crewplus 0.2.1__tar.gz → 0.2.3__tar.gz

{crewplus-0.2.1 → crewplus-0.2.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: crewplus
-Version: 0.2.1
+Version: 0.2.3
 Summary: Base services for CrewPlus AI applications
 Author-Email: Tim Liu <tim@opsmateai.com>
 License: MIT
@@ -45,7 +45,7 @@ CrewPlus is designed as a modular and extensible ecosystem of packages. This all
 
 -   **Chat Services:** A unified interface for interacting with various chat models (e.g., `GeminiChatModel`).
 -   **Model Load Balancer:** Intelligently distribute requests across multiple LLM endpoints.
--   **Vector DB Services:** Abstractions for working with popular vector stores for retrieval-augmented generation (RAG).
+-   **Vector DB Services:** working with popular vector stores (e.g. Milvus, Zilliz Cloud) for retrieval-augmented generation (RAG) and agent memory.
 
 
 ## Documentation

{crewplus-0.2.1 → crewplus-0.2.3}/README.md

@@ -25,7 +25,7 @@ CrewPlus is designed as a modular and extensible ecosystem of packages. This all
 
 -   **Chat Services:** A unified interface for interacting with various chat models (e.g., `GeminiChatModel`).
 -   **Model Load Balancer:** Intelligently distribute requests across multiple LLM endpoints.
--   **Vector DB Services:** Abstractions for working with popular vector stores for retrieval-augmented generation (RAG).
+-   **Vector DB Services:** working with popular vector stores (e.g. Milvus, Zilliz Cloud) for retrieval-augmented generation (RAG) and agent memory.
 
 
 ## Documentation

{crewplus-0.2.1 → crewplus-0.2.3}/crewplus/vectorstores/milvus/vdb_service.py

@@ -22,17 +22,22 @@ class VDBService(object):
     and provides helper methods to get embedding functions and vector store instances.
 
     Args:
-        settings (dict): A dictionary containing configuration for the vector store
+        settings (dict, optional): A dictionary containing configuration for the vector store
                          and embedding models.
+        endpoint (str, optional): The URI for the Zilliz cluster. Can be used for simple
+                                  initialization instead of `settings`.
+        token (str, optional): The token for authenticating with Zilliz. Must be provided
+                               with `endpoint`.
         schema (str, optional): The schema definition for a collection. Defaults to None.
         logger (logging.Logger, optional): An optional logger instance. Defaults to None.
 
     Raises:
-        ValueError: If required configurations are missing from the settings dictionary.
+        ValueError: If required configurations are missing.
         NotImplementedError: If an unsupported provider is specified.
         RuntimeError: If the MilvusClient fails to initialize after a retry.
 
     Example:
+        >>> # Initialize with a full settings dictionary
         >>> settings = {
         ...     "embedder": {
         ...         "provider": "azure-openai",
@@ -61,6 +66,10 @@ class VDBService(object):
         ...     }
         ... }
         >>> vdb_service = VDBService(settings=settings)
+        >>>
+        >>> # Alternatively, initialize with an endpoint and token for Zilliz
+        >>> # vdb_service_zilliz = VDBService(endpoint="YOUR_ZILLIZ_ENDPOINT", token="YOUR_ZILLIZ_TOKEN")
+        >>>
         >>> # Get the raw Milvus client
         >>> client = vdb_service.get_vector_client()
         >>> print(client.list_collections())
@@ -82,17 +91,41 @@ class VDBService(object):
     connection_args: dict
     settings: dict
     
-    def __init__(self, settings: dict, schema: str = None, logger: logging.Logger = None):
+    def __init__(self, settings: dict = None, endpoint: str = None, token: str = None, schema: str = None, logger: logging.Logger = None):
         """
         Initializes the VDBService.
+
+        Can be initialized in two ways:
+        1. By providing a full `settings` dictionary for complex configurations.
+        2. By providing `endpoint` and `token` for a direct Zilliz connection.
+           Note: When using this method, an `embedder` configuration is not created.
+           You must either use the `ModelLoadBalancer` or pass an `Embeddings` object
+           directly to methods like `get_vector_store`.
         
         Args:
-            settings (dict): Configuration dictionary for the service.
+            settings (dict, optional): Configuration dictionary for the service. Defaults to None.
+            endpoint (str, optional): The URI for the Zilliz cluster. Used if `settings` is not provided.
+            token (str, optional): The token for authenticating with the Zilliz cluster.
             schema (str, optional): Default schema for new collections. Defaults to None.
             logger (logging.Logger, optional): Logger instance. Defaults to None.
         """
         self.logger = logger or logging.getLogger(__name__)
-        self.settings = settings
+
+        if settings:
+            self.settings = settings
+        elif endpoint and token:
+            self.logger.info("Initializing VDBService with endpoint and token for a Zilliz connection.")
+            self.settings = {
+                "vector_store": {
+                    "provider": "zilliz",
+                    "config": {
+                        "uri": endpoint,
+                        "token": token
+                    }
+                }
+            }
+        else:
+            raise ValueError("VDBService must be initialized with either a 'settings' dictionary or both 'endpoint' and 'token'.")
 
         vector_store_settings = self.settings.get("vector_store")
         if not vector_store_settings:
@@ -257,43 +290,48 @@ class VDBService(object):
 
         return vdb
 
-    def delete_old_indexes(self, url: str = None, vdb: Zilliz = None) -> None:
+    def delete_old_indexes(self, url: str = None, vdb: Zilliz = None) -> (bool | None):
         """ Delete old indexes of the same source_url
 
         Args:
             url (str): source url
+            vdb (Zilliz): Zilliz instance
         """
+        self.logger.info(f"Delete old indexes of the same source_url:{url}")
+
         if url is None or vdb is None:
-            return
+            return None
 
         # Delete indexes of the same source_url
-        expr = "source in [\"" + url + "\"]"
+        expr = f'source_url == "{url}" or source == "{url}"'
         pks = vdb.get_pks(expr)
 
         # Delete entities by pks
         if pks is not None and len(pks) > 0 :
-            old_items = vdb.delete(pks)
-            self.logger.info("ingesting document -- delete old indexes -- " + str(old_items))
+            res = vdb.delete(pks)
+            self.logger.info("Deleted old indexes result: " + str(res))
+            return res
 
-    def delete_old_indexes_by_id(self, id: str = None, vdb: Zilliz = None) -> None:
+    def delete_old_indexes_by_id(self, source_id: str = None, vdb: Zilliz = None) -> (bool | None):
         """ Delete old indexes of the same source_id
 
         Args:
-            id (str): source id
+            source_id (str): source id
         """
-        self.logger.info(f"Delete old indexes of the same source_id:{id}")
+        self.logger.info(f"Delete old indexes of the same source_id:{source_id}")
 
-        if id is None or vdb is None:
-            return
+        if source_id is None or vdb is None:
+            return None
 
         # Delete indexes of the same source_id
-        expr = "source_id in [\"" + id + "\"]"
+        expr = f'source_id == "{source_id}"'
         pks = vdb.get_pks(expr)
 
         # Delete entities by pks
         if pks is not None and len(pks) > 0 :
-            old_items = vdb.delete(pks)
-            self.logger.info("ingesting document -- delete old indexes -- " + str(old_items))
+            res = vdb.delete(pks)
+            self.logger.info("Deleted old indexes result: " + str(res))
+            return res
 
     def drop_collection(self, collection_name: str) -> None:
         """
@@ -326,12 +364,13 @@ class VDBService(object):
                 self.logger.info(f"Removed '{collection_name}' from instance cache.")
 
     def delete_data_by_filter(self, collection_name: str = None, filter: str = None) -> None:
-        """ Delete a collection
+        """ Delete data by filter
 
         Args:
-            collection_name (str): scollection_name
+            collection_name (str): collection_name
+            filter (str): filter
         """
-        self.logger.info(f"drop a collection by name:{collection_name}")
+        self.logger.info(f"Delete data by filter:{filter}")
 
         try:
             client=self.get_vector_client()

{crewplus-0.2.1 → crewplus-0.2.3}/pyproject.toml

@@ -6,7 +6,7 @@ build-backend = "pdm.backend"
 
 [project]
 name = "crewplus"
-version = "0.2.1"
+version = "0.2.3"
 description = "Base services for CrewPlus AI applications"
 authors = [
     { name = "Tim Liu", email = "tim@opsmateai.com" },