Spaces:
Sleeping
Sleeping
File size: 2,956 Bytes
bc9904d | 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 | from fastapi import APIRouter
from fastapi.responses import StreamingResponse
from src.models.agent_models import QuestionRequest
from src.agents import AGENT_MANAGER
from langchain_core.messages import HumanMessage, AIMessage
import re
router = APIRouter()
@router.post("/ask-stream")
def ask_stream(req: QuestionRequest):
agent = AGENT_MANAGER.get_agent(req.agent_id)
if not agent:
return {"error": "Agente no encontrado"}
if "get_context" in agent and "prompt" in agent:
chat_history = "\n".join([
f"Usuario: {msg.content}" if isinstance(msg, HumanMessage) else f"Asistente: {msg.content}"
for msg in agent["history"].messages
])
context_result = agent["get_context"](req.question)
# Handle both old and new context return formats
if isinstance(context_result, tuple):
context, language = context_result
else:
context = context_result
language = "ESPAÑOL" # Default to Spanish if no language detection
prompt_text = agent["prompt"].format(
context=context,
question=req.question,
chat_history=chat_history,
language=language
)
def generate():
full_response = ""
for chunk in agent["llm"].stream(prompt_text):
content = chunk.content
full_response += content
yield content
agent["history"].add_user_message(req.question)
agent["history"].add_ai_message(full_response)
return StreamingResponse(generate(), media_type="text/plain")
elif "agent" in agent and "history" in agent:
def generate():
inputs = {
"input": req.question,
"chat_history": agent["history"].messages,
"intermediate_steps": []
}
response = agent["agent"].invoke(inputs)
raw_output = getattr(response, "output", str(response)).strip()
raw_output = re.sub(r"return_values=\{.*?['\"]output['\"]:\s*['\"]", "", raw_output, flags=re.DOTALL)
raw_output = re.sub(r"['\"]\}\s*(log=.*)?", "", raw_output, flags=re.DOTALL)
lines = [line.strip() for line in raw_output.splitlines() if line.strip()]
lines_with_price = [line for line in lines if re.search(r"\$\d{3,6}", line)]
final_line = lines_with_price[-1] if lines_with_price else lines[-1] if lines else "⚠️ No se pudo generar respuesta."
print("💬 Respuesta final mostrada al usuario:")
print(final_line)
final_msg = f"\n💵 {final_line}"
agent["history"].add_user_message(req.question)
agent["history"].add_ai_message(final_msg)
yield final_msg
return StreamingResponse(generate(), media_type="text/plain")
return {"error": "Estructura de agente no válida"}
|