mirror of
https://github.com/microsoft/agent-framework.git
synced 2026-06-16 21:04:09 +08:00
Python: [BREAKING] PR2 — Wire context provider pipeline, remove old types, update all consumers (#3850)
* PR2: Wire context provider pipeline and update all internal consumers - Replace AgentThread with AgentSession across all packages - Replace ContextProvider with BaseContextProvider across all packages - Replace context_provider param with context_providers (Sequence) - Replace thread= with session= in run() signatures - Replace get_new_thread() with create_session() - Add get_session(service_session_id) to agent interface - DurableAgentThread -> DurableAgentSession - Remove _notify_thread_of_new_messages from WorkflowAgent - Wire before_run/after_run context provider pipeline in RawAgent - Auto-inject InMemoryHistoryProvider when no providers configured * fix: update all tests for context provider pipeline, fix lazy-loaders, remove old test files * refactor: update all sample files for context provider pipeline (AgentThread→AgentSession, ContextProvider→BaseContextProvider) * fix: update remaining ag-ui references (client docstring, getting_started sample) * fix: make get_session service_session_id keyword-only to avoid confusion with session_id * refactor: rename _RunContext.thread_messages to session_messages * refactor: remove _threads.py, _memory.py, and old provider files; migrate devui to use plain message lists * rename: remove _new_ prefix from test files * refactor: rewrite SlidingWindowChatMessageStore as SlidingWindowHistoryProvider(InMemoryHistoryProvider) * fix: read full history from session state directly instead of reaching into provider internals * fix: update stale .pyi stubs, sample imports, and README references for new provider types * fix: remove stale message_store, _notify_thread_of_new_messages, and session_id.key references in samples * refactor: merge context_providers and sessions sample folders into sessions, remove aggregate_context_provider * refactor: UserInfoMemory stores state in session.state instead of instance attributes * feat: add Pydantic BaseModel support to session state serialization Pydantic models stored in session.state are now automatically serialized via model_dump() and restored via model_validate() during to_dict()/from_dict() round-trips. Models are auto-registered on first serialization; use register_state_type() for cold-start deserialization. Also export register_state_type as a public API. * fix mem0 * Update sample README links and descriptions for session terminology - Replace 'thread' with 'session' in sample descriptions across all READMEs - Update file links for renamed samples (mem0_sessions, redis_sessions, etc.) - Fix Threads section → Sessions section in main samples/README.md - Update tools, middleware, workflows, durabletask, azure_functions READMEs - Update architecture diagrams in concepts/tools/README.md - Update migration guides (autogen, semantic-kernel) * Fix broken Redis README link to renamed sample * Fix Mem0 OSS client search: pass scoping params as direct kwargs AsyncMemory (OSS) expects user_id/agent_id/run_id as direct kwargs, while AsyncMemoryClient (Platform) expects them in a filters dict. Adds tests for both client types. Port of fix from #3844 to new Mem0ContextProvider. * Fix rebase issues: restore missing _conversation_state.py and checkpoint decode logic - Add back _conversation_state.py (encode/decode_chat_messages) lost in rebase - Fix on_checkpoint_restore to decode cache/conversation with decode_chat_messages - Fix on_checkpoint_restore to use decode_checkpoint_value for pending requests - Add tests/workflow/__init__.py for relative import support - Fix test_agent_executor checkpoint selection (checkpoints[1] not superstep) * Add STORES_BY_DEFAULT ClassVar to skip redundant InMemoryHistoryProvider injection Chat clients that store history server-side by default (OpenAI Responses API, Azure AI Agent) now declare STORES_BY_DEFAULT = True. The agent checks this during auto-injection and skips InMemoryHistoryProvider unless the user explicitly sets store=False. * Fix broken markdown links in azure_ai and redis READMEs * Fix getting-started samples to use session API instead of removed thread/ContextProvider API * updates to workflow as agent * fix group chat import * Rename Thread→Session throughout, fix service_session_id propagation, remove stale AGUIThread - Fix: Propagate conversation_id from ChatResponse back to session.service_session_id in both streaming and non-streaming paths in _agents.py - Rename AgentThreadException → AgentSessionException - Remove stale AGUIThread from ag_ui lazy-loader - Rename use_service_thread → use_service_session in ag-ui package - Rename test functions from *_thread_* to *_session_* - Rename sample files from *_thread* to *_session* - Update docstrings and comments: thread → session - Update _mcp.py kwargs filter: add 'session' alongside 'thread' - Fix ContinuationToken docstring example: thread=thread → session=session - Fix _clients.py docstring: 'Agent threads' → 'Agent sessions' * Fix broken markdown links after thread→session file renames * fix azure ai test
This commit is contained in:
Eduard van Valkenburg
committed by
GitHub
Unverified
GitHub
parent
0c67dbbce5
commit
1e350ea22f
@@ -8,8 +8,7 @@ import os
|
||||
if os.environ.get("MEM0_TELEMETRY") is None:
|
||||
os.environ["MEM0_TELEMETRY"] = "false"
|
||||
|
||||
from ._context_provider import _Mem0ContextProvider
|
||||
from ._provider import Mem0Provider
|
||||
from ._context_provider import Mem0ContextProvider
|
||||
|
||||
try:
|
||||
__version__ = importlib.metadata.version(__name__)
|
||||
@@ -17,7 +16,6 @@ except importlib.metadata.PackageNotFoundError:
|
||||
__version__ = "0.0.0" # Fallback for development mode
|
||||
|
||||
__all__ = [
|
||||
"Mem0Provider",
|
||||
"_Mem0ContextProvider",
|
||||
"Mem0ContextProvider",
|
||||
"__version__",
|
||||
]
|
||||
|
||||
@@ -2,9 +2,8 @@
|
||||
|
||||
"""New-pattern Mem0 context provider using BaseContextProvider.
|
||||
|
||||
This module provides ``_Mem0ContextProvider``, a side-by-side implementation of
|
||||
:class:`Mem0Provider` built on the new :class:`BaseContextProvider` hooks pattern.
|
||||
It will be renamed to ``Mem0ContextProvider`` in PR2 when the old class is removed.
|
||||
This module provides ``Mem0ContextProvider``, built on the new
|
||||
:class:`BaseContextProvider` hooks pattern.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
@@ -35,17 +34,11 @@ class _MemorySearchResponse_v1_1(TypedDict):
|
||||
_MemorySearchResponse_v2 = list[dict[str, Any]]
|
||||
|
||||
|
||||
class _Mem0ContextProvider(BaseContextProvider):
|
||||
class Mem0ContextProvider(BaseContextProvider):
|
||||
"""Mem0 context provider using the new BaseContextProvider hooks pattern.
|
||||
|
||||
Integrates Mem0 for persistent semantic memory, searching and storing
|
||||
memories via the Mem0 API. This is the new-pattern equivalent of
|
||||
:class:`Mem0Provider`.
|
||||
|
||||
Note:
|
||||
This class uses a temporary ``_`` prefix to coexist with the existing
|
||||
:class:`Mem0Provider`. It will be renamed to ``Mem0ContextProvider``
|
||||
in PR2.
|
||||
memories via the Mem0 API.
|
||||
"""
|
||||
|
||||
DEFAULT_CONTEXT_PROMPT = "## Memories\nConsider the following memories when answering user questions:"
|
||||
@@ -115,9 +108,16 @@ class _Mem0ContextProvider(BaseContextProvider):
|
||||
|
||||
filters = self._build_filters(session_id=context.session_id)
|
||||
|
||||
# AsyncMemory (OSS) expects user_id/agent_id/run_id as direct kwargs
|
||||
# AsyncMemoryClient (Platform) expects them in a filters dict
|
||||
search_kwargs: dict[str, Any] = {"query": input_text}
|
||||
if isinstance(self.mem0_client, AsyncMemory):
|
||||
search_kwargs.update(filters)
|
||||
else:
|
||||
search_kwargs["filters"] = filters
|
||||
|
||||
search_response: _MemorySearchResponse_v1_1 | _MemorySearchResponse_v2 = await self.mem0_client.search( # type: ignore[misc]
|
||||
query=input_text,
|
||||
filters=filters,
|
||||
**search_kwargs,
|
||||
)
|
||||
|
||||
if isinstance(search_response, list):
|
||||
@@ -190,4 +190,4 @@ class _Mem0ContextProvider(BaseContextProvider):
|
||||
return filters
|
||||
|
||||
|
||||
__all__ = ["_Mem0ContextProvider"]
|
||||
__all__ = ["Mem0ContextProvider"]
|
||||
|
||||
@@ -1,239 +0,0 @@
|
||||
# Copyright (c) Microsoft. All rights reserved.
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import sys
|
||||
from collections.abc import MutableSequence, Sequence
|
||||
from contextlib import AbstractAsyncContextManager
|
||||
from typing import Any
|
||||
|
||||
from agent_framework import Context, ContextProvider, Message
|
||||
from agent_framework.exceptions import ServiceInitializationError
|
||||
from mem0 import AsyncMemory, AsyncMemoryClient
|
||||
|
||||
if sys.version_info >= (3, 12):
|
||||
from typing import override # type: ignore # pragma: no cover
|
||||
else:
|
||||
from typing_extensions import override # type: ignore[import] # pragma: no cover
|
||||
|
||||
if sys.version_info >= (3, 11):
|
||||
from typing import NotRequired, Self, TypedDict # pragma: no cover
|
||||
else:
|
||||
from typing_extensions import NotRequired, Self, TypedDict # pragma: no cover
|
||||
|
||||
|
||||
# Type aliases for Mem0 search response formats (v1.1 and v2; v1 is deprecated, but matches the type definition for v2)
|
||||
class MemorySearchResponse_v1_1(TypedDict):
|
||||
results: list[dict[str, Any]]
|
||||
relations: NotRequired[list[dict[str, Any]]]
|
||||
|
||||
|
||||
MemorySearchResponse_v2 = list[dict[str, Any]]
|
||||
|
||||
|
||||
class Mem0Provider(ContextProvider):
|
||||
"""Mem0 Context Provider.
|
||||
|
||||
Note:
|
||||
Mem0's telemetry is disabled by default when using this package.
|
||||
To enable telemetry, set the environment variable ``MEM0_TELEMETRY=true`` before
|
||||
importing this package.
|
||||
"""
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
mem0_client: AsyncMemory | AsyncMemoryClient | None = None,
|
||||
api_key: str | None = None,
|
||||
application_id: str | None = None,
|
||||
agent_id: str | None = None,
|
||||
thread_id: str | None = None,
|
||||
user_id: str | None = None,
|
||||
scope_to_per_operation_thread_id: bool = False,
|
||||
context_prompt: str = ContextProvider.DEFAULT_CONTEXT_PROMPT,
|
||||
) -> None:
|
||||
"""Initializes a new instance of the Mem0Provider class.
|
||||
|
||||
Args:
|
||||
mem0_client: A pre-created Mem0 MemoryClient or None to create a default client.
|
||||
api_key: The API key for authenticating with the Mem0 API. If not
|
||||
provided, it will attempt to use the MEM0_API_KEY environment variable.
|
||||
application_id: The application ID for scoping memories or None.
|
||||
agent_id: The agent ID for scoping memories or None.
|
||||
thread_id: The thread ID for scoping memories or None.
|
||||
user_id: The user ID for scoping memories or None.
|
||||
scope_to_per_operation_thread_id: Whether to scope memories to per-operation thread ID.
|
||||
context_prompt: The prompt to prepend to retrieved memories.
|
||||
"""
|
||||
should_close_client = False
|
||||
if mem0_client is None:
|
||||
mem0_client = AsyncMemoryClient(api_key=api_key)
|
||||
should_close_client = True
|
||||
|
||||
self.api_key = api_key
|
||||
self.application_id = application_id
|
||||
self.agent_id = agent_id
|
||||
self.thread_id = thread_id
|
||||
self.user_id = user_id
|
||||
self.scope_to_per_operation_thread_id = scope_to_per_operation_thread_id
|
||||
self.context_prompt = context_prompt
|
||||
self.mem0_client = mem0_client
|
||||
self._per_operation_thread_id: str | None = None
|
||||
self._should_close_client = should_close_client
|
||||
|
||||
async def __aenter__(self) -> Self:
|
||||
"""Async context manager entry."""
|
||||
if self.mem0_client and isinstance(self.mem0_client, AbstractAsyncContextManager):
|
||||
await self.mem0_client.__aenter__()
|
||||
return self
|
||||
|
||||
async def __aexit__(self, exc_type: type[BaseException] | None, exc_val: BaseException | None, exc_tb: Any) -> None:
|
||||
"""Async context manager exit."""
|
||||
if self._should_close_client and self.mem0_client and isinstance(self.mem0_client, AbstractAsyncContextManager):
|
||||
await self.mem0_client.__aexit__(exc_type, exc_val, exc_tb)
|
||||
|
||||
async def thread_created(self, thread_id: str | None = None) -> None:
|
||||
"""Called when a new thread is created.
|
||||
|
||||
Args:
|
||||
thread_id: The ID of the thread or None.
|
||||
"""
|
||||
self._validate_per_operation_thread_id(thread_id)
|
||||
self._per_operation_thread_id = self._per_operation_thread_id or thread_id
|
||||
|
||||
@override
|
||||
async def invoked(
|
||||
self,
|
||||
request_messages: Message | Sequence[Message],
|
||||
response_messages: Message | Sequence[Message] | None = None,
|
||||
invoke_exception: Exception | None = None,
|
||||
**kwargs: Any,
|
||||
) -> None:
|
||||
self._validate_filters()
|
||||
|
||||
request_messages_list = [request_messages] if isinstance(request_messages, Message) else list(request_messages)
|
||||
response_messages_list = (
|
||||
[response_messages]
|
||||
if isinstance(response_messages, Message)
|
||||
else list(response_messages)
|
||||
if response_messages
|
||||
else []
|
||||
)
|
||||
messages_list = [*request_messages_list, *response_messages_list]
|
||||
|
||||
# Extract role value - it may be a Role enum or a string
|
||||
def get_role_value(role: Any) -> str:
|
||||
return role.value if hasattr(role, "value") else str(role)
|
||||
|
||||
messages: list[dict[str, str]] = [
|
||||
{"role": get_role_value(message.role), "content": message.text}
|
||||
for message in messages_list
|
||||
if get_role_value(message.role) in {"user", "assistant", "system"} and message.text and message.text.strip()
|
||||
]
|
||||
|
||||
if messages:
|
||||
await self.mem0_client.add( # type: ignore[misc]
|
||||
messages=messages,
|
||||
user_id=self.user_id,
|
||||
agent_id=self.agent_id,
|
||||
run_id=self._per_operation_thread_id if self.scope_to_per_operation_thread_id else self.thread_id,
|
||||
metadata={"application_id": self.application_id},
|
||||
)
|
||||
|
||||
@override
|
||||
async def invoking(self, messages: Message | MutableSequence[Message], **kwargs: Any) -> Context:
|
||||
"""Called before invoking the AI model to provide context.
|
||||
|
||||
Args:
|
||||
messages: List of new messages in the thread.
|
||||
|
||||
Keyword Args:
|
||||
**kwargs: not used at present.
|
||||
|
||||
Returns:
|
||||
Context: Context object containing instructions with memories.
|
||||
"""
|
||||
self._validate_filters()
|
||||
messages_list = [messages] if isinstance(messages, Message) else list(messages)
|
||||
input_text = "\n".join(msg.text for msg in messages_list if msg and msg.text and msg.text.strip())
|
||||
|
||||
# Validate input text is not empty before searching (possible for function approval responses)
|
||||
if not input_text.strip():
|
||||
return Context(messages=None)
|
||||
|
||||
# Build filters from init parameters
|
||||
filters = self._build_filters()
|
||||
|
||||
search_response: MemorySearchResponse_v1_1 | MemorySearchResponse_v2 = await self.mem0_client.search( # type: ignore[misc]
|
||||
query=input_text,
|
||||
filters=filters,
|
||||
)
|
||||
|
||||
# Depending on the API version, the response schema varies slightly
|
||||
if isinstance(search_response, list):
|
||||
memories = search_response
|
||||
elif isinstance(search_response, dict) and "results" in search_response:
|
||||
memories = search_response["results"]
|
||||
else:
|
||||
# Fallback for unexpected schema - return response as text as-is
|
||||
memories = [search_response]
|
||||
|
||||
line_separated_memories = "\n".join(memory.get("memory", "") for memory in memories)
|
||||
|
||||
return Context(
|
||||
messages=[Message(role="user", text=f"{self.context_prompt}\n{line_separated_memories}")]
|
||||
if line_separated_memories
|
||||
else None
|
||||
)
|
||||
|
||||
def _validate_filters(self) -> None:
|
||||
"""Validates that at least one filter is provided.
|
||||
|
||||
Raises:
|
||||
ServiceInitializationError: If no filters are provided.
|
||||
"""
|
||||
if not self.agent_id and not self.user_id and not self.application_id and not self.thread_id:
|
||||
raise ServiceInitializationError(
|
||||
"At least one of the filters: agent_id, user_id, application_id, or thread_id is required."
|
||||
)
|
||||
|
||||
def _build_filters(self) -> dict[str, Any]:
|
||||
"""Build search filters from initialization parameters.
|
||||
|
||||
Returns:
|
||||
Filter dictionary for mem0 v2 search API containing initialization parameters.
|
||||
In the v2 API, filters holds the user_id, agent_id, run_id (thread_id), and app_id
|
||||
(application_id) which are required for scoping memory search operations.
|
||||
"""
|
||||
filters: dict[str, Any] = {}
|
||||
|
||||
if self.user_id:
|
||||
filters["user_id"] = self.user_id
|
||||
if self.agent_id:
|
||||
filters["agent_id"] = self.agent_id
|
||||
if self.scope_to_per_operation_thread_id and self._per_operation_thread_id:
|
||||
filters["run_id"] = self._per_operation_thread_id
|
||||
elif self.thread_id:
|
||||
filters["run_id"] = self.thread_id
|
||||
if self.application_id:
|
||||
filters["app_id"] = self.application_id
|
||||
|
||||
return filters
|
||||
|
||||
def _validate_per_operation_thread_id(self, thread_id: str | None) -> None:
|
||||
"""Validates that a new thread ID doesn't conflict with an existing one when scoped.
|
||||
|
||||
Args:
|
||||
thread_id: The new thread ID or None.
|
||||
|
||||
Raises:
|
||||
ValueError: If a new thread ID is provided when one already exists.
|
||||
"""
|
||||
if (
|
||||
self.scope_to_per_operation_thread_id
|
||||
and thread_id
|
||||
and self._per_operation_thread_id
|
||||
and thread_id != self._per_operation_thread_id
|
||||
):
|
||||
raise ValueError(
|
||||
"Mem0Provider can only be used with one thread at a time when scope_to_per_operation_thread_id is True."
|
||||
)
|
||||
Reference in New Issue
Block a user