llama-index-vector-stores-opensearch 0.1.9__tar.gz → 0.1.11__tar.gz

{llama_index_vector_stores_opensearch-0.1.9 → llama_index_vector_stores_opensearch-0.1.11}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: llama-index-vector-stores-opensearch
-Version: 0.1.9
+Version: 0.1.11
 Summary: llama-index vector_stores opensearch integration
 License: MIT
 Author: Your Name
@@ -11,6 +11,7 @@ Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.9
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
 Requires-Dist: llama-index-core (>=0.10.1,<0.11.0)
 Requires-Dist: opensearch-py[async] (>=2.4.2,<3.0.0)
 Description-Content-Type: text/markdown

{llama_index_vector_stores_opensearch-0.1.9 → llama_index_vector_stores_opensearch-0.1.11}/llama_index/vector_stores/opensearch/base.py

@@ -239,12 +239,12 @@ class OpensearchVectorClient:
         Returns:
             Up to k docs closest to query_embedding
         """
-        if filters is None:
+        pre_filter = self._parse_filters(filters)
+        if not pre_filter:
             search_query = self._default_approximate_search_query(
                 query_embedding, k, vector_field=embedding_field
             )
         else:
-            pre_filter = self._parse_filters(filters)
             # https://opensearch.org/docs/latest/search-plugins/knn/painless-functions/
             search_query = self._default_painless_scripting_query(
                 query_embedding,
@@ -265,17 +265,34 @@ class OpensearchVectorClient:
         k: int,
         filters: Optional[MetadataFilters] = None,
     ) -> Dict:
-        knn_query = self._knn_search_query(
-            embedding_field, query_embedding, k, filters
-        )["query"]
-        lexical_query = {"must": {"match": {text_field: {"query": query_str}}}}
+        knn_query = self._knn_search_query(embedding_field, query_embedding, k, filters)
+        lexical_query = self._lexical_search_query(text_field, query_str, k, filters)
+
+        return {
+            "size": k,
+            "query": {
+                "hybrid": {"queries": [lexical_query["query"], knn_query["query"]]}
+            },
+        }
+
+    def _lexical_search_query(
+        self,
+        text_field: str,
+        query_str: str,
+        k: int,
+        filters: Optional[MetadataFilters] = None,
+    ) -> Dict:
+        lexical_query = {
+            "bool": {"must": {"match": {text_field: {"query": query_str}}}}
+        }
 
         parsed_filters = self._parse_filters(filters)
         if len(parsed_filters) > 0:
-            lexical_query["filter"] = parsed_filters
+            lexical_query["bool"]["filter"] = parsed_filters
+
         return {
             "size": k,
-            "query": {"hybrid": {"queries": [{"bool": lexical_query}, knn_query]}},
+            "query": lexical_query,
         }
 
     def __get_painless_scripting_source(
@@ -388,17 +405,25 @@ class OpensearchVectorClient:
             )
             params = {
                 "search_pipeline": self._search_pipeline,
-                "_source_excludes": ["embedding"],
             }
+        elif query_mode == VectorStoreQueryMode.TEXT_SEARCH:
+            search_query = self._lexical_search_query(
+                self._text_field, query_str, k, filters=filters
+            )
+            params = None
         else:
             search_query = self._knn_search_query(
                 self._embedding_field, query_embedding, k, filters=filters
             )
-            params = {"_source_excludes": ["embedding"]}
+            params = None
 
         res = await self._os_client.search(
             index=self._index, body=search_query, params=params
         )
+
+        return self._to_query_result(res)
+
+    def _to_query_result(self, res) -> VectorStoreQueryResult:
         nodes = []
         ids = []
         scores = []
@@ -433,6 +458,7 @@ class OpensearchVectorClient:
             ids.append(node_id)
             nodes.append(node)
             scores.append(hit["_score"])
+
         return VectorStoreQueryResult(nodes=nodes, ids=ids, similarities=scores)
 
 

{llama_index_vector_stores_opensearch-0.1.9 → llama_index_vector_stores_opensearch-0.1.11}/pyproject.toml

@@ -27,7 +27,7 @@ exclude = ["**/BUILD"]
 license = "MIT"
 name = "llama-index-vector-stores-opensearch"
 readme = "README.md"
-version = "0.1.9"
+version = "0.1.11"
 
 [tool.poetry.dependencies]
 python = ">=3.8.1,<4.0"