Merge pull request #213 from marqo-ai/space_filtering_fix

Space filtering fix
marqo-ai · Dec 8, 2022 · bb54a78 · bb54a78
2 parents 0b31f02 + 000443c
commit bb54a78
Show file tree

Hide file tree

Showing 2 changed files with 34 additions and 1 deletion.
diff --git a/src/marqo/tensor_search/utils.py b/src/marqo/tensor_search/utils.py
@@ -99,7 +99,11 @@ def contextualise_filter(filter_string: str, simple_properties: typing.Iterable)
     """
     contextualised_filter = filter_string
     for field in simple_properties:
-        contextualised_filter = contextualised_filter.replace(f'{field}:', f'{enums.TensorField.chunks}.{field}:')
+        if ' ' in field:
+            field_with_escaped_space = field.replace(' ', '\ ')
+            contextualised_filter = contextualised_filter.replace(f'{field_with_escaped_space}:', f'{enums.TensorField.chunks}.{field_with_escaped_space}:')
+        else:
+            contextualised_filter = contextualised_filter.replace(f'{field}:', f'{enums.TensorField.chunks}.{field}:')
     return contextualised_filter
 
 

diff --git a/tests/tensor_search/test_search.py b/tests/tensor_search/test_search.py
@@ -382,6 +382,35 @@ def test_filtering(self):
             filter="*:*", verbose=1
         )["hits"])
 
+    def test_filter_spaced_fields(self):
+        tensor_search.add_documents(
+            config=self.config, index_name=self.index_name_1, docs=[
+                {"abc": "some text", "other field": "baaadd", "_id": "5678", "my_string": "b"},
+                {"abc": "some text", "other field": "Close match hehehe", "_id": "1234", "an_int": 2},
+                {"abc": "some text", "other field": "Close match hehehe", "_id": "1233", "my_bool": True},
+                {"abc": "some text", "Floaty Field": 0.548, "_id": "344", "my_bool": True},
+            ], auto_refresh=True)
+
+        res = tensor_search.search(
+            config=self.config, index_name=self.index_name_1, text='', filter="other\ field:baaadd")
+
+        assert len(res['hits']) == 1
+        assert res['hits'][0]['_id'] == "5678"
+
+        res_mult = tensor_search.search(
+            config=self.config, index_name=self.index_name_1, text='', filter="other\ field:(Close match hehehe)")
+        assert len(res_mult['hits']) == 2
+        assert res_mult['hits'][0]['_id'] in {'1234', '1233'}
+        assert res_mult['hits'][1]['_id'] in {'1234', '1233'}
+        assert res_mult['hits'][1]['_id'] != res_mult['hits'][0]['_id']
+
+        res_float = tensor_search.search(
+            config=self.config, index_name=self.index_name_1, text='', filter="(Floaty\ Field:[0 TO 1]) AND (abc:(some text))")
+        get_res = tensor_search.get_document_by_id(config=self.config, index_name=self.index_name_1, document_id='344')
+
+        assert len(res_float['hits']) == 1
+        assert res_float['hits'][0]['_id'] == "344"
+
     def test_filtering_bad_syntax(self):
         tensor_search.add_documents(
             config=self.config, index_name=self.index_name_1, docs=[