Method test_case_7

test/test_index_management.py:300–362

Driver handles index types.

(self)

        self.assertEqual(index["latestDefinition"], model["definition"])

    def test_case_7(self):
        """Driver handles index types."""

        # Create a collection with the "create" command using a randomly generated name (referred to as ``coll0``).
        coll0 = self.db[f"col{uuid.uuid4()}"]
        coll0.insert_one({})

        # Use these search and vector search definitions for indexes.
        search_definition = {"mappings": {"dynamic": False}}
        vector_search_definition = {
            "fields": [
                {
                    "type": "vector",
                    "path": "plot_embedding",
                    "numDimensions": 1536,
                    "similarity": "euclidean",
                },
            ]
        }

        # Create a new search index on ``coll0`` that implicitly passes its type.
        implicit_search_resp = coll0.create_search_index(
            model={"name": _NAME + "-implicit", "definition": search_definition}
        )

        # Get the index definition.
        resp = (coll0.list_search_indexes(name=implicit_search_resp)).next()

        # Assert that the index model contains the correct index type: ``"search"``.
        self.assertEqual(resp["type"], "search")

        # Create a new search index on ``coll0`` that explicitly passes its type.
        explicit_search_resp = coll0.create_search_index(
            model={"name": _NAME + "-explicit", "type": "search", "definition": search_definition}
        )

        # Get the index definition.
        resp = (coll0.list_search_indexes(name=explicit_search_resp)).next()

        # Assert that the index model contains the correct index type: ``"search"``.
        self.assertEqual(resp["type"], "search")

        # Create a new vector search index on ``coll0`` that explicitly passes its type.
        explicit_vector_resp = coll0.create_search_index(
            model={
                "name": _NAME + "-vector",
                "type": "vectorSearch",
                "definition": vector_search_definition,
            }
        )

        # Get the index definition.
        resp = (coll0.list_search_indexes(name=explicit_vector_resp)).next()

        # Assert that the index model contains the correct index type: ``"vectorSearch"``.
        self.assertEqual(resp["type"], "vectorSearch")

        # Catch the error raised when trying to create a vector search index without specifying the type