Fix proximity search with quotation

UCL · Oct 4, 2024 · be1d4dd · be1d4dd
1 parent 1c2e186
commit be1d4dd
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 3 deletions.
diff --git a/src/rard/research/tests/views/test_search.py b/src/rard/research/tests/views/test_search.py
@@ -484,6 +484,14 @@ def do_search(search_function, keywords):
         self.assertEqual(do_search(view.fragment_search, "Lorem ~3:5 sit"), [])
         self.assertEqual(do_search(view.fragment_search, "Lorem ~3:5 elit"), [])
 
+        # Test wildcards work in combination
+        self.assertEqual(do_search(view.fragment_search, "Lor*m ~3:5 consect*"), [f1])
+
+        # Test quotation works
+        self.assertEqual(
+            do_search(view.fragment_search, 'Lorem ~3:5 "consectetur adipiscing"'), [f1]
+        )
+
     def test_search_snippets(self):
         raw_content = (
             "Lorem ipsum dolor sit amet, <span class='test consectatur'>"

diff --git a/src/rard/research/views/search.py b/src/rard/research/views/search.py
@@ -217,10 +217,10 @@ def transform_keywords_to_regex(self, keywords):
             """
             # Proximity searches have one keyword and contain tilde character
             if len(keywords) == 1 and "~" in keywords[0]:
+                # Remove any '"' characters
+                kw = keywords[0].replace('"', "")
                 # Split keyword around proximity search
-                [(fore, prox_op, aft)] = re.findall(
-                    r"(.*)\s(~\d?:?\d?)\s(.*)", keywords[0]
-                )
+                [(fore, prox_op, aft)] = re.findall(r"(.*)\s(~\d?:?\d?)\s(.*)", kw)
                 # Fore and aft can be multi-word strings containing wildcards so loop back
                 fore = self.transform_keywords_to_regex([fore])
                 aft = self.transform_keywords_to_regex([aft])