Merge pull request #4834 from freelawproject/4707-case-name-search-bug

4707 Fixed Case name filter search bug
freelawproject · Dec 17, 2024 · b0764a4 · b0764a4
2 parents 88924a3 + 336869f
commit b0764a4
Show file tree

Hide file tree

Showing 3 changed files with 40 additions and 25 deletions.
diff --git a/cl/lib/elasticsearch_utils.py b/cl/lib/elasticsearch_utils.py
@@ -395,7 +395,13 @@ def build_fulltext_query(
             q_should.append(
                 Q(
                     "match_phrase",
-                    caseName={"query": query_value, "boost": 2, "slop": 1},
+                    **{
+                        "caseName.exact": {
+                            "query": query_value,
+                            "boost": 2,
+                            "slop": 1,
+                        }
+                    },
                 )
             )
 
@@ -481,26 +487,14 @@ def build_text_filter(field: str, value: str) -> List:
         if isinstance(value, str):
             validate_query_syntax(value, QueryType.FILTER)
 
-        base_query_string = {
-            "query_string": {
-                "query": value,
-                "fields": [field],
-                "default_operator": "AND",
-            }
-        }
-        if "caseName" in field and '"' not in value:
-            # Use phrase with slop, and give it a boost to prioritize the
-            # phrase match to ensure that caseName filtering returns exactly
-            # this order as the first priority.
-            # Avoid applying slop to quoted queries, as they expect exact matches.
-            base_query_string["query_string"].update(
-                {
-                    "type": "phrase",
-                    "phrase_slop": "1",
-                    "boost": "2",  # Boosting the phrase match to ensure it's ranked higher than individual term matches
-                }
+        return [
+            Q(
+                "query_string",
+                query=value,
+                fields=[field],
+                default_operator="AND",
             )
-        return [Q(base_query_string)]
+        ]
     return []
 
 

diff --git a/cl/search/tests/tests_es_opinion.py b/cl/search/tests/tests_es_opinion.py
@@ -2231,7 +2231,8 @@ def test_uses_exact_version_for_case_name_field(self) -> None:
         r = async_to_sync(self._test_article_count)(
             search_params, 1, "case_name exact filter"
         )
-        self.assertIn("<mark>Maecenas Howell</mark>", r.content.decode())
+        self.assertIn("<mark>Maecenas</mark>", r.content.decode())
+        self.assertIn("<mark>Howell</mark>", r.content.decode())
 
         # case_name filter: Howells
         search_params = {
@@ -2241,7 +2242,8 @@ def test_uses_exact_version_for_case_name_field(self) -> None:
         r = async_to_sync(self._test_article_count)(
             search_params, 1, "case_name exact filter"
         )
-        self.assertIn("<mark>Maecenas Howells</mark>", r.content.decode())
+        self.assertIn("<mark>Maecenas</mark>", r.content.decode())
+        self.assertIn("<mark>Howells</mark>", r.content.decode())
 
         # text query: Howell
         search_params = {

diff --git a/cl/search/tests/tests_es_recap.py b/cl/search/tests/tests_es_recap.py
@@ -1795,8 +1795,9 @@ async def test_results_highlights(self) -> None:
             "case_name": "SUBPOENAS SERVED ON",
         }
         r = await self._test_article_count(params, 1, "highlights caseName")
-        # Count child documents under docket.
-        self.assertIn("<mark>SUBPOENAS SERVED ON</mark>", r.content.decode())
+        self.assertIn("<mark>SUBPOENAS</mark>", r.content.decode())
+        self.assertIn("<mark>SERVED</mark>", r.content.decode())
+        self.assertIn("<mark>ON</mark>", r.content.decode())
 
         # Highlight filter: description
         params = {
@@ -2806,6 +2807,24 @@ def test_uses_exact_version_for_case_name_field(self) -> None:
         r = async_to_sync(self._test_article_count)(cd, 1, "Disable stemming")
         self.assertIn("<mark>Howells v. Indiana</mark>", r.content.decode())
 
+        # No quoted case_name filter: A match for 'Indiana Howell' is expected,
+        # as the order is not mandatory
+        cd = {
+            "type": SEARCH_TYPES.RECAP,
+            "case_name": "Indiana Howell",
+        }
+        r = async_to_sync(self._test_article_count)(cd, 1, "No phrase search.")
+        self.assertIn("<mark>Howell</mark>", r.content.decode())
+        self.assertIn("<mark>Indiana</mark>", r.content.decode())
+
+        # Quoted case_name filter: "Indiana v. Howells" match is not expected
+        # as order is mandatory
+        cd = {
+            "type": SEARCH_TYPES.RECAP,
+            "case_name": '"Indiana v. Howells"',
+        }
+        async_to_sync(self._test_article_count)(cd, 0, "Phrase search.")
+
         # text query: Howell
         cd = {
             "type": SEARCH_TYPES.RECAP,
@@ -3718,7 +3737,7 @@ async def test_results_api_highlighted_fields(self) -> None:
             "result": self.rd_api,
             "V4": True,
             "assignedTo": "<mark>George</mark> Doe II",
-            "caseName": "<mark>America vs API</mark> Lorem",
+            "caseName": "<mark>America</mark> <mark>vs</mark> <mark>API</mark> Lorem",
             "cause": "<mark>401</mark> <mark>Civil</mark>",
             "court_citation_string": "<mark>Appeals</mark>. CA9.",
             "docketNumber": "<mark>1:24-bk-0000</mark>",