alita-sdk 0.3.218__tar.gz → 0.3.220__tar.gz

{alita_sdk-0.3.218 → alita_sdk-0.3.220}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: alita_sdk
-Version: 0.3.218
+Version: 0.3.220
 Summary: SDK for building langchain agents using resources from Alita
 Author-email: Artem Rozumenko <artyom.rozumenko@gmail.com>, Mikalai Biazruchka <mikalai_biazruchka@epam.com>, Roman Mitusov <roman_mitusov@epam.com>, Ivan Krakhmaliuk <lifedjik@gmail.com>, Artem Dubrovskiy <ad13box@gmail.com>
 License-Expression: Apache-2.0

{alita_sdk-0.3.218 → alita_sdk-0.3.220}/alita_sdk/runtime/tools/vectorstore.py

@@ -188,6 +188,59 @@ class VectorStoreWrapper(BaseToolApiWrapper):
             except Exception as e:
                 logger.error(f"Failed to initialize PGVectorSearch: {str(e)}")
 
+    def _remove_collection(self):
+        """
+        Remove the vectorstore collection entirely.
+        """
+        self._log_data(
+            f"Remove collection '{self.dataset}'",
+            tool_name="_remove_collection"
+        )
+        from sqlalchemy import text
+        from sqlalchemy.orm import Session
+
+        schema_name = self.vectorstore.collection_name
+        with Session(self.vectorstore.session_maker.bind) as session:
+            drop_schema_query = text(f"DROP SCHEMA IF EXISTS {schema_name} CASCADE;")
+            session.execute(drop_schema_query)
+            session.commit()
+            logger.info(f"Schema '{schema_name}' has been dropped.")
+        self._log_data(
+            f"Collection '{self.dataset}' has been removed. ",
+            tool_name="_remove_collection"
+        )
+
+    def _get_indexed_ids(self, store):
+        """Get all indexed document IDs from vectorstore"""
+
+        # Check if this is a PGVector store
+        if hasattr(store, 'session_maker') and hasattr(store, 'EmbeddingStore'):
+            return self._get_pgvector_indexed_ids(store)
+        else:
+            # Fall back to Chroma implementation
+            return self._get_chroma_indexed_ids(store)
+
+    def _get_pgvector_indexed_ids(self, store):
+        """Get all indexed document IDs from PGVector"""
+        from sqlalchemy.orm import Session
+
+        try:
+            with Session(store.session_maker.bind) as session:
+                ids = session.query(store.EmbeddingStore.id).all()
+            return [str(id_tuple[0]) for id_tuple in ids]
+        except Exception as e:
+            logger.error(f"Failed to get indexed IDs from PGVector: {str(e)}")
+            return []
+
+    def _get_chroma_indexed_ids(self, store):
+        """Get all indexed document IDs from Chroma"""
+        try:
+            data = store.get(include=[])  # Only get IDs, no metadata
+            return data.get('ids', [])
+        except Exception as e:
+            logger.error(f"Failed to get indexed IDs from Chroma: {str(e)}")
+            return []
+
     def _clean_collection(self):
         """
         Clean the vectorstore collection by deleting all indexed data.
@@ -196,13 +249,10 @@ class VectorStoreWrapper(BaseToolApiWrapper):
             f"Cleaning collection '{self.dataset}'",
             tool_name="_clean_collection"
         )
-        # This logic deletes the entire collection
-        # Works for PGVector and Chroma
-        self.vectoradapter.vectorstore.delete_collection()
         # This logic deletes all data from the vectorstore collection without removal of collection.
-        # data = self.vectoradapter.vectorstore.get(include=['metadatas'])
-        # if data['ids']:
-        #     self.vectoradapter.vectorstore.delete(ids=data['ids'])
+        # Collection itself remains available for future indexing.
+        self.vectoradapter.vectorstore.delete(ids=self._get_indexed_ids(self.vectoradapter.vectorstore))
+
         self._log_data(
             f"Collection '{self.dataset}' has been cleaned. ",
             tool_name="_clean_collection"

{alita_sdk-0.3.218 → alita_sdk-0.3.220}/alita_sdk/tools/elitea_base.py

@@ -375,12 +375,30 @@ class BaseVectorStoreToolApiWrapper(BaseToolApiWrapper):
             Cleans the indexed data in the collection
         """
 
-        self._init_vector_store(collection_suffix)._clean_collection()
+        self._init_vector_store(collection_suffix)._remove_collection()
 
     def list_collections(self):
         """
             Lists all collections in the vector store
         """
+        if self.vectorstore_type.lower() == 'PGVector'.lower():
+            from sqlalchemy import text
+            from sqlalchemy.orm import Session
+
+            # schema_name = self.vectorstore.collection_name
+            with Session(self._init_vector_store().vectorstore.session_maker.bind) as session:
+                get_collections = text("""
+                    SELECT table_schema
+                    FROM information_schema.columns
+                    WHERE udt_name = 'vector';
+                """)
+
+                # Execute the raw SQL query
+                result = session.execute(get_collections)
+
+                # Fetch all rows from the result
+                docs = result.fetchall()
+            return str(docs)
         vector_client = self._init_vector_store().vectoradapter.vectorstore._client
         return ','.join([collection.name for collection in vector_client.list_collections()])
 

{alita_sdk-0.3.218 → alita_sdk-0.3.220}/alita_sdk.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: alita_sdk
-Version: 0.3.218
+Version: 0.3.220
 Summary: SDK for building langchain agents using resources from Alita
 Author-email: Artem Rozumenko <artyom.rozumenko@gmail.com>, Mikalai Biazruchka <mikalai_biazruchka@epam.com>, Roman Mitusov <roman_mitusov@epam.com>, Ivan Krakhmaliuk <lifedjik@gmail.com>, Artem Dubrovskiy <ad13box@gmail.com>
 License-Expression: Apache-2.0

{alita_sdk-0.3.218 → alita_sdk-0.3.220}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "alita_sdk"
-version = "0.3.218"
+version = "0.3.220"
 description = "SDK for building langchain agents using resources from Alita"
 readme = "README.md"
 requires-python = ">=3.10"