From 8b1e5d9c6d654be95159519e8c233e8868694e1b Mon Sep 17 00:00:00 2001
From: ad hoc <postma.marin@protonmail.com>
Date: Tue, 22 Mar 2022 09:55:49 +0100
Subject: [PATCH] add test for exact words

---
 milli/src/search/query_tree.rs | 25 ++++++++++++++++++++++---
 milli/src/update/settings.rs   |  9 ++++-----
 2 files changed, 26 insertions(+), 8 deletions(-)

diff --git a/milli/src/search/query_tree.rs b/milli/src/search/query_tree.rs
index a31a71590..0014075d4 100644
--- a/milli/src/search/query_tree.rs
+++ b/milli/src/search/query_tree.rs
@@ -584,6 +584,8 @@ mod test {
     struct TestContext {
         synonyms: HashMap<Vec<String>, Vec<Vec<String>>>,
         postings: HashMap<String, RoaringBitmap>,
+        // Raw bytes for the exact word fst Set
+        exact_words: Vec<u8>,
     }
 
     impl TestContext {
@@ -620,9 +622,7 @@ mod test {
         }
 
         fn exact_words(&self) -> crate::Result<fst::Set<Cow<[u8]>>> {
-            let builder = fst::SetBuilder::new(Vec::new()).unwrap();
-            let data = builder.into_inner().unwrap();
-            Ok(fst::Set::new(Cow::Owned(data)).unwrap())
+            Ok(fst::Set::new(Cow::Borrowed(self.exact_words.as_slice())).unwrap())
         }
     }
 
@@ -640,6 +640,8 @@ mod test {
                 RoaringBitmap::from_sorted_iter(values.into_iter()).unwrap()
             }
 
+            let exact_words = fst::SetBuilder::new(Vec::new()).unwrap().into_inner().unwrap();
+
             TestContext {
                 synonyms: hashmap! {
                     vec![String::from("hello")] => vec![
@@ -679,6 +681,7 @@ mod test {
                     String::from("good")       => random_postings(rng,   1250),
                     String::from("morning")    => random_postings(rng,    125),
                 },
+                exact_words,
             }
         }
     }
@@ -1263,4 +1266,20 @@ mod test {
             QueryKind::Tolerant { typo: 2, word: "verylongword".to_string() }
         );
     }
+    #[test]
+    fn disable_typo_on_word() {
+        let query = "goodbye";
+        let analyzer = Analyzer::new(AnalyzerConfig::<Vec<u8>>::default());
+        let result = analyzer.analyze(query);
+
+        let tokens = result.tokens();
+        let exact_words = fst::Set::from_iter(Some("goodbye")).unwrap().into_fst().into_inner();
+        let context = TestContext { exact_words, ..Default::default() };
+        let (query_tree, _) = context.build(false, true, Some(2), tokens).unwrap().unwrap();
+
+        assert!(matches!(
+            query_tree,
+            Operation::Query(Query { prefix: true, kind: QueryKind::Exact { .. } })
+        ));
+    }
 }
diff --git a/milli/src/update/settings.rs b/milli/src/update/settings.rs
index 513dee42c..503fbd06e 100644
--- a/milli/src/update/settings.rs
+++ b/milli/src/update/settings.rs
@@ -92,7 +92,7 @@ pub struct Settings<'a, 't, 'u, 'i> {
     authorize_typos: Setting<bool>,
     min_word_len_two_typos: Setting<u8>,
     min_word_len_one_typo: Setting<u8>,
-    exact_words: Setting<Vec<String>>,
+    exact_words: Setting<BTreeSet<String>>,
 }
 
 impl<'a, 't, 'u, 'i> Settings<'a, 't, 'u, 'i> {
@@ -115,9 +115,9 @@ impl<'a, 't, 'u, 'i> Settings<'a, 't, 'u, 'i> {
             primary_key: Setting::NotSet,
             authorize_typos: Setting::NotSet,
             exact_words: Setting::NotSet,
-            indexer_config,
             min_word_len_two_typos: Setting::Reset,
             min_word_len_one_typo: Setting::Reset,
+            indexer_config,
         }
     }
 
@@ -218,7 +218,7 @@ impl<'a, 't, 'u, 'i> Settings<'a, 't, 'u, 'i> {
         self.min_word_len_one_typo = Setting::Reset;
     }
 
-    pub fn set_exact_words(&mut self, words: Vec<String>) {
+    pub fn set_exact_words(&mut self, words: BTreeSet<String>) {
         self.exact_words = Setting::Set(words);
     }
 
@@ -539,8 +539,7 @@ impl<'a, 't, 'u, 'i> Settings<'a, 't, 'u, 'i> {
     fn update_exact_words(&mut self) -> Result<()> {
         match self.exact_words {
             Setting::Set(ref mut words) => {
-                words.sort_unstable();
-                let words = fst::Set::from_iter(words)?;
+                let words = fst::Set::from_iter(words.iter())?;
                 self.index.put_exact_words(&mut self.wtxn, &words)?;
             }
             Setting::Reset => {