BTreeSet , HashMap , HashSet } ;
2021-06-14 22:46:19 +08:00
use std ::result ::Result as StdResult ;
2020-11-12 00:08:18 +08:00
2021-01-21 00:27:43 +08:00
use itertools ::Itertools ;
2021-04-10 03:56:20 +08:00
use meilisearch_tokenizer ::{ Analyzer , AnalyzerConfig } ;
2021-04-07 19:33:44 +08:00
use serde ::{ Deserialize , Deserializer , Serialize , Serializer } ;
2022-02-15 18:41:55 +08:00
use time ::OffsetDateTime ;
2020-11-02 22:31:20 +08:00
2021-12-08 21:12:07 +08:00
use super ::index_documents ::{ IndexDocumentsConfig , Transform } ;
use super ::IndexerConfig ;
2021-06-08 23:03:27 +08:00
use crate ::criterion ::Criterion ;
2021-06-14 22:46:19 +08:00
use crate ::error ::UserError ;
2021-12-08 21:12:07 +08:00
use crate ::update ::index_documents ::IndexDocumentsMethod ;
2021-06-08 23:03:27 +08:00
use crate ::update ::{ ClearDocuments , IndexDocuments , UpdateIndexingStep } ;
2021-06-14 22:46:19 +08:00
use crate ::{ FieldsIdsMap , Index , Result } ;
2021-04-07 19:33:44 +08:00
2022-04-01 00:44:51 +08:00
#[ derive(Debug, Clone, PartialEq, Copy) ]
2021-04-07 19:33:44 +08:00
pub enum Setting < T > {
Set ( T ) ,
Reset ,
2021-04-07 20:06:14 +08:00
NotSet ,
2021-04-07 19:33:44 +08:00
}
impl < T > Default for Setting < T > {
fn default ( ) -> Self {
Self ::NotSet
}
}
impl < T > Setting < T > {
2021-10-13 19:05:07 +08:00
pub fn set ( self ) -> Option < T > {
match self {
Self ::Set ( value ) = > Some ( value ) ,
_ = > None ,
}
}
pub const fn as_ref ( & self ) -> Setting < & T > {
match * self {
Self ::Set ( ref value ) = > Setting ::Set ( value ) ,
Self ::Reset = > Setting ::Reset ,
Self ::NotSet = > Setting ::NotSet ,
}
}
2021-04-07 19:33:44 +08:00
pub const fn is_not_set ( & self ) -> bool {
matches! ( self , Self ::NotSet )
}
}
impl < T : Serialize > Serialize for Setting < T > {
2021-06-17 00:33:33 +08:00
fn serialize < S > ( & self , serializer : S ) -> StdResult < S ::Ok , S ::Error >
where
S : Serializer ,
{
2021-04-07 19:33:44 +08:00
match self {
Self ::Set ( value ) = > Some ( value ) ,
// Usually not_set isn't serialized by setting skip_serializing_if field attribute
Self ::NotSet | Self ::Reset = > None ,
2021-06-17 00:33:33 +08:00
}
. serialize ( serializer )
2021-04-07 19:33:44 +08:00
}
}
impl < ' de , T : Deserialize < ' de > > Deserialize < ' de > for Setting < T > {
2021-06-17 00:33:33 +08:00
fn deserialize < D > ( deserializer : D ) -> StdResult < Self , D ::Error >
where
D : Deserializer < ' de > ,
{
2021-04-07 19:33:44 +08:00
Deserialize ::deserialize ( deserializer ) . map ( | x | match x {
Some ( x ) = > Self ::Set ( x ) ,
None = > Self ::Reset , // Reset is forced by sending null value
} )
}
}
2020-11-03 20:20:11 +08:00
pub struct Settings < ' a , ' t , ' u , ' i > {
2020-11-02 22:31:20 +08:00
wtxn : & ' t mut heed ::RwTxn < ' i , ' u > ,
index : & ' i Index ,
2021-12-08 21:12:07 +08:00
indexer_config : & ' a IndexerConfig ,
2020-11-03 20:20:11 +08:00
2021-04-07 19:33:44 +08:00
searchable_fields : Setting < Vec < String > > ,
displayed_fields : Setting < Vec < String > > ,
2021-06-01 18:19:55 +08:00
filterable_fields : Setting < HashSet < String > > ,
2021-08-23 17:37:18 +08:00
sortable_fields : Setting < HashSet < String > > ,
2021-04-07 19:33:44 +08:00
criteria : Setting < Vec < String > > ,
stop_words : Setting < BTreeSet < String > > ,
2021-06-01 22:29:14 +08:00
distinct_field : Setting < String > ,
2021-04-07 16:53:57 +08:00
synonyms : Setting < HashMap < String , Vec < String > > > ,
2021-06-15 19:45:20 +08:00
primary_key : Setting < String > ,
2022-03-16 17:03:18 +08:00
authorize_typos : Setting < bool > ,
2022-04-01 00:23:12 +08:00
min_word_len_two_typos : Setting < u8 > ,
min_word_len_one_typo : Setting < u8 > ,
2020-11-02 22:31:20 +08:00
}
2020-11-03 20:20:11 +08:00
impl < ' a , ' t , ' u , ' i > Settings < ' a , ' t , ' u , ' i > {
2021-12-08 21:12:07 +08:00
pub fn new (
wtxn : & ' t mut heed ::RwTxn < ' i , ' u > ,
index : & ' i Index ,
indexer_config : & ' a IndexerConfig ,
) -> Settings < ' a , ' t , ' u , ' i > {
2020-11-03 20:20:11 +08:00
Settings {
wtxn ,
index ,
2021-04-07 19:33:44 +08:00
searchable_fields : Setting ::NotSet ,
displayed_fields : Setting ::NotSet ,
2021-06-01 18:19:55 +08:00
filterable_fields : Setting ::NotSet ,
2021-08-23 17:37:18 +08:00
sortable_fields : Setting ::NotSet ,
2021-04-07 19:33:44 +08:00
criteria : Setting ::NotSet ,
stop_words : Setting ::NotSet ,
2021-06-01 22:29:14 +08:00
distinct_field : Setting ::NotSet ,
2021-04-07 16:53:57 +08:00
synonyms : Setting ::NotSet ,
2021-06-15 19:45:20 +08:00
primary_key : Setting ::NotSet ,
2022-03-16 17:03:18 +08:00
authorize_typos : Setting ::NotSet ,
2021-12-08 21:12:07 +08:00
indexer_config ,
2022-04-01 00:23:12 +08:00
min_word_len_two_typos : Setting ::Reset ,
min_word_len_one_typo : Setting ::Reset ,
2020-11-03 20:20:11 +08:00
}
2020-11-02 22:31:20 +08:00
}
2020-11-03 20:42:29 +08:00
pub fn reset_searchable_fields ( & mut self ) {
2021-04-07 19:33:44 +08:00
self . searchable_fields = Setting ::Reset ;
2020-11-03 20:42:29 +08:00
}
pub fn set_searchable_fields ( & mut self , names : Vec < String > ) {
2021-04-07 19:33:44 +08:00
self . searchable_fields = Setting ::Set ( names ) ;
2020-11-03 20:42:29 +08:00
}
2020-11-02 22:31:20 +08:00
pub fn reset_displayed_fields ( & mut self ) {
2021-04-07 19:33:44 +08:00
self . displayed_fields = Setting ::Reset ;
2020-11-02 22:31:20 +08:00
}
pub fn set_displayed_fields ( & mut self , names : Vec < String > ) {
2021-04-07 19:33:44 +08:00
self . displayed_fields = Setting ::Set ( names ) ;
2020-11-02 22:31:20 +08:00
}
2021-06-01 18:19:55 +08:00
pub fn reset_filterable_fields ( & mut self ) {
self . filterable_fields = Setting ::Reset ;
2021-01-21 00:27:43 +08:00
}
2021-06-01 18:19:55 +08:00
pub fn set_filterable_fields ( & mut self , names : HashSet < String > ) {
self . filterable_fields = Setting ::Set ( names ) ;
2020-11-12 00:08:18 +08:00
}
2021-08-23 17:37:18 +08:00
pub fn set_sortable_fields ( & mut self , names : HashSet < String > ) {
self . sortable_fields = Setting ::Set ( names ) ;
}
2021-08-25 23:44:16 +08:00
pub fn reset_sortable_fields ( & mut self ) {
self . sortable_fields = Setting ::Reset ;
}
2020-12-04 19:02:22 +08:00
pub fn reset_criteria ( & mut self ) {
2021-04-07 19:33:44 +08:00
self . criteria = Setting ::Reset ;
2020-12-04 19:02:22 +08:00
}
pub fn set_criteria ( & mut self , criteria : Vec < String > ) {
2021-04-07 19:33:44 +08:00
self . criteria = Setting ::Set ( criteria ) ;
2020-12-04 19:02:22 +08:00
}
2021-03-30 01:15:47 +08:00
pub fn reset_stop_words ( & mut self ) {
2021-04-07 19:33:44 +08:00
self . stop_words = Setting ::Reset ;
2021-03-30 01:15:47 +08:00
}
pub fn set_stop_words ( & mut self , stop_words : BTreeSet < String > ) {
2021-06-17 00:33:33 +08:00
self . stop_words =
if stop_words . is_empty ( ) { Setting ::Reset } else { Setting ::Set ( stop_words ) }
2021-03-30 01:15:47 +08:00
}
2021-06-01 22:29:14 +08:00
pub fn reset_distinct_field ( & mut self ) {
self . distinct_field = Setting ::Reset ;
2021-04-07 16:53:57 +08:00
}
2021-06-01 22:29:14 +08:00
pub fn set_distinct_field ( & mut self , distinct_field : String ) {
self . distinct_field = Setting ::Set ( distinct_field ) ;
2021-04-07 18:38:48 +08:00
}
2021-04-07 16:53:57 +08:00
pub fn reset_synonyms ( & mut self ) {
self . synonyms = Setting ::Reset ;
}
pub fn set_synonyms ( & mut self , synonyms : HashMap < String , Vec < String > > ) {
2021-06-17 00:33:33 +08:00
self . synonyms = if synonyms . is_empty ( ) { Setting ::Reset } else { Setting ::Set ( synonyms ) }
2021-04-07 18:38:48 +08:00
}
2021-06-15 19:45:20 +08:00
pub fn reset_primary_key ( & mut self ) {
self . primary_key = Setting ::Reset ;
}
pub fn set_primary_key ( & mut self , primary_key : String ) {
self . primary_key = Setting ::Set ( primary_key ) ;
}
2022-03-16 17:03:18 +08:00
pub fn set_autorize_typos ( & mut self , val : bool ) {
self . authorize_typos = Setting ::Set ( val ) ;
}
pub fn reset_authorize_typos ( & mut self ) {
self . authorize_typos = Setting ::Reset ;
}
2022-04-01 00:23:12 +08:00
pub fn set_min_word_len_two_typos ( & mut self , val : u8 ) {
self . min_word_len_two_typos = Setting ::Set ( val ) ;
2022-03-21 20:03:06 +08:00
}
2022-04-01 00:23:12 +08:00
pub fn reset_min_word_len_two_typos ( & mut self ) {
self . min_word_len_two_typos = Setting ::Reset ;
2022-03-21 20:03:06 +08:00
}
2022-04-01 00:23:12 +08:00
pub fn set_min_word_len_one_typo ( & mut self , val : u8 ) {
self . min_word_len_one_typo = Setting ::Set ( val ) ;
2022-03-21 20:03:06 +08:00
}
2022-04-01 00:23:12 +08:00
pub fn reset_min_word_len_one_typo ( & mut self ) {
self . min_word_len_one_typo = Setting ::Reset ;
2022-03-21 20:03:06 +08:00
}
2021-06-14 22:46:19 +08:00
fn reindex < F > ( & mut self , cb : & F , old_fields_ids_map : FieldsIdsMap ) -> Result < ( ) >
2021-06-01 22:29:14 +08:00
where
2021-11-03 20:12:01 +08:00
F : Fn ( UpdateIndexingStep ) + Sync ,
2020-11-03 20:20:11 +08:00
{
2020-11-14 05:35:02 +08:00
let fields_ids_map = self . index . fields_ids_map ( self . wtxn ) ? ;
2021-01-21 00:27:43 +08:00
// if the settings are set before any document update, we don't need to do anything, and
// will set the primary key during the first document addition.
if self . index . number_of_documents ( & self . wtxn ) ? = = 0 {
2021-04-07 19:33:44 +08:00
return Ok ( ( ) ) ;
2021-01-21 00:27:43 +08:00
}
2020-11-14 05:35:02 +08:00
2021-12-08 21:12:07 +08:00
let transform = Transform ::new (
& self . index ,
& self . indexer_config ,
IndexDocumentsMethod ::ReplaceDocuments ,
false ,
) ;
2021-01-21 00:27:43 +08:00
// We remap the documents fields based on the new `FieldsIdsMap`.
let output = transform . remap_index_documents (
2021-12-08 21:12:07 +08:00
self . wtxn ,
2021-01-21 00:27:43 +08:00
old_fields_ids_map ,
2021-06-01 22:29:14 +08:00
fields_ids_map . clone ( ) ,
) ? ;
2021-01-21 00:27:43 +08:00
// We clear the full database (words-fst, documents ids and documents content).
2021-11-03 20:12:01 +08:00
ClearDocuments ::new ( self . wtxn , self . index ) . execute ( ) ? ;
2021-01-21 00:27:43 +08:00
// We index the generated `TransformOutput` which must contain
// all the documents with fields in the newly defined searchable order.
2021-12-08 21:12:07 +08:00
let indexing_builder = IndexDocuments ::new (
self . wtxn ,
self . index ,
& self . indexer_config ,
IndexDocumentsConfig ::default ( ) ,
& cb ,
) ;
indexing_builder . execute_raw ( output ) ? ;
2021-06-01 22:29:14 +08:00
2021-01-21 00:27:43 +08:00
Ok ( ( ) )
}
2020-11-12 00:08:18 +08:00
2021-06-14 22:46:19 +08:00
fn update_displayed ( & mut self ) -> Result < bool > {
2021-01-21 00:27:43 +08:00
match self . displayed_fields {
2021-04-07 19:33:44 +08:00
Setting ::Set ( ref fields ) = > {
2021-01-21 00:27:43 +08:00
// fields are deduplicated, only the first occurrence is taken into account
2021-06-17 00:33:33 +08:00
let names : Vec < _ > = fields . iter ( ) . unique ( ) . map ( String ::as_str ) . collect ( ) ;
2021-01-21 00:27:43 +08:00
self . index . put_displayed_fields ( self . wtxn , & names ) ? ;
2020-11-14 05:35:02 +08:00
}
2021-06-17 00:33:33 +08:00
Setting ::Reset = > {
self . index . delete_displayed_fields ( self . wtxn ) ? ;
}
2021-04-07 19:33:44 +08:00
Setting ::NotSet = > return Ok ( false ) ,
2020-11-14 05:35:02 +08:00
}
2021-01-21 00:27:43 +08:00
Ok ( true )
}
2020-11-03 20:20:11 +08:00
2021-06-14 22:46:19 +08:00
fn update_distinct_field ( & mut self ) -> Result < bool > {
2021-06-01 22:29:14 +08:00
match self . distinct_field {
2021-04-07 18:38:48 +08:00
Setting ::Set ( ref attr ) = > {
2021-06-01 22:29:14 +08:00
self . index . put_distinct_field ( self . wtxn , & attr ) ? ;
2021-04-07 18:38:48 +08:00
}
2021-06-17 00:33:33 +08:00
Setting ::Reset = > {
self . index . delete_distinct_field ( self . wtxn ) ? ;
}
2021-04-07 18:38:48 +08:00
Setting ::NotSet = > return Ok ( false ) ,
}
Ok ( true )
}
2021-03-31 23:14:23 +08:00
/// Updates the index's searchable attributes. This causes the field map to be recomputed to
2021-01-21 00:27:43 +08:00
/// reflect the order of the searchable attributes.
2021-06-14 22:46:19 +08:00
fn update_searchable ( & mut self ) -> Result < bool > {
2021-01-21 00:27:43 +08:00
match self . searchable_fields {
2021-04-07 19:33:44 +08:00
Setting ::Set ( ref fields ) = > {
2021-01-21 00:27:43 +08:00
// every time the searchable attributes are updated, we need to update the
2021-07-15 16:19:35 +08:00
// ids for any settings that uses the facets. (distinct_fields, filterable_fields).
2021-01-21 00:27:43 +08:00
let old_fields_ids_map = self . index . fields_ids_map ( self . wtxn ) ? ;
2020-12-04 19:02:22 +08:00
2021-01-21 00:27:43 +08:00
let mut new_fields_ids_map = FieldsIdsMap ::new ( ) ;
// fields are deduplicated, only the first occurrence is taken into account
2021-06-17 00:33:33 +08:00
let names = fields . iter ( ) . unique ( ) . map ( String ::as_str ) . collect ::< Vec < _ > > ( ) ;
2021-01-21 00:27:43 +08:00
// Add all the searchable attributes to the field map, and then add the
// remaining fields from the old field map to the new one
for name in names . iter ( ) {
2021-06-17 00:33:33 +08:00
new_fields_ids_map . insert ( & name ) . ok_or ( UserError ::AttributeLimitReached ) ? ;
2021-01-21 00:27:43 +08:00
}
2020-11-03 20:20:11 +08:00
2021-01-21 00:27:43 +08:00
for ( _ , name ) in old_fields_ids_map . iter ( ) {
2021-06-17 00:33:33 +08:00
new_fields_ids_map . insert ( & name ) . ok_or ( UserError ::AttributeLimitReached ) ? ;
2020-11-14 05:35:02 +08:00
}
2020-11-03 20:20:11 +08:00
2021-01-21 00:27:43 +08:00
self . index . put_searchable_fields ( self . wtxn , & names ) ? ;
self . index . put_fields_ids_map ( self . wtxn , & new_fields_ids_map ) ? ;
}
2021-06-17 00:33:33 +08:00
Setting ::Reset = > {
self . index . delete_searchable_fields ( self . wtxn ) ? ;
}
2021-04-07 19:33:44 +08:00
Setting ::NotSet = > return Ok ( false ) ,
2020-11-03 20:20:11 +08:00
}
2021-01-21 00:27:43 +08:00
Ok ( true )
}
2020-11-03 20:20:11 +08:00
2021-06-14 22:46:19 +08:00
fn update_stop_words ( & mut self ) -> Result < bool > {
2021-03-30 01:15:47 +08:00
match self . stop_words {
2021-04-07 19:33:44 +08:00
Setting ::Set ( ref stop_words ) = > {
2021-03-30 01:15:47 +08:00
let current = self . index . stop_words ( self . wtxn ) ? ;
// since we can't compare a BTreeSet with an FST we are going to convert the
// BTreeSet to an FST and then compare bytes per bytes the two FSTs.
2021-04-07 16:53:57 +08:00
let fst = fst ::Set ::from_iter ( stop_words ) ? ;
2021-03-30 01:15:47 +08:00
// Does the new FST differ from the previous one?
2021-06-17 00:33:33 +08:00
if current
. map_or ( true , | current | current . as_fst ( ) . as_bytes ( ) ! = fst . as_fst ( ) . as_bytes ( ) )
{
2021-03-30 01:15:47 +08:00
// we want to re-create our FST.
self . index . put_stop_words ( self . wtxn , & fst ) ? ;
Ok ( true )
} else {
Ok ( false )
}
}
2021-04-07 19:33:44 +08:00
Setting ::Reset = > Ok ( self . index . delete_stop_words ( self . wtxn ) ? ) ,
Setting ::NotSet = > Ok ( false ) ,
2021-03-30 01:15:47 +08:00
}
}
2021-06-14 22:46:19 +08:00
fn update_synonyms ( & mut self ) -> Result < bool > {
2021-04-07 16:53:57 +08:00
match self . synonyms {
Setting ::Set ( ref synonyms ) = > {
2021-04-10 03:56:20 +08:00
fn normalize ( analyzer : & Analyzer < & [ u8 ] > , text : & str ) -> Vec < String > {
analyzer
. analyze ( text )
. tokens ( )
2021-06-17 00:33:33 +08:00
. filter_map ( | token | {
if token . is_word ( ) {
Some ( token . text ( ) . to_string ( ) )
} else {
None
}
} )
2021-04-10 03:56:20 +08:00
. collect ::< Vec < _ > > ( )
}
2021-04-07 16:53:57 +08:00
let mut config = AnalyzerConfig ::default ( ) ;
let stop_words = self . index . stop_words ( self . wtxn ) ? ;
if let Some ( stop_words ) = & stop_words {
config . stop_words ( stop_words ) ;
}
let analyzer = Analyzer ::new ( config ) ;
2021-04-10 03:56:20 +08:00
let mut new_synonyms = HashMap ::new ( ) ;
for ( word , synonyms ) in synonyms {
// Normalize both the word and associated synonyms.
let normalized_word = normalize ( & analyzer , word ) ;
2021-06-17 00:33:33 +08:00
let normalized_synonyms =
synonyms . iter ( ) . map ( | synonym | normalize ( & analyzer , synonym ) ) ;
2021-04-10 03:56:20 +08:00
// Store the normalized synonyms under the normalized word,
// merging the possible duplicate words.
2021-06-17 00:33:33 +08:00
let entry = new_synonyms . entry ( normalized_word ) . or_insert_with ( Vec ::new ) ;
2021-04-10 03:56:20 +08:00
entry . extend ( normalized_synonyms ) ;
}
2021-04-07 16:53:57 +08:00
2021-04-10 03:56:20 +08:00
// Make sure that we don't have duplicate synonyms.
2021-06-17 00:33:33 +08:00
new_synonyms . iter_mut ( ) . for_each ( | ( _ , synonyms ) | {
synonyms . sort_unstable ( ) ;
synonyms . dedup ( ) ;
} ) ;
2021-04-10 03:56:20 +08:00
let old_synonyms = self . index . synonyms ( self . wtxn ) ? ;
2021-04-07 16:53:57 +08:00
if new_synonyms ! = old_synonyms {
self . index . put_synonyms ( self . wtxn , & new_synonyms ) ? ;
Ok ( true )
} else {
Ok ( false )
}
}
Setting ::Reset = > Ok ( self . index . delete_synonyms ( self . wtxn ) ? ) ,
Setting ::NotSet = > Ok ( false ) ,
}
}
2021-06-14 22:46:19 +08:00
fn update_filterable ( & mut self ) -> Result < ( ) > {
2021-06-01 18:19:55 +08:00
match self . filterable_fields {
2021-04-07 19:33:44 +08:00
Setting ::Set ( ref fields ) = > {
2021-04-28 23:58:16 +08:00
let mut new_facets = HashSet ::new ( ) ;
for name in fields {
new_facets . insert ( name . clone ( ) ) ;
2021-01-21 00:27:43 +08:00
}
2021-06-01 18:19:55 +08:00
self . index . put_filterable_fields ( self . wtxn , & new_facets ) ? ;
2020-11-02 22:31:20 +08:00
}
2021-06-17 00:33:33 +08:00
Setting ::Reset = > {
self . index . delete_filterable_fields ( self . wtxn ) ? ;
}
2021-06-01 22:29:14 +08:00
Setting ::NotSet = > ( ) ,
2020-11-02 22:31:20 +08:00
}
2021-06-01 22:29:14 +08:00
Ok ( ( ) )
2021-01-21 00:27:43 +08:00
}
2020-11-02 22:31:20 +08:00
2021-08-23 17:37:18 +08:00
fn update_sortable ( & mut self ) -> Result < ( ) > {
match self . sortable_fields {
Setting ::Set ( ref fields ) = > {
let mut new_fields = HashSet ::new ( ) ;
for name in fields {
new_fields . insert ( name . clone ( ) ) ;
}
self . index . put_sortable_fields ( self . wtxn , & new_fields ) ? ;
}
Setting ::Reset = > {
self . index . delete_sortable_fields ( self . wtxn ) ? ;
}
Setting ::NotSet = > ( ) ,
}
Ok ( ( ) )
}
2021-06-14 22:46:19 +08:00
fn update_criteria ( & mut self ) -> Result < ( ) > {
2021-01-21 00:27:43 +08:00
match self . criteria {
2021-04-07 19:33:44 +08:00
Setting ::Set ( ref fields ) = > {
2021-01-21 00:27:43 +08:00
let mut new_criteria = Vec ::new ( ) ;
for name in fields {
2021-06-08 23:03:27 +08:00
let criterion : Criterion = name . parse ( ) ? ;
2021-01-21 00:27:43 +08:00
new_criteria . push ( criterion ) ;
}
self . index . put_criteria ( self . wtxn , & new_criteria ) ? ;
2020-12-04 19:02:22 +08:00
}
2021-06-17 00:33:33 +08:00
Setting ::Reset = > {
self . index . delete_criteria ( self . wtxn ) ? ;
}
2021-04-07 19:33:44 +08:00
Setting ::NotSet = > ( ) ,
2020-12-04 19:02:22 +08:00
}
2020-11-02 22:31:20 +08:00
Ok ( ( ) )
}
2021-01-21 00:27:43 +08:00
2021-06-15 19:45:20 +08:00
fn update_primary_key ( & mut self ) -> Result < ( ) > {
match self . primary_key {
Setting ::Set ( ref primary_key ) = > {
if self . index . number_of_documents ( & self . wtxn ) ? = = 0 {
let mut fields_ids_map = self . index . fields_ids_map ( self . wtxn ) ? ;
fields_ids_map . insert ( primary_key ) . ok_or ( UserError ::AttributeLimitReached ) ? ;
self . index . put_fields_ids_map ( self . wtxn , & fields_ids_map ) ? ;
self . index . put_primary_key ( self . wtxn , primary_key ) ? ;
Ok ( ( ) )
} else {
2021-10-26 23:49:35 +08:00
let primary_key = self . index . primary_key ( self . wtxn ) ? . unwrap ( ) ;
Err ( UserError ::PrimaryKeyCannotBeChanged ( primary_key . to_string ( ) ) . into ( ) )
2021-06-15 19:45:20 +08:00
}
2021-06-17 00:33:33 +08:00
}
2021-06-15 19:45:20 +08:00
Setting ::Reset = > {
if self . index . number_of_documents ( & self . wtxn ) ? = = 0 {
self . index . delete_primary_key ( self . wtxn ) ? ;
Ok ( ( ) )
} else {
2021-10-26 23:49:35 +08:00
let primary_key = self . index . primary_key ( self . wtxn ) ? . unwrap ( ) ;
Err ( UserError ::PrimaryKeyCannotBeChanged ( primary_key . to_string ( ) ) . into ( ) )
2021-06-15 19:45:20 +08:00
}
2021-06-17 00:33:33 +08:00
}
2021-06-15 19:45:20 +08:00
Setting ::NotSet = > Ok ( ( ) ) ,
}
}
2022-03-16 17:03:18 +08:00
fn update_authorize_typos ( & mut self ) -> Result < ( ) > {
match self . authorize_typos {
Setting ::Set ( flag ) = > {
self . index . put_authorize_typos ( self . wtxn , flag ) ? ;
Ok ( ( ) )
}
Setting ::Reset = > {
self . index . put_authorize_typos ( self . wtxn , true ) ? ;
Ok ( ( ) )
}
Setting ::NotSet = > Ok ( ( ) ) ,
}
}
2022-03-21 20:03:06 +08:00
fn update_min_typo_word_len ( & mut self ) -> Result < ( ) > {
2022-04-01 00:44:51 +08:00
match ( self . min_word_len_one_typo , self . min_word_len_two_typos ) {
2022-03-21 20:03:06 +08:00
( Setting ::Set ( one ) , Setting ::Set ( two ) ) = > {
2022-03-31 19:50:18 +08:00
if one > two {
2022-04-01 00:44:51 +08:00
return Err ( UserError ::InvalidMinTypoWordLenSetting ( one , two ) . into ( ) ) ;
2022-03-31 19:50:18 +08:00
} else {
2022-04-01 00:44:51 +08:00
self . index . put_min_word_len_one_typo ( & mut self . wtxn , one ) ? ;
self . index . put_min_word_len_two_typos ( & mut self . wtxn , two ) ? ;
2022-03-21 20:03:06 +08:00
}
}
( Setting ::Set ( one ) , _ ) = > {
2022-04-01 00:23:12 +08:00
let two = self . index . min_word_len_two_typos ( & self . wtxn ) ? ;
2022-04-01 00:44:51 +08:00
if one > two {
return Err ( UserError ::InvalidMinTypoWordLenSetting ( one , two ) . into ( ) ) ;
2022-03-31 19:50:18 +08:00
} else {
2022-04-01 00:44:51 +08:00
self . index . put_min_word_len_one_typo ( & mut self . wtxn , one ) ? ;
2022-03-21 20:03:06 +08:00
}
}
( _ , Setting ::Set ( two ) ) = > {
2022-04-01 00:23:12 +08:00
let one = self . index . min_word_len_one_typo ( & self . wtxn ) ? ;
2022-04-01 00:44:51 +08:00
if one > two {
return Err ( UserError ::InvalidMinTypoWordLenSetting ( one , two ) . into ( ) ) ;
2022-03-31 19:50:18 +08:00
} else {
2022-04-01 00:44:51 +08:00
self . index . put_min_word_len_two_typos ( & mut self . wtxn , two ) ? ;
2022-03-21 20:03:06 +08:00
}
}
_ = > ( ) ,
}
Ok ( ( ) )
}
2021-06-14 22:46:19 +08:00
pub fn execute < F > ( mut self , progress_callback : F ) -> Result < ( ) >
2021-06-17 00:33:33 +08:00
where
2021-11-03 20:12:01 +08:00
F : Fn ( UpdateIndexingStep ) + Sync ,
2021-03-30 01:15:47 +08:00
{
2022-02-15 18:41:55 +08:00
self . index . set_updated_at ( self . wtxn , & OffsetDateTime ::now_utc ( ) ) ? ;
2021-06-01 22:29:14 +08:00
let old_faceted_fields = self . index . faceted_fields ( & self . wtxn ) ? ;
2021-03-30 01:15:47 +08:00
let old_fields_ids_map = self . index . fields_ids_map ( & self . wtxn ) ? ;
2021-06-01 22:29:14 +08:00
2021-03-30 01:15:47 +08:00
self . update_displayed ( ) ? ;
2021-06-01 22:29:14 +08:00
self . update_filterable ( ) ? ;
2021-08-23 17:37:18 +08:00
self . update_sortable ( ) ? ;
2021-06-01 22:29:14 +08:00
self . update_distinct_field ( ) ? ;
2021-03-30 01:15:47 +08:00
self . update_criteria ( ) ? ;
2021-06-15 19:45:20 +08:00
self . update_primary_key ( ) ? ;
2022-03-16 17:03:18 +08:00
self . update_authorize_typos ( ) ? ;
2022-03-21 20:03:06 +08:00
self . update_min_typo_word_len ( ) ? ;
2021-06-01 22:29:14 +08:00
// If there is new faceted fields we indicate that we must reindex as we must
// index new fields as facets. It means that the distinct attribute,
// an Asc/Desc criterion or a filtered attribute as be added or removed.
let new_faceted_fields = self . index . faceted_fields ( & self . wtxn ) ? ;
let faceted_updated = old_faceted_fields ! = new_faceted_fields ;
let stop_words_updated = self . update_stop_words ( ) ? ;
2021-04-07 16:53:57 +08:00
let synonyms_updated = self . update_synonyms ( ) ? ;
2021-03-30 01:15:47 +08:00
let searchable_updated = self . update_searchable ( ) ? ;
2021-06-01 22:29:14 +08:00
if stop_words_updated | | faceted_updated | | synonyms_updated | | searchable_updated {
2021-03-30 01:15:47 +08:00
self . reindex ( & progress_callback , old_fields_ids_map ) ? ;
2021-01-21 00:27:43 +08:00
}
2021-06-01 22:29:14 +08:00
2021-03-30 01:15:47 +08:00
Ok ( ( ) )
}
2020-11-02 22:31:20 +08:00
}
#[ cfg(test) ]
mod tests {
2021-06-17 00:33:33 +08:00
use big_s ::S ;
2021-05-03 21:58:47 +08:00
use heed ::types ::ByteSlice ;
2021-06-17 00:33:33 +08:00
use heed ::EnvOpenOptions ;
2021-05-03 21:58:47 +08:00
use maplit ::{ btreeset , hashmap , hashset } ;
2020-11-02 22:31:20 +08:00
2021-06-17 00:33:33 +08:00
use super ::* ;
2021-06-15 19:45:20 +08:00
use crate ::error ::Error ;
2022-03-31 20:12:00 +08:00
use crate ::index ::tests ::TempIndex ;
2021-08-31 17:44:15 +08:00
use crate ::update ::IndexDocuments ;
2021-11-06 08:32:12 +08:00
use crate ::{ Criterion , Filter , SearchResult } ;
2021-01-21 00:27:43 +08:00
2020-11-04 01:22:40 +08:00
#[ test ]
fn set_and_reset_searchable_fields ( ) {
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 ( ) ;
// First we send 3 documents with ids from 1 to 3.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-08-31 17:44:15 +08:00
let content = documents! ( [
{ " id " : 1 , " name " : " kevin " , " age " : 23 } ,
{ " id " : 2 , " name " : " kevina " , " age " : 21 } ,
{ " id " : 3 , " name " : " benoit " , " age " : 34 }
] ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
let indexing_config = IndexDocumentsConfig ::default ( ) ;
let mut builder =
IndexDocuments ::new ( & mut wtxn , & index , & config , indexing_config . clone ( ) , | _ | ( ) ) ;
builder . add_documents ( content ) . unwrap ( ) ;
builder . execute ( ) . unwrap ( ) ;
2020-11-04 01:22:40 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// We change the searchable fields to be the "name" field only.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2020-11-04 01:22:40 +08:00
builder . set_searchable_fields ( vec! [ " name " . into ( ) ] ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2020-11-04 01:22:40 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Check that the searchable field is correctly set to "name" only.
let rtxn = index . read_txn ( ) . unwrap ( ) ;
// When we search for something that is not in
// the searchable fields it must not return any document.
let result = index . search ( & rtxn ) . query ( " 23 " ) . execute ( ) . unwrap ( ) ;
assert! ( result . documents_ids . is_empty ( ) ) ;
// When we search for something that is in the searchable fields
// we must find the appropriate document.
let result = index . search ( & rtxn ) . query ( r # ""kevin""# ) . execute ( ) . unwrap ( ) ;
let documents = index . documents ( & rtxn , result . documents_ids ) . unwrap ( ) ;
assert_eq! ( documents . len ( ) , 1 ) ;
assert_eq! ( documents [ 0 ] . 1. get ( 0 ) , Some ( & br # ""kevin""# [ .. ] ) ) ;
drop ( rtxn ) ;
// We change the searchable fields to be the "name" field only.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2020-11-04 01:22:40 +08:00
builder . reset_searchable_fields ( ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2020-11-04 01:22:40 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Check that the searchable field have been reset and documents are found now.
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let searchable_fields = index . searchable_fields ( & rtxn ) . unwrap ( ) ;
assert_eq! ( searchable_fields , None ) ;
let result = index . search ( & rtxn ) . query ( " 23 " ) . execute ( ) . unwrap ( ) ;
assert_eq! ( result . documents_ids . len ( ) , 1 ) ;
let documents = index . documents ( & rtxn , result . documents_ids ) . unwrap ( ) ;
assert_eq! ( documents [ 0 ] . 1. get ( 0 ) , Some ( & br # ""kevin""# [ .. ] ) ) ;
}
#[ test ]
fn mixup_searchable_with_displayed_fields ( ) {
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 ( ) ;
// First we send 3 documents with ids from 1 to 3.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-08-31 17:44:15 +08:00
let content = documents! ( [
{ " name " : " kevin " , " age " : 23 } ,
{ " name " : " kevina " , " age " : 21 } ,
{ " name " : " benoit " , " age " : 34 }
] ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
let indexing_config =
IndexDocumentsConfig { autogenerate_docids : true , .. Default ::default ( ) } ;
let mut builder =
IndexDocuments ::new ( & mut wtxn , & index , & config , indexing_config . clone ( ) , | _ | ( ) ) ;
builder . add_documents ( content ) . unwrap ( ) ;
builder . execute ( ) . unwrap ( ) ;
2020-11-04 01:22:40 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// In the same transaction we change the displayed fields to be only the "age".
// We also change the searchable fields to be the "name" field only.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2020-11-04 01:22:40 +08:00
builder . set_displayed_fields ( vec! [ " age " . into ( ) ] ) ;
builder . set_searchable_fields ( vec! [ " name " . into ( ) ] ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2020-11-04 01:22:40 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Check that the displayed fields are correctly set to `None` (default value).
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let fields_ids = index . displayed_fields ( & rtxn ) . unwrap ( ) ;
2021-01-21 00:27:43 +08:00
assert_eq! ( fields_ids . unwrap ( ) , ( & [ " age " ] [ .. ] ) ) ;
2020-11-04 01:22:40 +08:00
drop ( rtxn ) ;
// We change the searchable fields to be the "name" field only.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2020-11-04 01:22:40 +08:00
builder . reset_searchable_fields ( ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2020-11-04 01:22:40 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Check that the displayed fields always contains only the "age" field.
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let fields_ids = index . displayed_fields ( & rtxn ) . unwrap ( ) ;
2021-01-21 00:27:43 +08:00
assert_eq! ( fields_ids . unwrap ( ) , & [ " age " ] [ .. ] ) ;
2020-11-04 01:22:40 +08:00
}
2020-11-02 22:31:20 +08:00
#[ test ]
fn default_displayed_fields ( ) {
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 ( ) ;
// First we send 3 documents with ids from 1 to 3.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-08-31 17:44:15 +08:00
let content = documents! ( [
{ " name " : " kevin " , " age " : 23 } ,
{ " name " : " kevina " , " age " : 21 } ,
{ " name " : " benoit " , " age " : 34 }
] ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
let indexing_config =
IndexDocumentsConfig { autogenerate_docids : true , .. Default ::default ( ) } ;
let mut builder =
IndexDocuments ::new ( & mut wtxn , & index , & config , indexing_config . clone ( ) , | _ | ( ) ) ;
builder . add_documents ( content ) . unwrap ( ) ;
builder . execute ( ) . unwrap ( ) ;
2020-11-02 22:31:20 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Check that the displayed fields are correctly set to `None` (default value).
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let fields_ids = index . displayed_fields ( & rtxn ) . unwrap ( ) ;
assert_eq! ( fields_ids , None ) ;
}
#[ test ]
fn set_and_reset_displayed_field ( ) {
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 ( ) ;
// First we send 3 documents with ids from 1 to 3.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-08-31 17:44:15 +08:00
let content = documents! ( [
{ " name " : " kevin " , " age " : 23 } ,
{ " name " : " kevina " , " age " : 21 } ,
{ " name " : " benoit " , " age " : 34 }
] ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
let indexing_config =
IndexDocumentsConfig { autogenerate_docids : true , .. Default ::default ( ) } ;
let mut builder =
IndexDocuments ::new ( & mut wtxn , & index , & config , indexing_config . clone ( ) , | _ | ( ) ) ;
builder . add_documents ( content ) . unwrap ( ) ;
builder . execute ( ) . unwrap ( ) ;
2020-11-02 22:31:20 +08:00
// In the same transaction we change the displayed fields to be only the age.
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2020-11-02 22:31:20 +08:00
builder . set_displayed_fields ( vec! [ " age " . into ( ) ] ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2020-11-02 22:31:20 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Check that the displayed fields are correctly set to only the "age" field.
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let fields_ids = index . displayed_fields ( & rtxn ) . unwrap ( ) ;
2021-01-21 00:27:43 +08:00
assert_eq! ( fields_ids . unwrap ( ) , & [ " age " ] [ .. ] ) ;
2020-11-02 22:31:20 +08:00
drop ( rtxn ) ;
// We reset the fields ids to become `None`, the default value.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2020-11-02 22:31:20 +08:00
builder . reset_displayed_fields ( ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2020-11-02 22:31:20 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Check that the displayed fields are correctly set to `None` (default value).
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let fields_ids = index . displayed_fields ( & rtxn ) . unwrap ( ) ;
assert_eq! ( fields_ids , None ) ;
}
2020-11-13 21:49:48 +08:00
#[ test ]
2021-06-01 18:19:55 +08:00
fn set_filterable_fields ( ) {
2020-11-13 21:49:48 +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 ( ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
2021-06-01 18:19:55 +08:00
// Set the filterable fields to be the age.
2020-11-13 21:49:48 +08:00
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-06-17 00:33:33 +08:00
builder . set_filterable_fields ( hashset! { S ( " age " ) } ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2020-11-13 21:49:48 +08:00
// Then index some documents.
2021-08-31 17:44:15 +08:00
let content = documents! ( [
{ " name " : " kevin " , " age " : 23 } ,
2021-05-03 21:58:47 +08:00
{ " name " : " kevina " , " age " : 21 } ,
{ " name " : " benoit " , " age " : 34 }
2021-08-31 17:44:15 +08:00
] ) ;
2021-12-08 21:12:07 +08:00
let indexing_config =
IndexDocumentsConfig { autogenerate_docids : true , .. Default ::default ( ) } ;
let mut builder =
IndexDocuments ::new ( & mut wtxn , & index , & config , indexing_config . clone ( ) , | _ | ( ) ) ;
builder . add_documents ( content ) . unwrap ( ) ;
builder . execute ( ) . unwrap ( ) ;
2020-11-13 21:49:48 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Check that the displayed fields are correctly set.
let rtxn = index . read_txn ( ) . unwrap ( ) ;
2021-06-01 18:19:55 +08:00
let fields_ids = index . filterable_fields ( & rtxn ) . unwrap ( ) ;
2021-06-17 00:33:33 +08:00
assert_eq! ( fields_ids , hashset! { S ( " age " ) } ) ;
2020-11-18 23:29:07 +08:00
// Only count the field_id 0 and level 0 facet values.
2021-05-03 21:58:47 +08:00
// TODO we must support typed CSVs for numbers to be understood.
2021-08-31 17:44:15 +08:00
let fidmap = index . fields_ids_map ( & rtxn ) . unwrap ( ) ;
println! ( " fidmap: {:?} " , fidmap ) ;
for document in index . all_documents ( & rtxn ) . unwrap ( ) {
let document = document . unwrap ( ) ;
let json = crate ::obkv_to_json ( & fidmap . ids ( ) . collect ::< Vec < _ > > ( ) , & fidmap , document . 1 )
. unwrap ( ) ;
println! ( " json: {:?} " , json ) ;
}
2021-06-17 00:33:33 +08:00
let count = index
. facet_id_f64_docids
2021-05-03 21:58:47 +08:00
. remap_key_type ::< ByteSlice > ( )
2021-08-31 17:44:15 +08:00
// The faceted field id is 1u16
. prefix_iter ( & rtxn , & [ 0 , 1 , 0 ] )
2021-06-17 00:33:33 +08:00
. unwrap ( )
. count ( ) ;
2020-11-15 18:06:51 +08:00
assert_eq! ( count , 3 ) ;
drop ( rtxn ) ;
// Index a little more documents with new and current facets values.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-08-31 17:44:15 +08:00
let content = documents! ( [
{ " name " : " kevin2 " , " age " : 23 } ,
2021-05-03 21:58:47 +08:00
{ " name " : " kevina2 " , " age " : 21 } ,
2021-08-31 17:44:15 +08:00
{ " name " : " benoit " , " age " : 35 }
] ) ;
2021-05-03 21:58:47 +08:00
2021-12-08 21:12:07 +08:00
let indexing_config =
IndexDocumentsConfig { autogenerate_docids : true , .. Default ::default ( ) } ;
let mut builder =
IndexDocuments ::new ( & mut wtxn , & index , & config , indexing_config . clone ( ) , | _ | ( ) ) ;
builder . add_documents ( content ) . unwrap ( ) ;
builder . execute ( ) . unwrap ( ) ;
2020-11-15 18:06:51 +08:00
wtxn . commit ( ) . unwrap ( ) ;
let rtxn = index . read_txn ( ) . unwrap ( ) ;
2020-11-18 23:29:07 +08:00
// Only count the field_id 0 and level 0 facet values.
2021-06-17 00:33:33 +08:00
let count = index
. facet_id_f64_docids
2021-05-03 21:58:47 +08:00
. remap_key_type ::< ByteSlice > ( )
2021-08-31 17:44:15 +08:00
. prefix_iter ( & rtxn , & [ 0 , 1 , 0 ] )
2021-06-17 00:33:33 +08:00
. unwrap ( )
. count ( ) ;
2020-11-15 18:06:51 +08:00
assert_eq! ( count , 4 ) ;
2021-03-30 01:15:47 +08:00
}
2021-06-01 22:29:14 +08:00
#[ test ]
fn set_asc_desc_field ( ) {
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 ( ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
2021-06-01 22:29:14 +08:00
// Set the filterable fields to be the age.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-06-01 22:29:14 +08:00
// Don't display the generated `id` field.
2021-06-08 23:55:08 +08:00
builder . set_displayed_fields ( vec! [ S ( " name " ) ] ) ;
2021-08-23 17:37:18 +08:00
builder . set_criteria ( vec! [ S ( " age:asc " ) ] ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2021-06-01 22:29:14 +08:00
// Then index some documents.
2021-08-31 17:44:15 +08:00
let content = documents! ( [
{ " name " : " kevin " , " age " : 23 } ,
2021-06-01 22:29:14 +08:00
{ " name " : " kevina " , " age " : 21 } ,
{ " name " : " benoit " , " age " : 34 }
2021-08-31 17:44:15 +08:00
] ) ;
2021-12-08 21:12:07 +08:00
let indexing_config =
IndexDocumentsConfig { autogenerate_docids : true , .. Default ::default ( ) } ;
let mut builder =
IndexDocuments ::new ( & mut wtxn , & index , & config , indexing_config . clone ( ) , | _ | ( ) ) ;
builder . add_documents ( content ) . unwrap ( ) ;
builder . execute ( ) . unwrap ( ) ;
2021-06-01 22:29:14 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Run an empty query just to ensure that the search results are ordered.
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let SearchResult { documents_ids , .. } = index . search ( & rtxn ) . execute ( ) . unwrap ( ) ;
let documents = index . documents ( & rtxn , documents_ids ) . unwrap ( ) ;
// Fetch the documents "age" field in the ordre in which the documents appear.
let age_field_id = index . fields_ids_map ( & rtxn ) . unwrap ( ) . id ( " age " ) . unwrap ( ) ;
let iter = documents . into_iter ( ) . map ( | ( _ , doc ) | {
let bytes = doc . get ( age_field_id ) . unwrap ( ) ;
let string = std ::str ::from_utf8 ( bytes ) . unwrap ( ) ;
string . parse ::< u32 > ( ) . unwrap ( )
} ) ;
assert_eq! ( iter . collect ::< Vec < _ > > ( ) , vec! [ 21 , 23 , 34 ] ) ;
}
#[ test ]
fn set_distinct_field ( ) {
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 ( ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
2021-06-01 22:29:14 +08:00
// Set the filterable fields to be the age.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-06-01 22:29:14 +08:00
// Don't display the generated `id` field.
builder . set_displayed_fields ( vec! [ S ( " name " ) , S ( " age " ) ] ) ;
builder . set_distinct_field ( S ( " age " ) ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2021-06-01 22:29:14 +08:00
// Then index some documents.
2021-08-31 17:44:15 +08:00
let content = documents! ( [
2021-06-01 22:29:14 +08:00
{ " name " : " kevin " , " age " : 23 } ,
{ " name " : " kevina " , " age " : 21 } ,
{ " name " : " benoit " , " age " : 34 } ,
{ " name " : " bernard " , " age " : 34 } ,
{ " name " : " bertrand " , " age " : 34 } ,
{ " name " : " bernie " , " age " : 34 } ,
{ " name " : " ben " , " age " : 34 }
2021-08-31 17:44:15 +08:00
] ) ;
2021-12-08 21:12:07 +08:00
let indexing_config =
IndexDocumentsConfig { autogenerate_docids : true , .. Default ::default ( ) } ;
let mut builder =
IndexDocuments ::new ( & mut wtxn , & index , & config , indexing_config . clone ( ) , | _ | ( ) ) ;
builder . add_documents ( content ) . unwrap ( ) ;
builder . execute ( ) . unwrap ( ) ;
2021-06-01 22:29:14 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Run an empty query just to ensure that the search results are ordered.
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let SearchResult { documents_ids , .. } = index . search ( & rtxn ) . execute ( ) . unwrap ( ) ;
// There must be at least one document with a 34 as the age.
assert_eq! ( documents_ids . len ( ) , 3 ) ;
}
2021-03-30 01:15:47 +08:00
#[ test ]
fn default_stop_words ( ) {
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 ( ) ;
// First we send 3 documents with ids from 1 to 3.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-08-31 17:44:15 +08:00
let content = documents! ( [
{ " name " : " kevin " , " age " : 23 } ,
{ " name " : " kevina " , " age " : 21 } ,
{ " name " : " benoit " , " age " : 34 }
] ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
let indexing_config =
IndexDocumentsConfig { autogenerate_docids : true , .. Default ::default ( ) } ;
let mut builder =
IndexDocuments ::new ( & mut wtxn , & index , & config , indexing_config . clone ( ) , | _ | ( ) ) ;
builder . add_documents ( content ) . unwrap ( ) ;
builder . execute ( ) . unwrap ( ) ;
2021-03-30 01:15:47 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Ensure there is no stop_words by default
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let stop_words = index . stop_words ( & rtxn ) . unwrap ( ) ;
assert! ( stop_words . is_none ( ) ) ;
}
#[ test ]
fn set_and_reset_stop_words ( ) {
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 ( ) ;
// First we send 3 documents with ids from 1 to 3.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-08-31 17:44:15 +08:00
let content = documents! ( [
{ " name " : " kevin " , " age " : 23 , " maxim " : " I love dogs " } ,
{ " name " : " kevina " , " age " : 21 , " maxim " : " Doggos are the best " } ,
{ " name " : " benoit " , " age " : 34 , " maxim " : " The crepes are really good " } ,
] ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
let indexing_config =
IndexDocumentsConfig { autogenerate_docids : true , .. Default ::default ( ) } ;
let mut builder =
IndexDocuments ::new ( & mut wtxn , & index , & config , indexing_config . clone ( ) , | _ | ( ) ) ;
builder . add_documents ( content ) . unwrap ( ) ;
builder . execute ( ) . unwrap ( ) ;
2021-03-30 01:15:47 +08:00
// In the same transaction we provide some stop_words
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-04-07 19:33:44 +08:00
let set = btreeset! { " i " . to_string ( ) , " the " . to_string ( ) , " are " . to_string ( ) } ;
2021-03-30 01:15:47 +08:00
builder . set_stop_words ( set . clone ( ) ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2021-03-30 01:15:47 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Ensure stop_words are effectively stored
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let stop_words = index . stop_words ( & rtxn ) . unwrap ( ) ;
assert! ( stop_words . is_some ( ) ) ; // at this point the index should return something
let stop_words = stop_words . unwrap ( ) ;
let expected = fst ::Set ::from_iter ( & set ) . unwrap ( ) ;
assert_eq! ( stop_words . as_fst ( ) . as_bytes ( ) , expected . as_fst ( ) . as_bytes ( ) ) ;
// when we search for something that is a non prefix stop_words it should be ignored
2021-04-09 03:21:20 +08:00
// thus we should get a placeholder search (all the results = 3)
2021-03-30 01:15:47 +08:00
let result = index . search ( & rtxn ) . query ( " the " ) . execute ( ) . unwrap ( ) ;
2021-04-09 03:21:20 +08:00
assert_eq! ( result . documents_ids . len ( ) , 3 ) ;
2021-03-30 01:15:47 +08:00
let result = index . search ( & rtxn ) . query ( " i " ) . execute ( ) . unwrap ( ) ;
2021-04-09 03:21:20 +08:00
assert_eq! ( result . documents_ids . len ( ) , 3 ) ;
2021-03-30 01:15:47 +08:00
let result = index . search ( & rtxn ) . query ( " are " ) . execute ( ) . unwrap ( ) ;
2021-04-09 03:21:20 +08:00
assert_eq! ( result . documents_ids . len ( ) , 3 ) ;
2021-03-30 01:15:47 +08:00
let result = index . search ( & rtxn ) . query ( " dog " ) . execute ( ) . unwrap ( ) ;
assert_eq! ( result . documents_ids . len ( ) , 2 ) ; // we have two maxims talking about doggos
let result = index . search ( & rtxn ) . query ( " benoît " ) . execute ( ) . unwrap ( ) ;
assert_eq! ( result . documents_ids . len ( ) , 1 ) ; // there is one benoit in our data
// now we'll reset the stop_words and ensure it's None
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-03-30 01:15:47 +08:00
builder . reset_stop_words ( ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2021-03-30 01:15:47 +08:00
wtxn . commit ( ) . unwrap ( ) ;
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let stop_words = index . stop_words ( & rtxn ) . unwrap ( ) ;
assert! ( stop_words . is_none ( ) ) ;
// now we can search for the stop words
let result = index . search ( & rtxn ) . query ( " the " ) . execute ( ) . unwrap ( ) ;
assert_eq! ( result . documents_ids . len ( ) , 2 ) ;
let result = index . search ( & rtxn ) . query ( " i " ) . execute ( ) . unwrap ( ) ;
assert_eq! ( result . documents_ids . len ( ) , 1 ) ;
let result = index . search ( & rtxn ) . query ( " are " ) . execute ( ) . unwrap ( ) ;
assert_eq! ( result . documents_ids . len ( ) , 2 ) ;
// the rest of the search is still not impacted
let result = index . search ( & rtxn ) . query ( " dog " ) . execute ( ) . unwrap ( ) ;
assert_eq! ( result . documents_ids . len ( ) , 2 ) ; // we have two maxims talking about doggos
let result = index . search ( & rtxn ) . query ( " benoît " ) . execute ( ) . unwrap ( ) ;
assert_eq! ( result . documents_ids . len ( ) , 1 ) ; // there is one benoit in our data
2020-11-13 21:49:48 +08:00
}
2021-01-21 00:27:43 +08:00
2021-04-10 03:56:20 +08:00
#[ test ]
fn set_and_reset_synonyms ( ) {
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 ( ) ;
// Send 3 documents with ids from 1 to 3.
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-08-31 17:44:15 +08:00
let content = documents! ( [
{ " name " : " kevin " , " age " : 23 , " maxim " : " I love dogs " } ,
{ " name " : " kevina " , " age " : 21 , " maxim " : " Doggos are the best " } ,
{ " name " : " benoit " , " age " : 34 , " maxim " : " The crepes are really good " } ,
] ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
let indexing_config =
IndexDocumentsConfig { autogenerate_docids : true , .. Default ::default ( ) } ;
let mut builder =
IndexDocuments ::new ( & mut wtxn , & index , & config , indexing_config . clone ( ) , | _ | ( ) ) ;
builder . add_documents ( content ) . unwrap ( ) ;
builder . execute ( ) . unwrap ( ) ;
2021-04-10 03:56:20 +08:00
// In the same transaction provide some synonyms
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-04-10 03:56:20 +08:00
builder . set_synonyms ( hashmap! {
" blini " . to_string ( ) = > vec! [ " crepes " . to_string ( ) ] ,
" super like " . to_string ( ) = > vec! [ " love " . to_string ( ) ] ,
" puppies " . to_string ( ) = > vec! [ " dogs " . to_string ( ) , " doggos " . to_string ( ) ]
} ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2021-04-10 03:56:20 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Ensure synonyms are effectively stored
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let synonyms = index . synonyms ( & rtxn ) . unwrap ( ) ;
assert! ( ! synonyms . is_empty ( ) ) ; // at this point the index should return something
// Check that we can use synonyms
let result = index . search ( & rtxn ) . query ( " blini " ) . execute ( ) . unwrap ( ) ;
assert_eq! ( result . documents_ids . len ( ) , 1 ) ;
let result = index . search ( & rtxn ) . query ( " super like " ) . execute ( ) . unwrap ( ) ;
assert_eq! ( result . documents_ids . len ( ) , 1 ) ;
let result = index . search ( & rtxn ) . query ( " puppies " ) . execute ( ) . unwrap ( ) ;
assert_eq! ( result . documents_ids . len ( ) , 2 ) ;
// Reset the synonyms
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-04-10 03:56:20 +08:00
builder . reset_synonyms ( ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2021-04-10 03:56:20 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// Ensure synonyms are reset
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let synonyms = index . synonyms ( & rtxn ) . unwrap ( ) ;
assert! ( synonyms . is_empty ( ) ) ;
// Check that synonyms are no longer work
let result = index . search ( & rtxn ) . query ( " blini " ) . execute ( ) . unwrap ( ) ;
assert! ( result . documents_ids . is_empty ( ) ) ;
let result = index . search ( & rtxn ) . query ( " super like " ) . execute ( ) . unwrap ( ) ;
assert! ( result . documents_ids . is_empty ( ) ) ;
let result = index . search ( & rtxn ) . query ( " puppies " ) . execute ( ) . unwrap ( ) ;
assert! ( result . documents_ids . is_empty ( ) ) ;
}
2021-01-21 00:27:43 +08:00
#[ test ]
fn setting_searchable_recomputes_other_settings ( ) {
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 ( ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
2021-01-21 00:27:43 +08:00
// Set all the settings except searchable
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-01-21 00:27:43 +08:00
builder . set_displayed_fields ( vec! [ " hello " . to_string ( ) ] ) ;
2021-06-17 00:33:33 +08:00
builder . set_filterable_fields ( hashset! { S ( " age " ) , S ( " toto " ) } ) ;
2021-08-23 17:37:18 +08:00
builder . set_criteria ( vec! [ " toto:asc " . to_string ( ) ] ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2021-01-21 00:27:43 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// check the output
let rtxn = index . read_txn ( ) . unwrap ( ) ;
assert_eq! ( & [ " hello " ] [ .. ] , index . displayed_fields ( & rtxn ) . unwrap ( ) . unwrap ( ) ) ;
// since no documents have been pushed the primary key is still unset
assert! ( index . primary_key ( & rtxn ) . unwrap ( ) . is_none ( ) ) ;
assert_eq! ( vec! [ Criterion ::Asc ( " toto " . to_string ( ) ) ] , index . criteria ( & rtxn ) . unwrap ( ) ) ;
drop ( rtxn ) ;
// We set toto and age as searchable to force reordering of the fields
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-01-21 00:27:43 +08:00
builder . set_searchable_fields ( vec! [ " toto " . to_string ( ) , " age " . to_string ( ) ] ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2021-01-21 00:27:43 +08:00
wtxn . commit ( ) . unwrap ( ) ;
let rtxn = index . read_txn ( ) . unwrap ( ) ;
assert_eq! ( & [ " hello " ] [ .. ] , index . displayed_fields ( & rtxn ) . unwrap ( ) . unwrap ( ) ) ;
assert! ( index . primary_key ( & rtxn ) . unwrap ( ) . is_none ( ) ) ;
assert_eq! ( vec! [ Criterion ::Asc ( " toto " . to_string ( ) ) ] , index . criteria ( & rtxn ) . unwrap ( ) ) ;
}
2021-06-01 21:48:38 +08:00
#[ test ]
fn setting_not_filterable_cant_filter ( ) {
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 ( ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
2021-06-01 21:48:38 +08:00
// Set all the settings except searchable
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-06-01 21:48:38 +08:00
builder . set_displayed_fields ( vec! [ " hello " . to_string ( ) ] ) ;
// It is only Asc(toto), there is a facet database but it is denied to filter with toto.
2021-08-23 17:37:18 +08:00
builder . set_criteria ( vec! [ " toto:asc " . to_string ( ) ] ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2021-06-01 21:48:38 +08:00
wtxn . commit ( ) . unwrap ( ) ;
let rtxn = index . read_txn ( ) . unwrap ( ) ;
2021-12-09 18:13:12 +08:00
let filter = Filter ::from_str ( " toto = 32 " ) . unwrap ( ) . unwrap ( ) ;
2021-11-06 08:32:12 +08:00
let _ = filter . evaluate ( & rtxn , & index ) . unwrap_err ( ) ;
2021-06-01 21:48:38 +08:00
}
2021-06-15 19:45:20 +08:00
#[ test ]
fn setting_primary_key ( ) {
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 ( ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
2021-06-15 19:45:20 +08:00
// Set the primary key settings
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-06-15 19:45:20 +08:00
builder . set_primary_key ( S ( " mykey " ) ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2021-06-15 19:45:20 +08:00
assert_eq! ( index . primary_key ( & wtxn ) . unwrap ( ) , Some ( " mykey " ) ) ;
// Then index some documents with the "mykey" primary key.
2021-08-31 17:44:15 +08:00
let content = documents! ( [
2021-06-15 19:45:20 +08:00
{ " mykey " : 1 , " name " : " kevin " , " age " : 23 } ,
{ " mykey " : 2 , " name " : " kevina " , " age " : 21 } ,
{ " mykey " : 3 , " name " : " benoit " , " age " : 34 } ,
{ " mykey " : 4 , " name " : " bernard " , " age " : 34 } ,
{ " mykey " : 5 , " name " : " bertrand " , " age " : 34 } ,
{ " mykey " : 6 , " name " : " bernie " , " age " : 34 } ,
{ " mykey " : 7 , " name " : " ben " , " age " : 34 }
2021-08-31 17:44:15 +08:00
] ) ;
2021-12-08 21:12:07 +08:00
let indexing_config =
IndexDocumentsConfig { autogenerate_docids : true , .. Default ::default ( ) } ;
let mut builder =
IndexDocuments ::new ( & mut wtxn , & index , & config , indexing_config . clone ( ) , | _ | ( ) ) ;
builder . add_documents ( content ) . unwrap ( ) ;
builder . execute ( ) . unwrap ( ) ;
2021-06-15 19:45:20 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// We now try to reset the primary key
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-06-15 19:45:20 +08:00
builder . reset_primary_key ( ) ;
2021-11-03 20:12:01 +08:00
let err = builder . execute ( | _ | ( ) ) . unwrap_err ( ) ;
2021-10-26 23:49:35 +08:00
assert! ( matches! ( err , Error ::UserError ( UserError ::PrimaryKeyCannotBeChanged ( _ ) ) ) ) ;
2021-06-15 23:49:15 +08:00
wtxn . abort ( ) . unwrap ( ) ;
2021-06-15 19:45:20 +08:00
// But if we clear the database...
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-11-03 20:12:01 +08:00
let builder = ClearDocuments ::new ( & mut wtxn , & index ) ;
2021-06-15 19:45:20 +08:00
builder . execute ( ) . unwrap ( ) ;
// ...we can change the primary key
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-06-15 19:45:20 +08:00
builder . set_primary_key ( S ( " myid " ) ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2021-06-15 23:49:15 +08:00
wtxn . commit ( ) . unwrap ( ) ;
2021-06-15 19:45:20 +08:00
}
2021-07-22 23:04:38 +08:00
#[ test ]
fn setting_impact_relevancy ( ) {
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 ( ) ;
2021-12-08 21:12:07 +08:00
let config = IndexerConfig ::default ( ) ;
2021-07-22 23:04:38 +08:00
// Set the genres setting
let mut wtxn = index . write_txn ( ) . unwrap ( ) ;
2021-12-08 21:12:07 +08:00
let mut builder = Settings ::new ( & mut wtxn , & index , & config ) ;
2021-07-22 23:04:38 +08:00
builder . set_filterable_fields ( hashset! { S ( " genres " ) } ) ;
2021-11-03 20:12:01 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
2021-07-22 23:04:38 +08:00
2021-08-31 17:44:15 +08:00
let content = documents! ( [
2021-07-22 23:04:38 +08:00
{
" id " : 11 ,
" title " : " Star Wars " ,
" overview " :
" Princess Leia is captured and held hostage by the evil Imperial forces in their effort to take over the galactic Empire. Venturesome Luke Skywalker and dashing captain Han Solo team together with the loveable robot duo R2-D2 and C-3PO to rescue the beautiful princess and restore peace and justice in the Empire. " ,
" genres " : [ " Adventure " , " Action " , " Science Fiction " ] ,
" poster " : " https://image.tmdb.org/t/p/w500/6FfCtAuVAW8XJjZ7eWeLibRLWTw.jpg " ,
" release_date " : 233366400
} ,
{
" id " : 30 ,
" title " : " Magnetic Rose " ,
" overview " : " " ,
" genres " : [ " Animation " , " Science Fiction " ] ,
" poster " : " https://image.tmdb.org/t/p/w500/gSuHDeWemA1menrwfMRChnSmMVN.jpg " ,
" release_date " : 819676800
}
2021-08-31 17:44:15 +08:00
] ) ;
2021-12-08 21:12:07 +08:00
let indexing_config =
IndexDocumentsConfig { autogenerate_docids : true , .. Default ::default ( ) } ;
let mut builder =
IndexDocuments ::new ( & mut wtxn , & index , & config , indexing_config . clone ( ) , | _ | ( ) ) ;
builder . add_documents ( content ) . unwrap ( ) ;
builder . execute ( ) . unwrap ( ) ;
2021-07-22 23:04:38 +08:00
wtxn . commit ( ) . unwrap ( ) ;
// We now try to reset the primary key
let rtxn = index . read_txn ( ) . unwrap ( ) ;
let SearchResult { documents_ids , .. } = index . search ( & rtxn ) . query ( " S " ) . execute ( ) . unwrap ( ) ;
let first_id = documents_ids [ 0 ] ;
let documents = index . documents ( & rtxn , documents_ids ) . unwrap ( ) ;
let ( _ , content ) = documents . iter ( ) . find ( | ( id , _ ) | * id = = first_id ) . unwrap ( ) ;
let fid = index . fields_ids_map ( & rtxn ) . unwrap ( ) . id ( " title " ) . unwrap ( ) ;
let line = std ::str ::from_utf8 ( content . get ( fid ) . unwrap ( ) ) . unwrap ( ) ;
assert_eq! ( line , r # ""Star Wars""# ) ;
}
2022-03-31 20:12:00 +08:00
#[ test ]
fn test_disable_typo ( ) {
let index = TempIndex ::new ( ) ;
let mut txn = index . write_txn ( ) . unwrap ( ) ;
let config = IndexerConfig ::default ( ) ;
assert! ( index . authorize_typos ( & txn ) . unwrap ( ) ) ;
let mut builder = Settings ::new ( & mut txn , & index , & config ) ;
builder . set_autorize_typos ( false ) ;
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
assert! ( ! index . authorize_typos ( & txn ) . unwrap ( ) ) ;
}
2022-03-31 19:50:18 +08:00
#[ test ]
fn update_min_word_len_for_typo ( ) {
let index = TempIndex ::new ( ) ;
let config = IndexerConfig ::default ( ) ;
// Set the genres setting
let mut txn = index . write_txn ( ) . unwrap ( ) ;
let mut builder = Settings ::new ( & mut txn , & index , & config ) ;
2022-04-01 00:23:12 +08:00
builder . set_min_word_len_one_typo ( 8 ) ;
builder . set_min_word_len_two_typos ( 8 ) ;
2022-03-31 19:50:18 +08:00
builder . execute ( | _ | ( ) ) . unwrap ( ) ;
txn . commit ( ) . unwrap ( ) ;
let txn = index . read_txn ( ) . unwrap ( ) ;
2022-04-01 00:23:12 +08:00
assert_eq! ( index . min_word_len_one_typo ( & txn ) . unwrap ( ) , 8 ) ;
assert_eq! ( index . min_word_len_two_typos ( & txn ) . unwrap ( ) , 8 ) ;
2022-03-31 19:50:18 +08:00
}
#[ test ]
fn update_invalid_min_word_len_for_typo ( ) {
let index = TempIndex ::new ( ) ;
let config = IndexerConfig ::default ( ) ;
// Set the genres setting
let mut txn = index . write_txn ( ) . unwrap ( ) ;
let mut builder = Settings ::new ( & mut txn , & index , & config ) ;
2022-04-01 00:23:12 +08:00
builder . set_min_word_len_one_typo ( 10 ) ;
builder . set_min_word_len_two_typos ( 7 ) ;
2022-03-31 19:50:18 +08:00
assert! ( builder . execute ( | _ | ( ) ) . is_err ( ) ) ;
}
2020-11-02 22:31:20 +08:00
}