MeiliSearch/milli/tests/search/distinct.rs

79 lines
3.2 KiB
Rust
Raw Normal View History

2021-06-17 14:24:59 +02:00
use std::collections::HashSet;
use big_s::S;
use milli::update::Settings;
2022-08-18 17:36:08 +02:00
use milli::{Criterion, Search, SearchResult, TermsMatchingStrategy};
2021-06-17 14:24:59 +02:00
use Criterion::*;
use crate::search::{self, EXTERNAL_DOCUMENTS_IDS};
macro_rules! test_distinct {
2022-04-09 14:50:43 +02:00
($func:ident, $distinct:ident, $criteria:expr, $n_res:expr) => {
2021-06-17 14:24:59 +02:00
#[test]
fn $func() {
let criteria = $criteria;
let index = search::setup_search_index_with_criteria(&criteria);
// update distinct attribute
let mut wtxn = index.write_txn().unwrap();
let config = milli::update::IndexerConfig::default();
let mut builder = Settings::new(&mut wtxn, &index, &config);
2021-06-17 14:24:59 +02:00
builder.set_distinct_field(S(stringify!($distinct)));
builder.execute(|_| (), || false).unwrap();
2021-06-17 14:24:59 +02:00
wtxn.commit().unwrap();
2021-06-17 15:19:03 +02:00
let rtxn = index.read_txn().unwrap();
2021-06-17 14:24:59 +02:00
2021-06-17 15:19:03 +02:00
let mut search = Search::new(&rtxn, &index);
2021-06-17 14:24:59 +02:00
search.query(search::TEST_QUERY);
search.limit(EXTERNAL_DOCUMENTS_IDS.len());
search.authorize_typos(true);
search.terms_matching_strategy(TermsMatchingStrategy::default());
2021-06-17 14:24:59 +02:00
2022-04-09 14:50:43 +02:00
let SearchResult { documents_ids, candidates, .. } = search.execute().unwrap();
assert_eq!(candidates.len(), $n_res);
2021-06-17 14:24:59 +02:00
let mut distinct_values = HashSet::new();
2022-08-18 17:36:08 +02:00
let expected_external_ids: Vec<_> =
search::expected_order(&criteria, true, TermsMatchingStrategy::default(), &[])
.into_iter()
.filter_map(|d| {
if distinct_values.contains(&d.$distinct) {
None
} else {
distinct_values.insert(d.$distinct.to_owned());
Some(d.id)
}
})
.collect();
2021-06-17 14:24:59 +02:00
let documents_ids = search::internal_to_external_ids(&index, &documents_ids);
assert_eq!(documents_ids, expected_external_ids);
}
};
}
2021-06-17 15:19:03 +02:00
test_distinct!(
distinct_string_default_criteria,
tag,
2022-04-09 14:50:43 +02:00
vec![Words, Typo, Proximity, Attribute, Exactness],
3
2021-06-17 15:19:03 +02:00
);
test_distinct!(
distinct_number_default_criteria,
asc_desc_rank,
2022-04-09 14:50:43 +02:00
vec![Words, Typo, Proximity, Attribute, Exactness],
7
2021-06-17 15:19:03 +02:00
);
2022-04-09 14:50:43 +02:00
test_distinct!(distinct_string_criterion_words, tag, vec![Words], 3);
test_distinct!(distinct_number_criterion_words, asc_desc_rank, vec![Words], 7);
test_distinct!(distinct_string_criterion_words_typo, tag, vec![Words, Typo], 3);
test_distinct!(distinct_number_criterion_words_typo, asc_desc_rank, vec![Words, Typo], 7);
test_distinct!(distinct_string_criterion_words_proximity, tag, vec![Words, Proximity], 3);
test_distinct!(distinct_number_criterion_words_proximity, asc_desc_rank, vec![Words, Proximity], 7);
test_distinct!(distinct_string_criterion_words_attribute, tag, vec![Words, Attribute], 3);
test_distinct!(distinct_number_criterion_words_attribute, asc_desc_rank, vec![Words, Attribute], 7);
test_distinct!(distinct_string_criterion_words_exactness, tag, vec![Words, Exactness], 3);
test_distinct!(distinct_number_criterion_words_exactness, asc_desc_rank, vec![Words, Exactness], 7);