bugfix tantivy query syntax error

c7nw3r · Sep 6, 2023 · 1f87b8f · 1f87b8f
1 parent 16a11ea
commit 1f87b8f
Showing 1 changed file with 16 additions and 1 deletion.
diff --git a/localsearch/searcher/tantivy_search.py b/localsearch/searcher/tantivy_search.py
@@ -94,7 +94,22 @@ def _canonicalize(self, text: str):
         text = remove_punctuation(text)
         text = remove_stopwords(text, self.config.lang)
         text = lemmatize(text, self.config.lang)
-        text = '"' + text + '"'
+        text = text.replace("+", " ")
+        text = text.replace("-", " ")
+        text = text.replace("^", " ")
+        text = text.replace("`", " ")
+        text = text.replace(":", " ")
+        text = text.replace("{", " ")
+        text = text.replace("}", " ")
+        text = text.replace('"', " ")
+        text = text.replace("[", " ")
+        text = text.replace("]", " ")
+        text = text.replace("(", " ")
+        text = text.replace(")", " ")
+        text = text.replace("~", " ")
+        text = text.replace("!", " ")
+        text = text.replace("*", " ")
+        text = text.replace("\\", " ")
         return text
 
     def search_by_source(self, source: str, n: Optional[int] = None) -> List[Document]: