Method get_client

AskAgent/python/core/retriever.py:502–581

Get or initialize the appropriate vector database client for a specific endpoint. Uses a cache to avoid creating duplicate client instances. Args: endpoint_name: Name of the endpoint to get client for Returns: Appropriate vector database cli

(self, endpoint_name: str)

            return False

    async def get_client(self, endpoint_name: str) -> VectorDBClientInterface:
        """
        Get or initialize the appropriate vector database client for a specific endpoint.
        Uses a cache to avoid creating duplicate client instances.

        Args:
            endpoint_name: Name of the endpoint to get client for

        Returns:
            Appropriate vector database client
        """
        if endpoint_name not in self.enabled_endpoints:
            raise ValueError(f"Endpoint {endpoint_name} is not in enabled endpoints")

        config = self.enabled_endpoints[endpoint_name]
        db_type = config.db_type

        # Use cache key combining db_type and endpoint
        cache_key = f"{db_type}_{endpoint_name}"

        # Check if client already exists in cache
        async with _client_cache_lock:
            if cache_key in _client_cache:
                return _client_cache[cache_key]

            # Ensure required packages are installed
            _ensure_package_installed(db_type)

            # Create the appropriate client with dynamic imports
            logger.debug(f"Creating new client for {db_type} with endpoint {endpoint_name}")

            try:
                # Use preloaded module if available, otherwise load on demand
                if db_type in _preloaded_modules:
                    client_class = _preloaded_modules[db_type]
                    client = client_class(endpoint_name)
                elif db_type == "azure_ai_search":
                    from retrieval_providers.azure_search_client import AzureSearchClient
                    client = AzureSearchClient(endpoint_name)
                elif db_type == "milvus":
                    from retrieval_providers.milvus_client import MilvusVectorClient
                    client = MilvusVectorClient(endpoint_name)
                elif db_type == "opensearch":
                    from retrieval_providers.opensearch_client import OpenSearchClient
                    client = OpenSearchClient(endpoint_name)
                elif db_type == "qdrant":
                    from retrieval_providers.qdrant import QdrantVectorClient
                    client = QdrantVectorClient(endpoint_name)
                elif db_type == "snowflake_cortex_search":
                    from retrieval_providers.snowflake_client import SnowflakeCortexSearchClient
                    client = SnowflakeCortexSearchClient(endpoint_name)
                elif db_type == "cloudflare_autorag":
                    from retrieval_providers.cf_autorag_client import (
                        CloudflareAutoRAGClient,
                    )

                    client = CloudflareAutoRAGClient(endpoint_name)
                elif db_type == "elasticsearch":