From 9bbffb8fee9ab73fb59eab731f1e739c85e536dd Mon Sep 17 00:00:00 2001
From: ad hoc <postma.marin@protonmail.com>
Date: Mon, 21 Mar 2022 14:03:31 +0100
Subject: [PATCH] add exact words setting

---
 milli/src/index.rs           | 22 ++++++++++++++++++++++
 milli/src/update/settings.rs | 27 +++++++++++++++++++++++++++
 2 files changed, 49 insertions(+)

diff --git a/milli/src/index.rs b/milli/src/index.rs
index 853e7537d..c0be985da 100644
--- a/milli/src/index.rs
+++ b/milli/src/index.rs
@@ -52,6 +52,7 @@ pub mod main_key {
     pub const AUTHORIZE_TYPOS: &str = "authorize-typos";
     pub const ONE_TYPO_WORD_LEN: &str = "one-typo-word-len";
     pub const TWO_TYPOS_WORD_LEN: &str = "two-typos-word-len";
+    pub const EXACT_WORDS: &str = "exact-words";
 }
 
 pub mod db_name {
@@ -927,6 +928,27 @@ impl Index {
         self.main.put::<_, Str, OwnedType<u8>>(txn, main_key::TWO_TYPOS_WORD_LEN, &val)?;
         Ok(())
     }
+
+    /// List the words on which typo are not allowed
+    pub fn exact_words<'t>(&self, txn: &'t RoTxn) -> Result<fst::Set<Cow<'t, [u8]>>> {
+        match self.main.get::<_, Str, ByteSlice>(txn, main_key::EXACT_WORDS)? {
+            Some(bytes) => Ok(fst::Set::new(bytes)?.map_data(Cow::Borrowed)?),
+            None => Ok(fst::Set::default().map_data(Cow::Owned)?),
+        }
+    }
+
+    pub(crate) fn put_exact_words<A: AsRef<[u8]>>(
+        &self,
+        txn: &mut RwTxn,
+        words: &fst::Set<A>,
+    ) -> Result<()> {
+        self.main.put::<_, Str, ByteSlice>(
+            txn,
+            main_key::EXACT_WORDS,
+            words.as_fst().as_bytes(),
+        )?;
+        Ok(())
+    }
 }
 
 #[cfg(test)]
diff --git a/milli/src/update/settings.rs b/milli/src/update/settings.rs
index c03d6e0ae..513dee42c 100644
--- a/milli/src/update/settings.rs
+++ b/milli/src/update/settings.rs
@@ -92,6 +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>>,
 }
 
 impl<'a, 't, 'u, 'i> Settings<'a, 't, 'u, 'i> {
@@ -113,6 +114,7 @@ impl<'a, 't, 'u, 'i> Settings<'a, 't, 'u, 'i> {
             synonyms: Setting::NotSet,
             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,
@@ -216,6 +218,14 @@ 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>) {
+        self.exact_words = Setting::Set(words);
+    }
+
+    pub fn reset_exact_words(&mut self) {
+        self.exact_words = Setting::Reset;
+    }
+
     fn reindex<F>(&mut self, cb: &F, old_fields_ids_map: FieldsIdsMap) -> Result<()>
     where
         F: Fn(UpdateIndexingStep) + Sync,
@@ -526,6 +536,22 @@ impl<'a, 't, 'u, 'i> Settings<'a, 't, 'u, 'i> {
         Ok(())
     }
 
+    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)?;
+                self.index.put_exact_words(&mut self.wtxn, &words)?;
+            }
+            Setting::Reset => {
+                self.index.put_exact_words(&mut self.wtxn, &fst::Set::default())?;
+            }
+            Setting::NotSet => (),
+        }
+
+        Ok(())
+    }
+
     pub fn execute<F>(mut self, progress_callback: F) -> Result<()>
     where
         F: Fn(UpdateIndexingStep) + Sync,
@@ -543,6 +569,7 @@ impl<'a, 't, 'u, 'i> Settings<'a, 't, 'u, 'i> {
         self.update_primary_key()?;
         self.update_authorize_typos()?;
         self.update_min_typo_word_len()?;
+        self.update_exact_words()?;
 
         // 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,