mirror of
https://github.com/meilisearch/MeiliSearch
synced 2025-03-13 11:33:02 +01:00
Revert metadata creation when computing facet search and distinct
This commit is contained in:
parent
b12ffd1356
commit
abef655849
@ -548,7 +548,7 @@ async fn facet_search_with_filterable_attributes_rules_errors() {
|
|||||||
&json!({"facetName": "invalid", "facetQuery": "a"}),
|
&json!({"facetName": "invalid", "facetQuery": "a"}),
|
||||||
|response, code| {
|
|response, code| {
|
||||||
snapshot!(code, @"400 Bad Request");
|
snapshot!(code, @"400 Bad Request");
|
||||||
snapshot!(response["message"], @r###""Attribute `invalid` is not facet-searchable. Available facet-searchable attributes are: `genres`. To make it facet-searchable add it to the `filterableAttributes` index settings.""###);
|
snapshot!(response["message"], @r###""Attribute `invalid` is not facet-searchable. Available facet-searchable attributes patterns are: `genres`. To make it facet-searchable add it to the `filterableAttributes` index settings.""###);
|
||||||
},
|
},
|
||||||
)
|
)
|
||||||
.await;
|
.await;
|
||||||
|
@ -158,28 +158,32 @@ and can not be more than 511 bytes.", .document_id.to_string()
|
|||||||
InvalidSortableAttribute { field: String, valid_fields: BTreeSet<String>, hidden_fields: bool },
|
InvalidSortableAttribute { field: String, valid_fields: BTreeSet<String>, hidden_fields: bool },
|
||||||
#[error("Attribute `{}` is not filterable and thus, cannot be used as distinct attribute. {}",
|
#[error("Attribute `{}` is not filterable and thus, cannot be used as distinct attribute. {}",
|
||||||
.field,
|
.field,
|
||||||
match .valid_fields.is_empty() {
|
match .valid_patterns.is_empty() {
|
||||||
true => "This index does not have configured filterable attributes.".to_string(),
|
true => "This index does not have configured filterable attributes.".to_string(),
|
||||||
false => format!("Available filterable attributes are: `{}{}`.",
|
false => format!("Available filterable attributes patterns are: `{}{}`.",
|
||||||
valid_fields.iter().map(AsRef::as_ref).collect::<Vec<&str>>().join(", "),
|
valid_patterns.iter().map(AsRef::as_ref).collect::<Vec<&str>>().join(", "),
|
||||||
.hidden_fields.then_some(", <..hidden-attributes>").unwrap_or(""),
|
.hidden_fields.then_some(", <..hidden-attributes>").unwrap_or(""),
|
||||||
),
|
),
|
||||||
}
|
}
|
||||||
)]
|
)]
|
||||||
InvalidDistinctAttribute { field: String, valid_fields: BTreeSet<String>, hidden_fields: bool },
|
InvalidDistinctAttribute {
|
||||||
|
field: String,
|
||||||
|
valid_patterns: BTreeSet<String>,
|
||||||
|
hidden_fields: bool,
|
||||||
|
},
|
||||||
#[error("Attribute `{}` is not facet-searchable. {}",
|
#[error("Attribute `{}` is not facet-searchable. {}",
|
||||||
.field,
|
.field,
|
||||||
match .valid_fields.is_empty() {
|
match .valid_patterns.is_empty() {
|
||||||
true => "This index does not have configured facet-searchable attributes. To make it facet-searchable add it to the `filterableAttributes` index settings.".to_string(),
|
true => "This index does not have configured facet-searchable attributes. To make it facet-searchable add it to the `filterableAttributes` index settings.".to_string(),
|
||||||
false => format!("Available facet-searchable attributes are: `{}{}`. To make it facet-searchable add it to the `filterableAttributes` index settings.",
|
false => format!("Available facet-searchable attributes patterns are: `{}{}`. To make it facet-searchable add it to the `filterableAttributes` index settings.",
|
||||||
valid_fields.iter().map(AsRef::as_ref).collect::<Vec<&str>>().join(", "),
|
valid_patterns.iter().map(AsRef::as_ref).collect::<Vec<&str>>().join(", "),
|
||||||
.hidden_fields.then_some(", <..hidden-attributes>").unwrap_or(""),
|
.hidden_fields.then_some(", <..hidden-attributes>").unwrap_or(""),
|
||||||
),
|
),
|
||||||
}
|
}
|
||||||
)]
|
)]
|
||||||
InvalidFacetSearchFacetName {
|
InvalidFacetSearchFacetName {
|
||||||
field: String,
|
field: String,
|
||||||
valid_fields: BTreeSet<String>,
|
valid_patterns: BTreeSet<String>,
|
||||||
hidden_fields: bool,
|
hidden_fields: bool,
|
||||||
},
|
},
|
||||||
#[error("Attribute `{}` is not searchable. Available searchable attributes are: `{}{}`.",
|
#[error("Attribute `{}` is not searchable. Available searchable attributes are: `{}{}`.",
|
||||||
|
@ -6,7 +6,6 @@ use utoipa::ToSchema;
|
|||||||
use crate::{
|
use crate::{
|
||||||
attribute_patterns::{match_distinct_field, match_field_legacy, PatternMatch},
|
attribute_patterns::{match_distinct_field, match_field_legacy, PatternMatch},
|
||||||
constants::RESERVED_GEO_FIELD_NAME,
|
constants::RESERVED_GEO_FIELD_NAME,
|
||||||
fields_ids_map::metadata::FieldIdMapWithMetadata,
|
|
||||||
AttributePatterns,
|
AttributePatterns,
|
||||||
};
|
};
|
||||||
|
|
||||||
@ -225,30 +224,6 @@ impl Default for FilterFeatures {
|
|||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
/// Match a field against a set of filterable attributes rules.
|
|
||||||
///
|
|
||||||
/// This function will return the set of field names that match the given filter.
|
|
||||||
///
|
|
||||||
/// # Arguments
|
|
||||||
///
|
|
||||||
/// * `filterable_attributes` - The set of filterable attributes rules to match against.
|
|
||||||
/// * `fields_ids_map` - The map of field names to field ids.
|
|
||||||
/// * `filter` - The filter function to apply to the filterable attributes rules.
|
|
||||||
pub fn filtered_matching_field_names<'fim>(
|
|
||||||
filterable_attributes: &[FilterableAttributesRule],
|
|
||||||
fields_ids_map: &'fim FieldIdMapWithMetadata,
|
|
||||||
filter: &impl Fn(FilterableAttributesFeatures) -> bool,
|
|
||||||
) -> BTreeSet<&'fim str> {
|
|
||||||
let mut result = BTreeSet::new();
|
|
||||||
for (_, field_name, metadata) in fields_ids_map.iter() {
|
|
||||||
let features = metadata.filterable_attributes_features(filterable_attributes);
|
|
||||||
if filter(features) {
|
|
||||||
result.insert(field_name);
|
|
||||||
}
|
|
||||||
}
|
|
||||||
result
|
|
||||||
}
|
|
||||||
|
|
||||||
/// Match a field against a set of filterable attributes rules.
|
/// Match a field against a set of filterable attributes rules.
|
||||||
///
|
///
|
||||||
/// This function will return the set of patterns that match the given filter.
|
/// This function will return the set of patterns that match the given filter.
|
||||||
@ -306,34 +281,6 @@ pub fn matching_features(
|
|||||||
None
|
None
|
||||||
}
|
}
|
||||||
|
|
||||||
/// Check if a field is filterable calling the method `FilterableAttributesFeatures::is_filterable()`.
|
|
||||||
///
|
|
||||||
/// # Arguments
|
|
||||||
///
|
|
||||||
/// * `field_name` - The field name to check.
|
|
||||||
/// * `filterable_attributes` - The set of filterable attributes rules to match against.
|
|
||||||
pub fn is_field_filterable(
|
|
||||||
field_name: &str,
|
|
||||||
filterable_attributes: &[FilterableAttributesRule],
|
|
||||||
) -> bool {
|
|
||||||
matching_features(field_name, filterable_attributes)
|
|
||||||
.map_or(false, |(_, features)| features.is_filterable())
|
|
||||||
}
|
|
||||||
|
|
||||||
/// Check if a field is facet searchable calling the method `FilterableAttributesFeatures::is_facet_searchable()`.
|
|
||||||
///
|
|
||||||
/// # Arguments
|
|
||||||
///
|
|
||||||
/// * `field_name` - The field name to check.
|
|
||||||
/// * `filterable_attributes` - The set of filterable attributes rules to match against.
|
|
||||||
pub fn is_field_facet_searchable(
|
|
||||||
field_name: &str,
|
|
||||||
filterable_attributes: &[FilterableAttributesRule],
|
|
||||||
) -> bool {
|
|
||||||
matching_features(field_name, filterable_attributes)
|
|
||||||
.map_or(false, |(_, features)| features.is_facet_searchable())
|
|
||||||
}
|
|
||||||
|
|
||||||
/// Match a field against a set of filterable, facet searchable fields, distinct field, sortable fields, and asc_desc fields.
|
/// Match a field against a set of filterable, facet searchable fields, distinct field, sortable fields, and asc_desc fields.
|
||||||
pub fn match_faceted_field(
|
pub fn match_faceted_field(
|
||||||
field_name: &str,
|
field_name: &str,
|
||||||
|
@ -10,9 +10,7 @@ use roaring::RoaringBitmap;
|
|||||||
use tracing::error;
|
use tracing::error;
|
||||||
|
|
||||||
use crate::error::UserError;
|
use crate::error::UserError;
|
||||||
use crate::filterable_attributes_rules::{
|
use crate::filterable_attributes_rules::{filtered_matching_patterns, matching_features};
|
||||||
filtered_matching_field_names, is_field_facet_searchable,
|
|
||||||
};
|
|
||||||
use crate::heed_codec::facet::{FacetGroupKey, FacetGroupValue};
|
use crate::heed_codec::facet::{FacetGroupKey, FacetGroupValue};
|
||||||
use crate::search::build_dfa;
|
use crate::search::build_dfa;
|
||||||
use crate::{DocumentId, FieldId, OrderBy, Result, Search};
|
use crate::{DocumentId, FieldId, OrderBy, Result, Search};
|
||||||
@ -77,37 +75,27 @@ impl<'a> SearchForFacetValues<'a> {
|
|||||||
let rtxn = self.search_query.rtxn;
|
let rtxn = self.search_query.rtxn;
|
||||||
|
|
||||||
let filterable_attributes_rules = index.filterable_attributes_rules(rtxn)?;
|
let filterable_attributes_rules = index.filterable_attributes_rules(rtxn)?;
|
||||||
let fields_ids_map = index.fields_ids_map_with_metadata(rtxn)?;
|
if !matching_features(&self.facet, &filterable_attributes_rules)
|
||||||
let fid = match fields_ids_map.id_with_metadata(&self.facet) {
|
.map_or(false, |(_, features)| features.is_facet_searchable())
|
||||||
Some((fid, metadata))
|
|
||||||
if metadata
|
|
||||||
.filterable_attributes_features(&filterable_attributes_rules)
|
|
||||||
.is_facet_searchable() =>
|
|
||||||
{
|
{
|
||||||
fid
|
let matching_field_names =
|
||||||
}
|
filtered_matching_patterns(&filterable_attributes_rules, &|features| {
|
||||||
// we return an empty list of results when the attribute has been
|
features.is_facet_searchable()
|
||||||
// set as filterable but no document contains this field (yet).
|
});
|
||||||
None if is_field_facet_searchable(&self.facet, &filterable_attributes_rules) => {
|
let (valid_patterns, hidden_fields) =
|
||||||
return Ok(Vec::new());
|
|
||||||
}
|
|
||||||
// we return an error when the attribute is not facet searchable
|
|
||||||
_otherwise => {
|
|
||||||
let matching_field_names = filtered_matching_field_names(
|
|
||||||
&filterable_attributes_rules,
|
|
||||||
&fields_ids_map,
|
|
||||||
&|features| features.is_facet_searchable(),
|
|
||||||
);
|
|
||||||
let (valid_fields, hidden_fields) =
|
|
||||||
index.remove_hidden_fields(rtxn, matching_field_names)?;
|
index.remove_hidden_fields(rtxn, matching_field_names)?;
|
||||||
|
|
||||||
return Err(UserError::InvalidFacetSearchFacetName {
|
return Err(UserError::InvalidFacetSearchFacetName {
|
||||||
field: self.facet.clone(),
|
field: self.facet.clone(),
|
||||||
valid_fields,
|
valid_patterns,
|
||||||
hidden_fields,
|
hidden_fields,
|
||||||
}
|
}
|
||||||
.into());
|
.into());
|
||||||
}
|
};
|
||||||
|
|
||||||
|
let fields_ids_map = index.fields_ids_map(rtxn)?;
|
||||||
|
let Some(fid) = fields_ids_map.id(&self.facet) else {
|
||||||
|
return Ok(Vec::new());
|
||||||
};
|
};
|
||||||
|
|
||||||
let fst = match self.search_query.index.facet_id_string_fst.get(rtxn, &fid)? {
|
let fst = match self.search_query.index.facet_id_string_fst.get(rtxn, &fid)? {
|
||||||
|
@ -9,7 +9,7 @@ use roaring::bitmap::RoaringBitmap;
|
|||||||
pub use self::facet::{FacetDistribution, Filter, OrderBy, DEFAULT_VALUES_PER_FACET};
|
pub use self::facet::{FacetDistribution, Filter, OrderBy, DEFAULT_VALUES_PER_FACET};
|
||||||
pub use self::new::matches::{FormatOptions, MatchBounds, MatcherBuilder, MatchingWords};
|
pub use self::new::matches::{FormatOptions, MatchBounds, MatcherBuilder, MatchingWords};
|
||||||
use self::new::{execute_vector_search, PartialSearchResult};
|
use self::new::{execute_vector_search, PartialSearchResult};
|
||||||
use crate::filterable_attributes_rules::{filtered_matching_field_names, is_field_filterable};
|
use crate::filterable_attributes_rules::{filtered_matching_patterns, matching_features};
|
||||||
use crate::score_details::{ScoreDetails, ScoringStrategy};
|
use crate::score_details::{ScoreDetails, ScoringStrategy};
|
||||||
use crate::vector::Embedder;
|
use crate::vector::Embedder;
|
||||||
use crate::{
|
use crate::{
|
||||||
@ -190,20 +190,20 @@ impl<'a> Search<'a> {
|
|||||||
if let Some(distinct) = &self.distinct {
|
if let Some(distinct) = &self.distinct {
|
||||||
let filterable_fields = ctx.index.filterable_attributes_rules(ctx.txn)?;
|
let filterable_fields = ctx.index.filterable_attributes_rules(ctx.txn)?;
|
||||||
// check if the distinct field is in the filterable fields
|
// check if the distinct field is in the filterable fields
|
||||||
if !is_field_filterable(distinct, &filterable_fields) {
|
if !matching_features(distinct, &filterable_fields)
|
||||||
|
.map_or(false, |(_, features)| features.is_filterable())
|
||||||
|
{
|
||||||
// if not, remove the hidden fields from the filterable fields to generate the error message
|
// if not, remove the hidden fields from the filterable fields to generate the error message
|
||||||
let fields_ids_map = ctx.index.fields_ids_map_with_metadata(ctx.txn)?;
|
let matching_patterns =
|
||||||
let matching_field_names = filtered_matching_field_names(
|
filtered_matching_patterns(&filterable_fields, &|features| {
|
||||||
&filterable_fields,
|
features.is_filterable()
|
||||||
&fields_ids_map,
|
});
|
||||||
&|features| features.is_filterable(),
|
let (valid_patterns, hidden_fields) =
|
||||||
);
|
ctx.index.remove_hidden_fields(ctx.txn, matching_patterns)?;
|
||||||
let (valid_fields, hidden_fields) =
|
|
||||||
ctx.index.remove_hidden_fields(ctx.txn, matching_field_names)?;
|
|
||||||
// and return the error
|
// and return the error
|
||||||
return Err(Error::UserError(UserError::InvalidDistinctAttribute {
|
return Err(Error::UserError(UserError::InvalidDistinctAttribute {
|
||||||
field: distinct.clone(),
|
field: distinct.clone(),
|
||||||
valid_fields,
|
valid_patterns,
|
||||||
hidden_fields,
|
hidden_fields,
|
||||||
}));
|
}));
|
||||||
}
|
}
|
||||||
|
@ -770,12 +770,11 @@ mod tests {
|
|||||||
use bumpalo::Bump;
|
use bumpalo::Bump;
|
||||||
use fst::IntoStreamer;
|
use fst::IntoStreamer;
|
||||||
use heed::RwTxn;
|
use heed::RwTxn;
|
||||||
use maplit::{btreeset, hashset};
|
use maplit::hashset;
|
||||||
|
|
||||||
use super::*;
|
use super::*;
|
||||||
use crate::constants::RESERVED_GEO_FIELD_NAME;
|
use crate::constants::RESERVED_GEO_FIELD_NAME;
|
||||||
use crate::documents::mmap_from_objects;
|
use crate::documents::mmap_from_objects;
|
||||||
use crate::filterable_attributes_rules::filtered_matching_field_names;
|
|
||||||
use crate::index::tests::TempIndex;
|
use crate::index::tests::TempIndex;
|
||||||
use crate::index::IndexEmbeddingConfig;
|
use crate::index::IndexEmbeddingConfig;
|
||||||
use crate::progress::Progress;
|
use crate::progress::Progress;
|
||||||
@ -1255,14 +1254,6 @@ mod tests {
|
|||||||
|
|
||||||
let rtxn = index.read_txn().unwrap();
|
let rtxn = index.read_txn().unwrap();
|
||||||
|
|
||||||
let filterable_fields = index.filterable_attributes_rules(&rtxn).unwrap();
|
|
||||||
let fields_ids_map = index.fields_ids_map_with_metadata(&rtxn).unwrap();
|
|
||||||
let facets =
|
|
||||||
filtered_matching_field_names(&filterable_fields, &fields_ids_map, &|features| {
|
|
||||||
features.is_filterable()
|
|
||||||
});
|
|
||||||
assert_eq!(facets, btreeset!("title", "nested.object", "nested.machin"));
|
|
||||||
|
|
||||||
// testing the simple query search
|
// testing the simple query search
|
||||||
let mut search = crate::Search::new(&rtxn, &index);
|
let mut search = crate::Search::new(&rtxn, &index);
|
||||||
search.query("document");
|
search.query("document");
|
||||||
@ -1478,15 +1469,6 @@ mod tests {
|
|||||||
|
|
||||||
let rtxn = index.read_txn().unwrap();
|
let rtxn = index.read_txn().unwrap();
|
||||||
|
|
||||||
let filterable_fields = index.filterable_attributes_rules(&rtxn).unwrap();
|
|
||||||
let fields_ids_map = index.fields_ids_map_with_metadata(&rtxn).unwrap();
|
|
||||||
let facets =
|
|
||||||
filtered_matching_field_names(&filterable_fields, &fields_ids_map, &|features| {
|
|
||||||
features.is_filterable()
|
|
||||||
});
|
|
||||||
|
|
||||||
assert_eq!(facets, btreeset!("dog", "dog.race", "dog.race.bernese mountain"));
|
|
||||||
|
|
||||||
for (s, i) in [("zeroth", 0), ("first", 1), ("second", 2), ("third", 3)] {
|
for (s, i) in [("zeroth", 0), ("first", 1), ("second", 2), ("third", 3)] {
|
||||||
let mut search = crate::Search::new(&rtxn, &index);
|
let mut search = crate::Search::new(&rtxn, &index);
|
||||||
let filter = format!(r#""dog.race.bernese mountain" = {s}"#);
|
let filter = format!(r#""dog.race.bernese mountain" = {s}"#);
|
||||||
@ -1504,17 +1486,6 @@ mod tests {
|
|||||||
db_snap!(index, facet_id_string_docids, @"");
|
db_snap!(index, facet_id_string_docids, @"");
|
||||||
db_snap!(index, field_id_docid_facet_strings, @"");
|
db_snap!(index, field_id_docid_facet_strings, @"");
|
||||||
|
|
||||||
let rtxn = index.read_txn().unwrap();
|
|
||||||
|
|
||||||
let filterable_fields = index.filterable_attributes_rules(&rtxn).unwrap();
|
|
||||||
let fields_ids_map = index.fields_ids_map_with_metadata(&rtxn).unwrap();
|
|
||||||
let facets =
|
|
||||||
filtered_matching_field_names(&filterable_fields, &fields_ids_map, &|features| {
|
|
||||||
features.is_filterable()
|
|
||||||
});
|
|
||||||
|
|
||||||
assert_eq!(facets, btreeset!());
|
|
||||||
|
|
||||||
// update the settings to test the sortable
|
// update the settings to test the sortable
|
||||||
index
|
index
|
||||||
.update_settings(|settings| {
|
.update_settings(|settings| {
|
||||||
@ -1744,13 +1715,6 @@ mod tests {
|
|||||||
|
|
||||||
let check_ok = |index: &Index| {
|
let check_ok = |index: &Index| {
|
||||||
let rtxn = index.read_txn().unwrap();
|
let rtxn = index.read_txn().unwrap();
|
||||||
let filterable_fields = index.filterable_attributes_rules(&rtxn).unwrap();
|
|
||||||
let fields_ids_map = index.fields_ids_map_with_metadata(&rtxn).unwrap();
|
|
||||||
let facets =
|
|
||||||
filtered_matching_field_names(&filterable_fields, &fields_ids_map, &|features| {
|
|
||||||
features.is_filterable()
|
|
||||||
});
|
|
||||||
assert_eq!(facets, btreeset!("colour", "colour.green", "colour.green.blue"));
|
|
||||||
|
|
||||||
let colour_id = index.fields_ids_map(&rtxn).unwrap().id("colour").unwrap();
|
let colour_id = index.fields_ids_map(&rtxn).unwrap().id("colour").unwrap();
|
||||||
let colour_green_id = index.fields_ids_map(&rtxn).unwrap().id("colour.green").unwrap();
|
let colour_green_id = index.fields_ids_map(&rtxn).unwrap().id("colour.green").unwrap();
|
||||||
@ -1855,13 +1819,6 @@ mod tests {
|
|||||||
|
|
||||||
let check_ok = |index: &Index| {
|
let check_ok = |index: &Index| {
|
||||||
let rtxn = index.read_txn().unwrap();
|
let rtxn = index.read_txn().unwrap();
|
||||||
let filterable_fields = index.filterable_attributes_rules(&rtxn).unwrap();
|
|
||||||
let fields_ids_map = index.fields_ids_map_with_metadata(&rtxn).unwrap();
|
|
||||||
let facets =
|
|
||||||
filtered_matching_field_names(&filterable_fields, &fields_ids_map, &|features| {
|
|
||||||
features.is_filterable()
|
|
||||||
});
|
|
||||||
assert_eq!(facets, btreeset!("colour", "colour.green", "colour.green.blue"));
|
|
||||||
|
|
||||||
let colour_id = index.fields_ids_map(&rtxn).unwrap().id("colour").unwrap();
|
let colour_id = index.fields_ids_map(&rtxn).unwrap().id("colour").unwrap();
|
||||||
let colour_green_id = index.fields_ids_map(&rtxn).unwrap().id("colour.green").unwrap();
|
let colour_green_id = index.fields_ids_map(&rtxn).unwrap().id("colour.green").unwrap();
|
||||||
@ -1924,13 +1881,6 @@ mod tests {
|
|||||||
|
|
||||||
let check_ok = |index: &Index| {
|
let check_ok = |index: &Index| {
|
||||||
let rtxn = index.read_txn().unwrap();
|
let rtxn = index.read_txn().unwrap();
|
||||||
let filterable_fields = index.filterable_attributes_rules(&rtxn).unwrap();
|
|
||||||
let fields_ids_map = index.fields_ids_map_with_metadata(&rtxn).unwrap();
|
|
||||||
let facets =
|
|
||||||
filtered_matching_field_names(&filterable_fields, &fields_ids_map, &|features| {
|
|
||||||
features.is_filterable()
|
|
||||||
});
|
|
||||||
assert_eq!(facets, btreeset!("tags", "tags.green", "tags.green.blue"));
|
|
||||||
|
|
||||||
let tags_id = index.fields_ids_map(&rtxn).unwrap().id("tags").unwrap();
|
let tags_id = index.fields_ids_map(&rtxn).unwrap().id("tags").unwrap();
|
||||||
let tags_green_id = index.fields_ids_map(&rtxn).unwrap().id("tags.green").unwrap();
|
let tags_green_id = index.fields_ids_map(&rtxn).unwrap().id("tags.green").unwrap();
|
||||||
|
Loading…
x
Reference in New Issue
Block a user