Merge #238

238: Integration tests on filters and distinct r=Kerollmops a=ManyTheFish

Fix #216 
Fix #120 

Co-authored-by: many <maxime@meilisearch.com>

milli/tests/assets/test_set.ndjson

@@ -2,7 +2,7 @@
 {"id":"B","word_rank":2,"typo_rank":0,"proximity_rank":0,"attribute_rank":0,"exact_rank":4,"asc_desc_rank":1,"title":"hello","description":"hello is a song recorded by english singer songwriter adele","tag":"red","":""}
 {"id":"C","word_rank":0,"typo_rank":1,"proximity_rank":8,"attribute_rank":336,"exact_rank":4,"asc_desc_rank":2,"title":"hell on earth","description":"hell on earth is the third studio album by american hip hop duo mobb deep","tag":"blue","":""}
 {"id":"D","word_rank":0,"typo_rank":1,"proximity_rank":10,"attribute_rank":757,"exact_rank":4,"asc_desc_rank":3,"title":"hell on wheels tv series","description":"the construction of the first transcontinental railroad across the united states in the world","tag":"red","":""}
-{"id":"E","word_rank":2,"typo_rank":0,"proximity_rank":0,"attribute_rank":0,"exact_rank":1,"asc_desc_rank":4,"title":"hello kitty","description":"also known by her full name kitty white is a fictional character produced by the japanese company sanrio","tag":"green","":""}
+{"id":"E","word_rank":2,"typo_rank":0,"proximity_rank":0,"attribute_rank":0,"exact_rank":4,"asc_desc_rank":4,"title":"hello kitty","description":"also known by her full name kitty white is a fictional character produced by the japanese company sanrio","tag":"green","":""}
 {"id":"F","word_rank":2,"typo_rank":1,"proximity_rank":0,"attribute_rank":1017,"exact_rank":5,"asc_desc_rank":5,"title":"laptop orchestra","description":"a laptop orchestra lork or lo is a chamber music ensemble consisting primarily of laptops like helo huddersfield experimental laptop orchestra","tag":"blue","":""}
 {"id":"G","word_rank":1,"typo_rank":0,"proximity_rank":0,"attribute_rank":0,"exact_rank":3,"asc_desc_rank":5,"title":"hello world film","description":"hello world is a 2019 japanese animated sci fi romantic drama film directed by tomohiko ito and produced by graphinica","tag":"red","":""}
 {"id":"H","word_rank":1,"typo_rank":0,"proximity_rank":1,"attribute_rank":0,"exact_rank":3,"asc_desc_rank":4,"title":"world hello day","description":"holiday observed on november 21 to express that conflicts should be resolved through communication rather than the use of force","tag":"green","":""}

milli/tests/search/distinct.rs

@@ -0,0 +1,72 @@
+use std::collections::HashSet;
+
+use big_s::S;
+use milli::update::Settings;
+use milli::{Criterion, Search, SearchResult};
+use Criterion::*;
+
+use crate::search::{self, EXTERNAL_DOCUMENTS_IDS};
+
+macro_rules! test_distinct {
+    ($func:ident, $distinct:ident, $criteria:expr) => {
+        #[test]
+        fn $func() {
+            let criteria = $criteria;
+            let index = search::setup_search_index_with_criteria(&criteria);
+
+            // update distinct attribute
+            let mut wtxn = index.write_txn().unwrap();
+            let mut builder = Settings::new(&mut wtxn, &index, 0);
+            builder.set_distinct_field(S(stringify!($distinct)));
+            builder.execute(|_, _| ()).unwrap();
+            wtxn.commit().unwrap();
+
+            let rtxn = index.read_txn().unwrap();
+
+            let mut search = Search::new(&rtxn, &index);
+            search.query(search::TEST_QUERY);
+            search.limit(EXTERNAL_DOCUMENTS_IDS.len());
+            search.authorize_typos(true);
+            search.optional_words(true);
+
+            let SearchResult { documents_ids, .. } = search.execute().unwrap();
+
+            let mut distinct_values = HashSet::new();
+            let expected_external_ids: Vec<_> = search::expected_order(&criteria, true, true)
+                .into_iter()
+                .filter_map(|d| {
+                    if distinct_values.contains(&d.$distinct) {
+                        None
+                    } else {
+                        distinct_values.insert(d.$distinct.to_owned());
+                        Some(d.id)
+                    }
+                })
+                .collect();
+
+            let documents_ids = search::internal_to_external_ids(&index, &documents_ids);
+            assert_eq!(documents_ids, expected_external_ids);
+        }
+    };
+}
+
+test_distinct!(
+    distinct_string_default_criteria,
+    tag,
+    vec![Words, Typo, Proximity, Attribute, Exactness]
+);
+test_distinct!(
+    distinct_number_default_criteria,
+    asc_desc_rank,
+    vec![Words, Typo, Proximity, Attribute, Exactness]
+);
+test_distinct!(distinct_string_criterion_words, tag, vec![Words]);
+test_distinct!(distinct_number_criterion_words, asc_desc_rank, vec![Words]);
+test_distinct!(distinct_string_criterion_words_typo, tag, vec![Words, Typo]);
+test_distinct!(distinct_number_criterion_words_typo, asc_desc_rank, vec![Words, Typo]);
+test_distinct!(distinct_string_criterion_words_proximity, tag, vec![Words, Proximity]);
+test_distinct!(distinct_number_criterion_words_proximity, asc_desc_rank, vec![Words, Proximity]);
+test_distinct!(distinct_string_criterion_words_attribute, tag, vec![Words, Attribute]);
+test_distinct!(distinct_number_criterion_words_attribute, asc_desc_rank, vec![Words, Attribute]);
+test_distinct!(distinct_string_criterion_words_exactness, tag, vec![Words, Exactness]);
+test_distinct!(distinct_number_criterion_words_exactness, asc_desc_rank, vec![Words, Exactness]);

milli/tests/search/filters.rs

@@ -0,0 +1,78 @@
+use either::{Either, Left, Right};
+use milli::{Criterion, FilterCondition, Search, SearchResult};
+use Criterion::*;
+
+use crate::search::{self, EXTERNAL_DOCUMENTS_IDS};
+
+macro_rules! test_filter {
+    ($func:ident, $filter:expr) => {
+        #[test]
+        fn $func() {
+            let criteria = vec![Words, Typo, Proximity, Attribute, Exactness];
+            let index = search::setup_search_index_with_criteria(&criteria);
+            let rtxn = index.read_txn().unwrap();
+
+            let filter_conditions =
+                FilterCondition::from_array::<Vec<Either<Vec<&str>, &str>>, _, _, _>(
+                    &rtxn, &index, $filter,
+                )
+                .unwrap()
+                .unwrap();
+
+            let mut search = Search::new(&rtxn, &index);
+            search.query(search::TEST_QUERY);
+            search.limit(EXTERNAL_DOCUMENTS_IDS.len());
+            search.authorize_typos(true);
+            search.optional_words(true);
+            search.filter(filter_conditions);
+
+            let SearchResult { documents_ids, .. } = search.execute().unwrap();
+
+            let filtered_ids = search::expected_filtered_ids($filter);
+            let expected_external_ids: Vec<_> = search::expected_order(&criteria, true, true)
+                .into_iter()
+                .filter_map(|d| if filtered_ids.contains(&d.id) { Some(d.id) } else { None })
+                .collect();
+
+            let documents_ids = search::internal_to_external_ids(&index, &documents_ids);
+            assert_eq!(documents_ids, expected_external_ids);
+        }
+    };
+}
+
+test_filter!(eq_simple_string_filter, vec![Right("tag=red")]);
+test_filter!(eq_simple_number_filter, vec![Right("asc_desc_rank=1")]);
+test_filter!(eq_string_and_filter_return_empty, vec![Right("tag=red"), Right("tag=green")]);
+test_filter!(eq_mix_and_filter, vec![Right("tag=red"), Right("asc_desc_rank=1")]);
+test_filter!(eq_string_or_filter, vec![Left(vec!["tag=red", "tag=green"])]);
+test_filter!(eq_mix_or_filter, vec![Left(vec!["tag=red", "asc_desc_rank=1"])]);
+test_filter!(eq_number_or_filter, vec![Left(vec!["asc_desc_rank=3", "asc_desc_rank=1"])]);
+test_filter!(eq_complex_filter, vec![Left(vec!["tag=red", "tag=green"]), Right("asc_desc_rank=3")]);
+test_filter!(
+    eq_complex_filter_2,
+    vec![Left(vec!["tag=red", "tag=green"]), Left(vec!["asc_desc_rank=3", "asc_desc_rank=1"])]
+);
+test_filter!(greater_simple_number_filter, vec![Right("asc_desc_rank>1")]);
+test_filter!(greater_mix_and_filter, vec![Right("tag=red"), Right("asc_desc_rank>1")]);
+test_filter!(greater_mix_or_filter, vec![Left(vec!["tag=red", "asc_desc_rank>1"])]);
+test_filter!(greater_number_or_filter, vec![Left(vec!["asc_desc_rank>3", "asc_desc_rank>1"])]);
+test_filter!(
+    greater_complex_filter,
+    vec![Left(vec!["tag=red", "tag=green"]), Right("asc_desc_rank>3")]
+);
+test_filter!(
+    greater_complex_filter_2,
+    vec![Left(vec!["tag=red", "tag=green"]), Left(vec!["asc_desc_rank>3", "asc_desc_rank>1"])]
+);
+test_filter!(lower_simple_number_filter, vec![Right("asc_desc_rank<1")]);
+test_filter!(lower_mix_and_filter, vec![Right("tag=red"), Right("asc_desc_rank<1")]);
+test_filter!(lower_mix_or_filter, vec![Left(vec!["tag=red", "asc_desc_rank<1"])]);
+test_filter!(lower_number_or_filter, vec![Left(vec!["asc_desc_rank<3", "asc_desc_rank<1"])]);
+test_filter!(
+    lower_complex_filter,
+    vec![Left(vec!["tag=red", "tag=green"]), Right("asc_desc_rank<3")]
+);
+test_filter!(
+    lower_complex_filter_2,
+    vec![Left(vec!["tag=red", "tag=green"]), Left(vec!["asc_desc_rank<3", "asc_desc_rank<1"])]
+);

milli/tests/search/mod.rs

@@ -1,4 +1,7 @@
+use std::collections::HashSet;
+
 use big_s::S;
+use either::{Either, Left, Right};
 use heed::EnvOpenOptions;
 use maplit::{hashmap, hashset};
 use milli::update::{IndexDocuments, Settings, UpdateFormat};
@@ -6,6 +9,8 @@ use milli::{Criterion, DocumentId, Index};
 use serde::Deserialize;
 use slice_group_by::GroupBy;
 
+mod distinct;
+mod filters;
 mod query_criteria;
 
 pub const TEST_QUERY: &'static str = "hello world america";
@@ -120,7 +125,58 @@ pub fn expected_order(
     }
 }
 
-#[derive(Debug, Clone, Deserialize)]
+fn execute_filter(filter: &str, document: &TestDocument) -> Option<String> {
+    let mut id = None;
+    if let Some((field, filter)) = filter.split_once("=") {
+        if field == "tag" && document.tag == filter {
+            id = Some(document.id.clone())
+        } else if field == "asc_desc_rank"
+            && document.asc_desc_rank == filter.parse::<u32>().unwrap()
+        {
+            id = Some(document.id.clone())
+        }
+    } else if let Some(("asc_desc_rank", filter)) = filter.split_once("<") {
+        if document.asc_desc_rank < filter.parse().unwrap() {
+            id = Some(document.id.clone())
+        }
+    } else if let Some(("asc_desc_rank", filter)) = filter.split_once(">") {
+        if document.asc_desc_rank > filter.parse().unwrap() {
+            id = Some(document.id.clone())
+        }
+    }
+    id
+}
+
+pub fn expected_filtered_ids(filters: Vec<Either<Vec<&str>, &str>>) -> HashSet<String> {
+    let dataset: HashSet<TestDocument> =
+        serde_json::Deserializer::from_str(CONTENT).into_iter().map(|r| r.unwrap()).collect();
+
+    let mut filtered_ids: HashSet<_> = dataset.iter().map(|d| d.id.clone()).collect();
+    for either in filters {
+        let ids = match either {
+            Left(array) => array
+                .into_iter()
+                .map(|f| {
+                    let ids: HashSet<String> =
+                        dataset.iter().filter_map(|d| execute_filter(f, d)).collect();
+                    ids
+                })
+                .reduce(|a, b| a.union(&b).cloned().collect())
+                .unwrap(),
+            Right(filter) => {
+                let ids: HashSet<String> =
+                    dataset.iter().filter_map(|d| execute_filter(filter, d)).collect();
+                ids
+            }
+        };
+
+        filtered_ids = filtered_ids.intersection(&ids).cloned().collect();
+    }
+
+    filtered_ids
+}
+
+#[derive(Debug, Clone, Deserialize, PartialEq, Eq, Hash)]
 pub struct TestDocument {
     pub id: String,
     pub word_rank: u32,

milli/tests/search/query_criteria.rs

@@ -11,14 +11,14 @@ const ALLOW_OPTIONAL_WORDS: bool = true;
 const DISALLOW_OPTIONAL_WORDS: bool = false;
 
 macro_rules! test_criterion {
-    ($func:ident, $optional_word:ident, $authorize_typos:ident $(, $criterion:expr)?) => {
+    ($func:ident, $optional_word:ident, $authorize_typos:ident, $criteria:expr) => {
         #[test]
         fn $func() {
-            let criteria = vec![$($criterion)?];
+            let criteria = $criteria;
             let index = search::setup_search_index_with_criteria(&criteria);
-            let mut rtxn = index.read_txn().unwrap();
+            let rtxn = index.read_txn().unwrap();
 
-            let mut search = Search::new(&mut rtxn, &index);
+            let mut search = Search::new(&rtxn, &index);
             search.query(search::TEST_QUERY);
             search.limit(EXTERNAL_DOCUMENTS_IDS.len());
             search.authorize_typos($authorize_typos);
@@ -26,49 +26,80 @@ macro_rules! test_criterion {
 
             let SearchResult { documents_ids, .. } = search.execute().unwrap();
 
-            let expected_external_ids: Vec<_> = search::expected_order(&criteria, $authorize_typos, $optional_word)
-                .into_iter()
-                .map(|d| d.id).collect();
+            let expected_external_ids: Vec<_> =
+                search::expected_order(&criteria, $authorize_typos, $optional_word)
+                    .into_iter()
+                    .map(|d| d.id)
+                    .collect();
             let documents_ids = search::internal_to_external_ids(&index, &documents_ids);
             assert_eq!(documents_ids, expected_external_ids);
         }
-    }
+    };
 }
 
-#[rustfmt::skip]
-test_criterion!(none_allow_typo,                     ALLOW_OPTIONAL_WORDS,      ALLOW_TYPOS);
-#[rustfmt::skip]
-test_criterion!(none_disallow_typo,                  DISALLOW_OPTIONAL_WORDS,   DISALLOW_TYPOS);
-#[rustfmt::skip]
-test_criterion!(words_allow_typo,                    ALLOW_OPTIONAL_WORDS,      ALLOW_TYPOS,    Words);
-#[rustfmt::skip]
-test_criterion!(attribute_allow_typo,                DISALLOW_OPTIONAL_WORDS,   ALLOW_TYPOS,    Attribute);
-#[rustfmt::skip]
-test_criterion!(attribute_disallow_typo,             DISALLOW_OPTIONAL_WORDS,   DISALLOW_TYPOS, Attribute);
-#[rustfmt::skip]
-test_criterion!(exactness_allow_typo,                DISALLOW_OPTIONAL_WORDS,   ALLOW_TYPOS,    Exactness);
-#[rustfmt::skip]
-test_criterion!(exactness_disallow_typo,             DISALLOW_OPTIONAL_WORDS,   DISALLOW_TYPOS, Exactness);
-#[rustfmt::skip]
-test_criterion!(proximity_allow_typo,                DISALLOW_OPTIONAL_WORDS,   ALLOW_TYPOS,    Proximity);
-#[rustfmt::skip]
-test_criterion!(proximity_disallow_typo,             DISALLOW_OPTIONAL_WORDS,   DISALLOW_TYPOS, Proximity);
-#[rustfmt::skip]
-test_criterion!(asc_allow_typo,                      DISALLOW_OPTIONAL_WORDS,   ALLOW_TYPOS,    Asc(S("asc_desc_rank")));
-#[rustfmt::skip]
-test_criterion!(asc_disallow_typo,                   DISALLOW_OPTIONAL_WORDS,   DISALLOW_TYPOS, Asc(S("asc_desc_rank")));
-#[rustfmt::skip]
-test_criterion!(desc_allow_typo,                     DISALLOW_OPTIONAL_WORDS,   ALLOW_TYPOS,    Desc(S("asc_desc_rank")));
-#[rustfmt::skip]
-test_criterion!(desc_disallow_typo,                  DISALLOW_OPTIONAL_WORDS,   DISALLOW_TYPOS, Desc(S("asc_desc_rank")));
-#[rustfmt::skip]
-test_criterion!(asc_unexisting_field_allow_typo,     DISALLOW_OPTIONAL_WORDS,   ALLOW_TYPOS,    Asc(S("unexisting_field")));
-#[rustfmt::skip]
-test_criterion!(asc_unexisting_field_disallow_typo,  DISALLOW_OPTIONAL_WORDS,   DISALLOW_TYPOS, Asc(S("unexisting_field")));
-#[rustfmt::skip]
-test_criterion!(desc_unexisting_field_allow_typo,    DISALLOW_OPTIONAL_WORDS,   ALLOW_TYPOS,    Desc(S("unexisting_field")));
-#[rustfmt::skip]
-test_criterion!(desc_unexisting_field_disallow_typo, DISALLOW_OPTIONAL_WORDS,   DISALLOW_TYPOS, Desc(S("unexisting_field")));
+test_criterion!(none_allow_typo, ALLOW_OPTIONAL_WORDS, ALLOW_TYPOS, vec![]);
+test_criterion!(none_disallow_typo, DISALLOW_OPTIONAL_WORDS, DISALLOW_TYPOS, vec![]);
+test_criterion!(words_allow_typo, ALLOW_OPTIONAL_WORDS, ALLOW_TYPOS, vec![Words]);
+test_criterion!(attribute_allow_typo, DISALLOW_OPTIONAL_WORDS, ALLOW_TYPOS, vec![Attribute]);
+test_criterion!(attribute_disallow_typo, DISALLOW_OPTIONAL_WORDS, DISALLOW_TYPOS, vec![Attribute]);
+test_criterion!(exactness_allow_typo, DISALLOW_OPTIONAL_WORDS, ALLOW_TYPOS, vec![Exactness]);
+test_criterion!(exactness_disallow_typo, DISALLOW_OPTIONAL_WORDS, DISALLOW_TYPOS, vec![Exactness]);
+test_criterion!(proximity_allow_typo, DISALLOW_OPTIONAL_WORDS, ALLOW_TYPOS, vec![Proximity]);
+test_criterion!(proximity_disallow_typo, DISALLOW_OPTIONAL_WORDS, DISALLOW_TYPOS, vec![Proximity]);
+test_criterion!(
+    asc_allow_typo,
+    DISALLOW_OPTIONAL_WORDS,
+    ALLOW_TYPOS,
+    vec![Asc(S("asc_desc_rank"))]
+);
+test_criterion!(
+    asc_disallow_typo,
+    DISALLOW_OPTIONAL_WORDS,
+    DISALLOW_TYPOS,
+    vec![Asc(S("asc_desc_rank"))]
+);
+test_criterion!(
+    desc_allow_typo,
+    DISALLOW_OPTIONAL_WORDS,
+    ALLOW_TYPOS,
+    vec![Desc(S("asc_desc_rank"))]
+);
+test_criterion!(
+    desc_disallow_typo,
+    DISALLOW_OPTIONAL_WORDS,
+    DISALLOW_TYPOS,
+    vec![Desc(S("asc_desc_rank"))]
+);
+test_criterion!(
+    asc_unexisting_field_allow_typo,
+    DISALLOW_OPTIONAL_WORDS,
+    ALLOW_TYPOS,
+    vec![Asc(S("unexisting_field"))]
+);
+test_criterion!(
+    asc_unexisting_field_disallow_typo,
+    DISALLOW_OPTIONAL_WORDS,
+    DISALLOW_TYPOS,
+    vec![Asc(S("unexisting_field"))]
+);
+test_criterion!(
+    desc_unexisting_field_allow_typo,
+    DISALLOW_OPTIONAL_WORDS,
+    ALLOW_TYPOS,
+    vec![Desc(S("unexisting_field"))]
+);
+test_criterion!(
+    desc_unexisting_field_disallow_typo,
+    DISALLOW_OPTIONAL_WORDS,
+    DISALLOW_TYPOS,
+    vec![Desc(S("unexisting_field"))]
+);
+test_criterion!(
+    default_criteria_order,
+    ALLOW_OPTIONAL_WORDS,
+    ALLOW_TYPOS,
+    vec![Words, Typo, Proximity, Attribute, Exactness]
+);
 
 #[test]
 fn criteria_mixup() {