Luxx/luxx/services/stream_service.py

473 lines
17 KiB
Python

"""Stream Service - handles SSE streaming logic"""
import json
import logging
from typing import List, Dict, Any, Optional, AsyncGenerator
from luxx.services.llm_service import LLMService
from luxx.services.message_service import MessageService
from luxx.tools.executor import ToolExecutor
from luxx.tools.core import registry
logger = logging.getLogger(__name__)
# Maximum iterations to prevent infinite loops
MAX_ITERATIONS = 10
def _sse_event(event: str, data: dict) -> str:
"""Format a Server-Sent Event string."""
return f"event: {event}\ndata: {json.dumps(data, ensure_ascii=False)}\n\n"
class StreamContext:
"""
Context for streaming response state management.
Encapsulates all state needed during a streaming session.
"""
def __init__(
self,
step_index: int = 0,
current_step_id: str = None,
current_step_idx: int = None,
current_stream_type: str = None,
full_content: str = "",
full_thinking: str = ""
):
self.step_index = step_index
self.current_step_id = current_step_id
self.current_step_idx = current_step_idx
self.current_stream_type = current_stream_type
self.full_content = full_content
self.full_thinking = full_thinking
self.all_steps: List[Dict] = []
self.all_tool_calls: List[Dict] = []
self.all_tool_results: List[Dict] = []
self.tool_calls_list: List[Dict] = []
def reset_iteration(self):
"""Reset streaming step tracker for new iteration."""
self.current_step_id = None
self.current_step_idx = None
self.current_stream_type = None
self.full_content = ""
self.full_thinking = ""
self.tool_calls_list = []
def start_stream_step(self, step_type: str) -> str:
"""Start a new streaming step. Returns the step_id."""
self.current_step_idx = self.step_index
self.current_step_id = f"step-{self.step_index}"
self.current_stream_type = step_type
self.step_index += 1
return self.current_step_id
def yield_stream_step(self, step_type: str, content: str) -> str:
"""Yield a streaming step event."""
return _sse_event("process_step", {
"step": {
"id": self.current_step_id,
"index": self.current_step_idx,
"type": step_type,
"content": content
}
})
def save_streaming_step(self):
"""Save the current streaming step to all_steps."""
if self.current_step_id is None:
return
if self.current_stream_type == "thinking":
self.all_steps.append({
"id": self.current_step_id,
"index": self.current_step_idx,
"type": "thinking",
"content": self.full_thinking
})
elif self.current_stream_type == "text":
self.all_steps.append({
"id": self.current_step_id,
"index": self.current_step_idx,
"type": "text",
"content": self.full_content
})
def handle_thinking_stream(self, delta: Dict) -> Optional[str]:
"""Handle reasoning/thinking delta. Returns SSE string if yielded."""
reasoning = delta.get("reasoning_content", "")
if not reasoning:
return None
prev_len = len(self.full_thinking)
self.full_thinking += reasoning
if prev_len == 0:
self.start_stream_step("thinking")
return self.yield_stream_step("thinking", self.full_thinking)
def handle_text_stream(self, delta: Dict) -> Optional[str]:
"""Handle content delta. Returns SSE string if yielded."""
content = delta.get("content", "")
if not content:
return None
prev_len = len(self.full_content)
self.full_content += content
if prev_len == 0:
self.start_stream_step("text")
return self.yield_stream_step("text", self.full_content)
def handle_tool_calls(self) -> tuple:
"""Handle tool calls accumulation. Returns (step_ids, steps, sse_strings)."""
tool_call_step_ids = []
tool_call_steps = []
yield_objs = []
for tc in self.tool_calls_list:
call_step_idx = self.step_index
call_step_id = f"step-{self.step_index}"
tool_call_step_ids.append(call_step_id)
self.step_index += 1
call_step = {
"id": call_step_id,
"index": call_step_idx,
"type": "tool_call",
"id_ref": tc.get("id", ""),
"name": tc["function"]["name"],
"arguments": tc["function"]["arguments"]
}
tool_call_steps.append(call_step)
yield_objs.append(_sse_event("process_step", {"step": call_step}))
return tool_call_step_ids, tool_call_steps, yield_objs
def handle_tool_result(self, tool_result: Dict, tool_call_step_id: str) -> tuple:
"""Handle single tool result. Returns (result_step, sse_string)."""
result_step_idx = self.step_index
result_step_id = f"step-{self.step_index}"
self.step_index += 1
content = tool_result.get("content", "")
success = True
try:
content_obj = json.loads(content)
if isinstance(content_obj, dict):
success = content_obj.get("success", True)
except (json.JSONDecodeError, TypeError):
pass
result_step = {
"id": result_step_id,
"index": result_step_idx,
"type": "tool_result",
"id_ref": tool_call_step_id,
"name": tool_result.get("name", ""),
"content": content,
"success": success
}
return result_step, _sse_event("process_step", {"step": result_step})
class StreamService:
"""
Service for handling streaming response logic.
Separated from ChatService for better separation of concerns.
"""
def __init__(
self,
llm_service: LLMService = None,
message_service: MessageService = None,
tool_executor: ToolExecutor = None
):
self.llm_service = llm_service or LLMService()
self.message_service = message_service or MessageService()
self.tool_executor = tool_executor or ToolExecutor()
def build_tool_context(
self,
workspace: str = None,
user_id: int = None,
username: str = None,
user_permission_level: int = 1
) -> Dict[str, Any]:
"""Build context dict for tool execution."""
return {
"workspace": workspace,
"user_id": user_id,
"username": username,
"user_permission_level": user_permission_level
}
def filter_tools(self, enabled_tools: List[str]) -> List[Dict]:
"""Filter tools by enabled list."""
if not enabled_tools:
return []
return [
t for t in registry.list_all()
if t.get("function", {}).get("name") in enabled_tools
]
async def stream(
self,
messages: List[Dict],
model: str,
tools: List[Dict],
temperature: float,
max_tokens: int,
thinking_enabled: bool,
llm_client=None,
conversation=None,
provider_id: int = None,
conversation_id: int = None,
workspace: str = None,
user_id: int = None,
username: str = None,
user_permission_level: int = 1
) -> AsyncGenerator[str, None]:
"""
Core streaming logic.
Args:
messages: Message list with conversation history
model: Model name
tools: Tool definitions
temperature: Sampling temperature
max_tokens: Max tokens
thinking_enabled: Enable reasoning
provider_id: LLM provider ID
conversation_id: Conversation ID for saving
workspace: Workspace path
user_id: User ID
username: Username
user_permission_level: Permission level
Yields:
SSE event strings
"""
# Get LLM client - use provided client or create from conversation/provider
llm = llm_client if llm_client else self.llm_service.get_client(
conversation=conversation, provider_id=provider_id
)[0]
# Token usage tracking
total_usage = {"prompt_tokens": 0, "completion_tokens": 0, "total_tokens": 0}
actual_token_count = 0
# Streaming context
ctx = StreamContext()
# Tool execution context
tool_context = self.build_tool_context(
workspace, user_id, username, user_permission_level
)
try:
for _ in range(MAX_ITERATIONS):
ctx.reset_iteration()
async for sse_line in llm.stream_call(
model=model,
messages=messages,
tools=tools,
temperature=temperature,
max_tokens=max_tokens or 8192,
thinking_enabled=thinking_enabled
):
# Parse SSE line
event_type, data_str = self._parse_sse_line(sse_line)
if data_str is None:
continue
# Handle error events
if event_type == 'error':
error_data = self._parse_json(data_str)
content = error_data.get("content", "Unknown error") if error_data else data_str
yield _sse_event("error", {"content": content})
return
# Parse data
chunk = self._parse_json(data_str)
if chunk is None:
yield _sse_event("error", {"content": f"Failed to parse: {data_str}"})
return
# Extract usage info
if "usage" in chunk:
usage = chunk["usage"]
total_usage["prompt_tokens"] = usage.get("prompt_tokens", 0)
total_usage["completion_tokens"] = usage.get("completion_tokens", 0)
total_usage["total_tokens"] = usage.get("total_tokens", 0)
# Check for error in response
if "error" in chunk:
error_msg = chunk["error"].get("message", str(chunk["error"]))
yield _sse_event("error", {"content": f"API Error: {error_msg}"})
return
# Get delta
choices = chunk.get("choices", [])
if not choices:
# Handle non-standard responses
content = chunk.get("content") or chunk.get("message", {}).get("content", "")
if content:
prev_len = len(ctx.full_content)
ctx.full_content += content
if prev_len == 0:
ctx.start_stream_step("text")
yield _sse_event("process_step", {
"step": {
"id": ctx.current_step_id if prev_len == 0 else f"step-{ctx.step_index - 1}",
"index": ctx.current_step_idx if prev_len == 0 else ctx.step_index - 1,
"type": "text",
"content": ctx.full_content
}
})
continue
delta = choices[0].get("delta", {})
# Handle thinking and text streams
yield_obj = ctx.handle_thinking_stream(delta)
if yield_obj:
yield yield_obj
yield_obj = ctx.handle_text_stream(delta)
if yield_obj:
yield yield_obj
# Accumulate tool calls
self._accumulate_tool_calls(ctx, delta)
# Save streaming step
ctx.save_streaming_step()
# Handle tool calls
if ctx.tool_calls_list:
# Yield tool execution results
async for event in self._handle_tool_execution(ctx, messages, tool_context):
yield event
continue
# No tool calls - final iteration
msg_id = self.message_service.create_message_id()
actual_token_count = total_usage.get("completion_tokens", 0) or len(ctx.full_content) // 4
logger.info(f"[TOKEN] total_usage: {total_usage}, actual_token_count: {actual_token_count}")
if conversation_id:
self.message_service.save_assistant_message(
conversation_id, msg_id, ctx.full_content,
ctx.all_tool_calls, ctx.all_tool_results, ctx.all_steps,
actual_token_count, total_usage
)
yield _sse_event("done", {
"message_id": msg_id,
"token_count": actual_token_count,
"usage": total_usage
})
return
# Max iterations exceeded
if conversation_id and (ctx.full_content or ctx.all_tool_calls):
msg_id = self.message_service.create_message_id()
self.message_service.save_assistant_message(
conversation_id, msg_id, ctx.full_content,
ctx.all_tool_calls, ctx.all_tool_results, ctx.all_steps,
actual_token_count, total_usage
)
yield _sse_event("error", {"content": "Exceeded maximum tool call iterations"})
except Exception as e:
logger.error(f"Stream error: {e}")
yield _sse_event("error", {"content": str(e)})
def _parse_sse_line(self, sse_line: str) -> tuple:
"""Parse SSE line. Returns (event_type, data_str)."""
event_type = None
data_str = None
for line in sse_line.strip().split('\n'):
if line.startswith('event: '):
event_type = line[7:].strip()
elif line.startswith('data: '):
data_str = line[6:].strip()
return event_type, data_str
def _parse_json(self, data_str: str) -> Optional[Dict]:
"""Parse JSON string safely."""
try:
return json.loads(data_str)
except json.JSONDecodeError:
return None
def _accumulate_tool_calls(self, ctx: StreamContext, delta: Dict):
"""Accumulate tool calls from delta."""
tool_calls_delta = delta.get("tool_calls", [])
for tc in tool_calls_delta:
idx = tc.get("index", 0)
if idx >= len(ctx.tool_calls_list):
ctx.tool_calls_list.append({
"id": tc.get("id", ""),
"type": "function",
"function": {"name": "", "arguments": ""}
})
func = tc.get("function", {})
if func.get("name"):
ctx.tool_calls_list[idx]["function"]["name"] += func["name"]
if func.get("arguments"):
ctx.tool_calls_list[idx]["function"]["arguments"] += func["arguments"]
async def _handle_tool_execution(
self,
ctx: StreamContext,
messages: List[Dict],
tool_context: Dict[str, Any]
) -> AsyncGenerator[str, None]:
"""Handle tool execution for one iteration. Yields SSE events."""
ctx.all_tool_calls.extend(ctx.tool_calls_list)
# Yield tool call steps
tool_call_step_ids, tool_call_steps, yield_objs = ctx.handle_tool_calls()
ctx.all_steps.extend(tool_call_steps)
for yield_obj in yield_objs:
yield yield_obj
# Execute tools
tool_results = self.tool_executor.process_tool_calls_parallel(
ctx.tool_calls_list, tool_context
)
# Yield tool result steps
for i, tr in enumerate(tool_results):
tool_call_step_id = tool_call_step_ids[i] if i < len(tool_call_step_ids) else f"step-{i}"
result_step, yield_obj = ctx.handle_tool_result(tr, tool_call_step_id)
ctx.all_steps.append(result_step)
yield yield_obj
ctx.all_tool_results.append({
"role": "tool",
"tool_call_id": tr.get("tool_call_id", ""),
"content": tr.get("content", "")
})
# Add messages for next iteration
messages.append({
"role": "assistant",
"content": ctx.full_content or "",
"tool_calls": ctx.tool_calls_list
})
messages.extend(ctx.all_tool_results[-len(tool_results):])
ctx.all_tool_results = []
# Global service instance
stream_service = StreamService()