Spaces:
Running
Running
File size: 13,246 Bytes
a57a50a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 |
"""ChatKit Server implementation for task management with Gemini LLM.
[Task]: T010
[From]: specs/010-chatkit-migration/contracts/backend.md - ChatKitServer Implementation
[From]: specs/010-chatkit-migration/research.md - Section 3
This module implements the ChatKitServer class which handles ChatKit protocol
requests and streams responses using Server-Sent Events (SSE).
The server integrates:
- ChatKit Python SDK for protocol handling
- OpenAI Agents SDK for agent orchestration
- Gemini LLM via OpenAI-compatible endpoint
- MCP tools wrapped as Agents SDK functions
Architecture:
Frontend (ChatKit.js)
↓ SSE with custom fetch
ChatKitServer (this module)
↓ Agents SDK
Gemini API (via AsyncOpenAI with base_url)
"""
import asyncio
import logging
from typing import Any, AsyncIterator, Optional
from uuid import UUID
from openai import AsyncOpenAI
from agents import Agent, set_default_openai_client, RunContextWrapper, Runner
from core.config import get_gemini_client, get_settings
from services.chatkit_store import PostgresChatKitStore
logger = logging.getLogger(__name__)
class AgentContext:
"""Context object passed to agent during execution.
Contains:
- thread_id: Current thread/conversation ID
- user_id: Authenticated user ID
- store: Database store for persistence
- request_context: Additional request metadata
"""
def __init__(
self,
thread_id: str,
user_id: str,
store: PostgresChatKitStore,
request_context: Optional[dict] = None,
):
self.thread_id = thread_id
self.user_id = user_id
self.store = store
self.request_context = request_context or {}
class TaskManagerChatKitServer:
"""ChatKit Server for task management with Gemini LLM.
[From]: specs/010-chatkit-migration/contracts/backend.md - ChatKitServer Implementation
This server extends the ChatKit protocol to work with:
- Custom authentication (JWT cookies)
- Gemini LLM via OpenAI-compatible endpoint
- Server-side tool execution (MCP tools)
- PostgreSQL thread/message persistence
Usage:
from fastapi import FastAPI, Request
from chatkit_server import TaskManagerChatKitServer
app = FastAPI()
server = TaskManagerChatKitServer(store=postgres_store)
@app.post("/api/chatkit")
async def chatkit_endpoint(request: Request):
body = await request.body()
user_id = get_current_user_id(request)
result = await server.process(body, {"user_id": user_id})
if hasattr(result, '__aiter__'):
return StreamingResponse(result, media_type="text/event-stream")
return Response(content=result.json, media_type="application/json")
"""
def __init__(self, store: PostgresChatKitStore):
"""Initialize the ChatKit server.
Args:
store: PostgresChatKitStore instance for persistence
"""
self.store = store
# Configure Gemini client as default for Agents SDK
# [From]: specs/010-chatkit-migration/research.md - Section 2
try:
gemini_client = get_gemini_client()
set_default_openai_client(gemini_client)
logger.info("Gemini client configured for Agents SDK")
except Exception as e:
logger.warning(f"Gemini client not configured: {e}")
# Assistant agent will be configured after tools are wrapped
# This placeholder will be replaced in T019 with actual tools
self.assistant_agent: Optional[Agent] = None
def set_agent(self, agent: Agent) -> None:
"""Set the assistant agent with tools.
[From]: specs/010-chatkit-migration/tasks.md - T019
Args:
agent: Configured Agent with tools and instructions
"""
self.assistant_agent = agent
logger.info(f"Agent configured: {agent.name} with model {agent.model}")
async def process(
self,
body: bytes,
context: dict[str, Any]
) -> Any:
"""Process ChatKit request and return streaming or non-streaming result.
[From]: specs/010-chatkit-migration/contracts/backend.md - ChatKit SSE Endpoint
Args:
body: Raw request body bytes from ChatKit.js
context: Request context containing user_id and auth info
Returns:
StreamingResult for SSE responses or dict for JSON responses
Note: This is a placeholder implementation. The actual implementation
would use the ChatKit Python SDK's process() method which handles
protocol parsing, event routing, and response formatting.
"""
import json
from fastapi.responses import StreamingResponse
# Parse ChatKit protocol request
try:
request_data = json.loads(body.decode('utf-8'))
except Exception as e:
logger.error(f"Failed to parse ChatKit request: {e}")
return {"error": "Invalid request format"}
# Extract thread ID and message content
conversation_id = request_data.get("conversation_id")
item = request_data.get("item", {})
event_type = request_data.get("event", "conversation_item_created")
logger.info(f"ChatKit request: event={event_type}, conversation_id={conversation_id}")
# Get or create thread
thread_id = conversation_id
if not thread_id:
# Create new thread for first message
user_id = context.get("user_id")
if not user_id:
return {"error": "Unauthorized: no user_id in context"}
thread_meta = await self.store.create_thread(
user_id=user_id,
title=None,
metadata={}
)
thread_id = thread_meta["id"]
logger.info(f"Created new thread: {thread_id}")
# Extract user message
user_message = self._extract_message_content(item)
if not user_message:
return {"error": "No message content provided"}
# Build agent context
user_id = context.get("user_id")
agent_context = AgentContext(
thread_id=thread_id,
user_id=user_id,
store=self.store,
request_context=context,
)
# Create user message in database
await self.store.create_message(
thread_id=thread_id,
item={
"type": "message",
"role": "user",
"content": [{"type": "text", "text": user_message}],
}
)
# Stream agent response
# [From]: specs/010-chatkit-migration/research.md - Section 3
# [Task]: T034 - Add timeout handling for long-running tool executions
async def stream_response():
"""Stream ChatKit events as SSE with timeout protection.
[Task]: T034 - Timeout handling for long-running tool executions
Implements a 120-second timeout for agent execution to prevent
indefinite hangs from slow tools or network issues.
"""
if not self.assistant_agent:
yield self._sse_event("error", {"message": "Agent not configured"})
return
try:
# Run agent with streaming and timeout
# [Task]: T034 - 120 second timeout for entire agent execution
# This covers LLM calls, tool executions, and any delays
async with asyncio.timeout(120):
result = Runner.run_streamed(
self.assistant_agent,
[{"role": "user", "content": user_message}],
context=agent_context,
)
# Collect assistant response
full_response = ""
async for chunk in result:
if hasattr(chunk, 'content'):
content = chunk.content
if content:
full_response += content
# Stream text delta
yield self._sse_event("message_delta", {
"type": "text",
"text": content
})
# Create assistant message in database
await self.store.create_message(
thread_id=thread_id,
item={
"type": "message",
"role": "assistant",
"content": [{"type": "text", "text": full_response}],
}
)
# Send message done event
yield self._sse_event("message_done", {
"message_id": thread_id,
"role": "assistant"
})
except TimeoutError:
# [Task]: T034 - Handle timeout gracefully
logger.error(f"Agent execution timeout for thread {thread_id}")
yield self._sse_event("error", {
"message": "Request timed out",
"detail": "The AI assistant took too long to respond. Please try again."
})
except Exception as e:
logger.error(f"Agent execution error: {e}", exc_info=True)
yield self._sse_event("error", {"message": str(e)})
# Return streaming result
class StreamingResult:
def __init__(self, generator):
self.generator = generator
self.json = json.dumps({"thread_id": thread_id})
def __aiter__(self):
return self.generator()
return StreamingResult(stream_response())
def _extract_message_content(self, item: dict) -> str:
"""Extract text content from ChatKit item.
Args:
item: ChatKit message item
Returns:
Extracted text content
"""
content_array = item.get("content", [])
for content_block in content_array:
if content_block.get("type") == "text":
return content_block.get("text", "")
return ""
def _sse_event(self, event_type: str, data: dict) -> str:
"""Format data as Server-Sent Event.
[From]: specs/010-chatkit-migration/contracts/backend.md - SSE Event Types
Args:
event_type: Event type name
data: Event data payload
Returns:
Formatted SSE string
"""
import json
return f"event: {event_type}\ndata: {json.dumps(data)}\n\n"
# Create singleton instance (will be configured with tools in T019)
_server_instance: Optional[TaskManagerChatKitServer] = None
def get_chatkit_server(store: PostgresChatKitStore) -> TaskManagerChatKitServer:
"""Get or create ChatKit server singleton with agent configuration.
[From]: specs/010-chatkit-migration/contracts/backend.md
[Task]: T019 - Configure TaskAssistant agent with Gemini model and wrapped tools
Args:
store: PostgresChatKitStore instance
Returns:
ChatKit server instance with configured agent
"""
global _server_instance
if _server_instance is None:
_server_instance = TaskManagerChatKitServer(store)
# Configure the assistant agent with tools
# [From]: specs/010-chatkit-migration/tasks.md - T019
# [From]: specs/010-chatkit-migration/contracts/backend.md - Tool Contracts
from ai_agent.tool_wrappers import TOOL_FUNCTIONS
from core.config import get_settings
settings = get_settings()
# Create the TaskAssistant agent with Gemini model
# [From]: specs/010-chatkit-migration/research.md - Section 3
assistant_agent = Agent[AgentContext](
name="TaskAssistant",
model=settings.gemini_model or "gemini-2.0-flash-exp",
instructions="""You are a helpful task management assistant. You help users create, list, update, complete, and delete tasks through natural language.
Available tools:
- create_task: Create a new task with title, description, due date, priority, tags
- list_tasks: List all tasks with optional filters
- update_task: Update an existing task
- delete_task: Delete a task
- complete_task: Mark a task as completed or incomplete
- complete_all_tasks: Mark all tasks as completed (requires confirmation)
- delete_all_tasks: Delete all tasks (requires confirmation)
When users ask about tasks, use the appropriate tool. Always confirm destructive actions (complete_all_tasks, delete_all_tasks) by requiring the confirm parameter.
Be concise and helpful. If a user's request is unclear, ask for clarification.""",
tools=TOOL_FUNCTIONS,
)
_server_instance.set_agent(assistant_agent)
logger.info(f"ChatKit server initialized with {len(TOOL_FUNCTIONS)} tools and model {settings.gemini_model}")
return _server_instance
|