Python: Add json schema handling. Add experimental tag to OpenAPI and Memory Connectors.

python/samples/concepts/auto_function_calling/chat_gpt_api_function_calling.py

@@ -64,7 +64,7 @@
     temperature=0.7,
     top_p=0.8,
     function_call_behavior=FunctionCallBehavior.EnableFunctions(
-        auto_invoke=True, filters={"included_plugins": ["math"]}
+        auto_invoke=True, filters={"included_plugins": ["math", "time"]}
     ),
 )
 

python/semantic_kernel/connectors/ai/open_ai/services/utils.py

@@ -13,16 +13,6 @@
 logger = logging.getLogger(__name__)
 
 
-TYPE_MAPPER = {
-    "str": "string",
-    "int": "number",
-    "float": "number",
-    "bool": "boolean",
-    "list": "array",
-    "dict": "object",
-}
-
-
 def update_settings_from_function_call_configuration(
     function_call_configuration: "FunctionCallConfiguration", settings: "OpenAIChatPromptExecutionSettings"
 ) -> None:
@@ -44,31 +34,31 @@ def update_settings_from_function_call_configuration(
 
 def kernel_function_metadata_to_openai_tool_format(metadata: KernelFunctionMetadata) -> dict[str, Any]:
     """Convert the kernel function metadata to OpenAI format."""
+
+    def parse_schema(schema_data):
+        """Recursively parse the schema data to include nested properties."""
+        if schema_data.get("type") == "object":
+            return {
+                "type": "object",
+                "properties": {key: parse_schema(value) for key, value in schema_data.get("properties", {}).items()},
+                "description": schema_data.get("description", ""),
+            }
+        else:
+            return {
+                "type": schema_data.get("type", "string"),
+                "description": schema_data.get("description", ""),
+                **({"enum": schema_data.get("enum")} if "enum" in schema_data else {}),
+            }
+
     return {
         "type": "function",
         "function": {
             "name": metadata.fully_qualified_name,
             "description": metadata.description or "",
             "parameters": {
                 "type": "object",
-                "properties": {
-                    param.name: {
-                        "description": param.description or "",
-                        "type": parse_parameter_type(param.type_),
-                        **({"enum": param.enum} if hasattr(param, "enum") else {}),  # Added support for enum
-                    }
-                    for param in metadata.parameters
-                },
+                "properties": {param.name: parse_schema(param.schema_data) for param in metadata.parameters},
                 "required": [p.name for p in metadata.parameters if p.is_required],
             },
         },
     }
-
-
-def parse_parameter_type(param_type: str | None) -> str:
-    """Parse the parameter type."""
-    if not param_type:
-        return "string"
-    if "," in param_type:
-        param_type = param_type.split(",", maxsplit=1)[0]
-    return TYPE_MAPPER.get(param_type, "string")

python/semantic_kernel/connectors/memory/astradb/astra_client.py

@@ -6,6 +6,7 @@
 from semantic_kernel.connectors.memory.astradb.utils import AsyncSession
 from semantic_kernel.connectors.telemetry import APP_INFO
 from semantic_kernel.exceptions import ServiceResponseException
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 ASTRA_CALLER_IDENTITY: str
 SEMANTIC_KERNEL_VERSION = APP_INFO.get("Semantic-Kernel-Version")
@@ -15,6 +16,7 @@
     ASTRA_CALLER_IDENTITY = "semantic-kernel"
 
 
+@experimental_class
 class AstraClient:
     def __init__(
         self,

python/semantic_kernel/connectors/memory/astradb/astradb_memory_store.py

@@ -17,6 +17,7 @@
 from semantic_kernel.exceptions import MemoryConnectorInitializationError
 from semantic_kernel.memory.memory_record import MemoryRecord
 from semantic_kernel.memory.memory_store_base import MemoryStoreBase
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 MAX_DIMENSIONALITY = 20000
 MAX_UPSERT_BATCH_SIZE = 100
@@ -28,6 +29,7 @@
 logger: logging.Logger = logging.getLogger(__name__)
 
 
+@experimental_class
 class AstraDBMemoryStore(MemoryStoreBase):
     """A memory store that uses Astra database as the backend."""
 

python/semantic_kernel/connectors/memory/astradb/astradb_settings.py

@@ -3,8 +3,10 @@
 from pydantic import SecretStr
 
 from semantic_kernel.connectors.memory.memory_settings_base import BaseModelSettings
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 
+@experimental_class
 class AstraDBSettings(BaseModelSettings):
     """AstraDB model settings
 

python/semantic_kernel/connectors/memory/azure_cognitive_search/azure_ai_search_settings.py

@@ -4,8 +4,10 @@
 
 from semantic_kernel.connectors.memory.memory_settings_base import BaseModelSettings
 from semantic_kernel.kernel_pydantic import HttpsUrl
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 
+@experimental_class
 class AzureAISearchSettings(BaseModelSettings):
     """Azure AI Search model settings currently used by the AzureCognitiveSearchMemoryStore connector
 

python/semantic_kernel/connectors/memory/azure_cognitive_search/azure_cognitive_search_memory_store.py

@@ -33,10 +33,12 @@
 from semantic_kernel.exceptions import MemoryConnectorInitializationError, MemoryConnectorResourceNotFound
 from semantic_kernel.memory.memory_record import MemoryRecord
 from semantic_kernel.memory.memory_store_base import MemoryStoreBase
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 logger: logging.Logger = logging.getLogger(__name__)
 
 
+@experimental_class
 class AzureCognitiveSearchMemoryStore(MemoryStoreBase):
     _search_index_client: SearchIndexClient = None
     _vector_size: int = None

python/semantic_kernel/connectors/memory/azure_cosmosdb/azure_cosmos_db_memory_store.py

@@ -17,10 +17,12 @@
 from semantic_kernel.exceptions import MemoryConnectorInitializationError
 from semantic_kernel.memory.memory_record import MemoryRecord
 from semantic_kernel.memory.memory_store_base import MemoryStoreBase
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 logger: logging.Logger = logging.getLogger(__name__)
 
 
+@experimental_class
 class AzureCosmosDBMemoryStore(MemoryStoreBase):
     """A memory store that uses AzureCosmosDB for MongoDB vCore, to perform vector similarity search on a fully
     managed MongoDB compatible database service.

python/semantic_kernel/connectors/memory/azure_cosmosdb/azure_cosmos_db_store_api.py

@@ -7,9 +7,11 @@
 from numpy import ndarray
 
 from semantic_kernel.memory.memory_record import MemoryRecord
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 
 # Abstract class similar to the original data store that allows API level abstraction
+@experimental_class
 class AzureCosmosDBStoreApi(ABC):
     @abstractmethod
     async def create_collection(self, collection_name: str) -> None:

python/semantic_kernel/connectors/memory/azure_cosmosdb/azure_cosmosdb_settings.py

@@ -3,8 +3,10 @@
 from pydantic import SecretStr
 
 from semantic_kernel.connectors.memory.memory_settings_base import BaseModelSettings
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 
+@experimental_class
 class AzureCosmosDBSettings(BaseModelSettings):
     """Azure CosmosDB model settings
 

python/semantic_kernel/connectors/memory/azure_cosmosdb/cosmosdb_utils.py

@@ -7,8 +7,10 @@
 
 from semantic_kernel.connectors.telemetry import HTTP_USER_AGENT
 from semantic_kernel.exceptions import ServiceInitializationError
+from semantic_kernel.utils.experimental_decorator import experimental_function
 
 
+@experimental_function
 class CosmosDBSimilarityType(str, Enum):
     """Cosmos DB Similarity Type as enumerator."""
 
@@ -20,6 +22,7 @@ class CosmosDBSimilarityType(str, Enum):
     """Euclidean distance"""
 
 
+@experimental_function
 class CosmosDBVectorSearchType(str, Enum):
     """Cosmos DB Vector Search Type as enumerator."""
 
@@ -29,6 +32,7 @@ class CosmosDBVectorSearchType(str, Enum):
     """HNSW vector index"""
 
 
+@experimental_function
 def get_mongodb_search_client(connection_string: str, application_name: str):
     """
     Returns a client for Azure Cosmos Mongo vCore Vector DB

python/semantic_kernel/connectors/memory/azure_cosmosdb/mongo_vcore_store_api.py

@@ -13,8 +13,10 @@
     CosmosDBVectorSearchType,
 )
 from semantic_kernel.memory.memory_record import MemoryRecord
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 
+@experimental_class
 class MongoStoreApi(AzureCosmosDBStoreApi):
     database = None
     collection_name: str

python/semantic_kernel/connectors/memory/azure_cosmosdb_no_sql/azure_cosmosdb_no_sql_memory_store.py

@@ -1,36 +1,38 @@
 # Copyright (c) Microsoft. All rights reserved.
 
 import json
-from typing import Any, Dict, List, Tuple
+from typing import Any, List, Tuple
 
 import numpy as np
 from azure.cosmos.aio import ContainerProxy, CosmosClient, DatabaseProxy
 from numpy import ndarray
 
 from semantic_kernel.memory.memory_record import MemoryRecord
 from semantic_kernel.memory.memory_store_base import MemoryStoreBase
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 
 # You can read more about vector search using AzureCosmosDBNoSQL here.
 # https://aka.ms/CosmosVectorSearch
+@experimental_class
 class AzureCosmosDBNoSQLMemoryStore(MemoryStoreBase):
     cosmos_client: CosmosClient = None
     database: DatabaseProxy
     container: ContainerProxy
     database_name: str = None
     partition_key: str = None
-    vector_embedding_policy: [Dict[str, Any]] = None
-    indexing_policy: [Dict[str, Any]] = None
-    cosmos_container_properties: [Dict[str, Any]] = None
+    vector_embedding_policy: dict[str, Any] | None = None
+    indexing_policy: dict[str, Any] | None = None
+    cosmos_container_properties: dict[str, Any] | None = None
 
     def __init__(
         self,
         cosmos_client: CosmosClient,
         database_name: str,
         partition_key: str,
-        vector_embedding_policy: [Dict[str, Any]],
-        indexing_policy: [Dict[str, Any]],
-        cosmos_container_properties: [Dict[str, Any]],
+        vector_embedding_policy: dict[str, Any] | None = None,
+        indexing_policy: dict[str, Any] | None = None,
+        cosmos_container_properties: dict[str, Any] | None = None,
     ):
         if indexing_policy["vectorIndexes"] is None or len(indexing_policy["vectorIndexes"]) == 0:
             raise ValueError("vectorIndexes cannot be null or empty in the indexing_policy.")

python/semantic_kernel/connectors/memory/chroma/chroma_memory_store.py

@@ -9,6 +9,7 @@
 from semantic_kernel.exceptions import ServiceInitializationError, ServiceResourceNotFoundError
 from semantic_kernel.memory.memory_record import MemoryRecord
 from semantic_kernel.memory.memory_store_base import MemoryStoreBase
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 if TYPE_CHECKING:
     import chromadb
@@ -18,6 +19,7 @@
 logger: logging.Logger = logging.getLogger(__name__)
 
 
+@experimental_class
 class ChromaMemoryStore(MemoryStoreBase):
     _client: "chromadb.Client"
 

python/semantic_kernel/connectors/memory/memory_settings_base.py

@@ -2,7 +2,10 @@
 
 from pydantic_settings import BaseSettings
 
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
+
+@experimental_class
 class BaseModelSettings(BaseSettings):
     env_file_path: str | None = None
 

python/semantic_kernel/connectors/memory/milvus/milvus_memory_store.py

@@ -10,6 +10,7 @@
 from semantic_kernel.exceptions import ServiceResourceNotFoundError, ServiceResponseException
 from semantic_kernel.memory.memory_record import MemoryRecord
 from semantic_kernel.memory.memory_store_base import MemoryStoreBase
+from semantic_kernel.utils.experimental_decorator import experimental_class, experimental_function
 
 logger: logging.Logger = logging.getLogger(__name__)
 
@@ -47,6 +48,7 @@
 ]
 
 
+@experimental_function
 def memoryrecord_to_milvus_dict(mem: MemoryRecord) -> Dict[str, Any]:
     """Convert a memoryrecord into a dict.
     Args:
@@ -66,6 +68,7 @@ def memoryrecord_to_milvus_dict(mem: MemoryRecord) -> Dict[str, Any]:
     return ret_dict
 
 
+@experimental_function
 def milvus_dict_to_memoryrecord(milvus_dict: Dict[str, Any]) -> MemoryRecord:
     """Convert Milvus search result dict into MemoryRecord.
 
@@ -92,6 +95,7 @@ def milvus_dict_to_memoryrecord(milvus_dict: Dict[str, Any]) -> MemoryRecord:
     )
 
 
+@experimental_function
 def create_fields(dimensions: int) -> List[FieldSchema]:
     return [
         FieldSchema(
@@ -138,6 +142,7 @@ def create_fields(dimensions: int) -> List[FieldSchema]:
     ]
 
 
+@experimental_class
 class MilvusMemoryStore(MemoryStoreBase):
     def __init__(
         self,

python/semantic_kernel/connectors/memory/mongodb_atlas/mongodb_atlas_memory_store.py

@@ -23,10 +23,12 @@
 from semantic_kernel.exceptions import ServiceResourceNotFoundError
 from semantic_kernel.memory.memory_record import MemoryRecord
 from semantic_kernel.memory.memory_store_base import MemoryStoreBase
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 logger: logging.Logger = logging.getLogger(__name__)
 
 
+@experimental_class
 class MongoDBAtlasMemoryStore(MemoryStoreBase):
     """Memory Store for MongoDB Atlas Vector Search Connections"""
 

python/semantic_kernel/connectors/memory/mongodb_atlas/mongodb_atlas_settings.py

@@ -3,8 +3,10 @@
 from pydantic import SecretStr
 
 from semantic_kernel.connectors.memory.memory_settings_base import BaseModelSettings
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 
+@experimental_class
 class MongoDBAtlasSettings(BaseModelSettings):
     """MongoDB Atlas model settings
 

python/semantic_kernel/connectors/memory/pinecone/pinecone_memory_store.py

@@ -20,6 +20,7 @@
 )
 from semantic_kernel.memory.memory_record import MemoryRecord
 from semantic_kernel.memory.memory_store_base import MemoryStoreBase
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 # Limitations set by Pinecone at https://docs.pinecone.io/reference/known-limitations
 MAX_DIMENSIONALITY = 20000
@@ -32,6 +33,7 @@
 logger: logging.Logger = logging.getLogger(__name__)
 
 
+@experimental_class
 class PineconeMemoryStore(MemoryStoreBase):
     """A memory store that uses Pinecone as the backend."""
 

python/semantic_kernel/connectors/memory/pinecone/pinecone_settings.py

@@ -3,8 +3,10 @@
 from pydantic import SecretStr
 
 from semantic_kernel.connectors.memory.memory_settings_base import BaseModelSettings
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 
+@experimental_class
 class PineconeSettings(BaseModelSettings):
     """Pinecone model settings
 

python/semantic_kernel/connectors/memory/postgres/postgres_memory_store.py

@@ -20,6 +20,7 @@
 )
 from semantic_kernel.memory.memory_record import MemoryRecord
 from semantic_kernel.memory.memory_store_base import MemoryStoreBase
+from semantic_kernel.utils.experimental_decorator import experimental_class
 
 # Limitation based on pgvector documentation https://github.com/pgvector/pgvector#what-if-i-want-to-index-vectors-with-more-than-2000-dimensions
 MAX_DIMENSIONALITY = 2000
@@ -28,6 +29,7 @@
 logger: logging.Logger = logging.getLogger(__name__)
 
 
+@experimental_class
 class PostgresMemoryStore(MemoryStoreBase):
     """A memory store that uses Postgres with pgvector as the backend."""