Spaces:
Sleeping
Sleeping
File size: 4,842 Bytes
6a9fed6 |
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 |
import os
import uuid
import logging
from datetime import datetime
from pathlib import Path
from fastapi import FastAPI, HTTPException, BackgroundTasks, Request
from fastapi.responses import StreamingResponse
from pydantic import BaseModel
from utils import (
sanitize_url,
crawl_documentation,
get_voice_prompt_style,
voice_map,
)
from ai_agents import Runner, setup_agents
from generate_audio import generate_audio
app = FastAPI()
Path("audio_outputs").mkdir(parents=True, exist_ok=True)
AUDIO_DIR = "audio_outputs"
logging.basicConfig(
filename="voice_agent.log",
filemode="w",
format="%(asctime)s | %(levelname)s | %(message)s",
level=logging.INFO,
)
logger = logging.getLogger(__name__)
class QueryRequest(BaseModel):
query: str
url: str = None
voice: str = None
file_text: str = None
from typing import Optional
class QueryResponse(BaseModel):
answer: str
audio_key: Optional[str] = None
sources: list = []
key_points: list[str] = []
@app.post("/process", response_model=QueryResponse)
async def process_query(req: QueryRequest, background_tasks: BackgroundTasks):
try:
start = datetime.now()
logger.info(f"🧠 Processing query: {req.query}")
logger.info(f"🌐 URL: {req.url}")
logger.info(f"📎 File text preview: {req.file_text[:100] if req.file_text else 'None'}")
logger.info(f"🎙️ Voice: {req.voice}")
key_points = []
if req.file_text:
from ai_agents import Agent
extract_agent = Agent(
name="KeyPointAgent",
instructions="Extract the 5–7 most important key points from this content. Respond only as a bullet list.",
model="gpt-4o"
)
key_points_raw = await extract_agent.run(req.file_text)
key_points = [line.strip('-•* ').strip() for line in key_points_raw.splitlines() if line.strip()]
if not key_points:
logger.info('⚠️ No bullet points detected from GPT, using fallback.')
key_points = [key_points_raw.strip()]
logger.info(f'🔎 Final key points: {key_points}')
if req.url:
try:
content = crawl_documentation(req.url)
context = f"{content}\n\nNow answer the user's question: {req.query}"
except Exception as e:
logger.warning(f"⚠️ URL crawl failed: {e}")
context = f"Answer the following using your general knowledge:\n\n{req.query}"
elif req.file_text:
context = f"{req.file_text}\n\nNow answer the user's question: {req.query}"
else:
context = f"Answer the following using your general knowledge:\n\n{req.query}"
tone = get_voice_prompt_style(req.voice or "")
if tone:
context = tone + "\n\n" + context
processor, _ = setup_agents()
logger.info("🧠 Sending context to GPT")
answer = await Runner.run(processor, context)
if not answer:
raise HTTPException(status_code=500, detail="No GPT response.")
logger.info(f"✅ GPT returned: {answer[:100]}...")
logger.info(f"🤖 GPT answer complete. ⏱️ {datetime.now() - start}")
audio_key = None
if req.voice and req.voice in voice_map:
voice_id = voice_map[req.voice]
audio_key = str(uuid.uuid4())
generate_audio(answer, voice_id, audio_key)
logger.info(f"🎙️ Audio generation triggered for voice: {req.voice}")
# ✅ Check if audio file actually exists
output_path = os.path.join(AUDIO_DIR, f"{audio_key}.mp3")
if not os.path.exists(output_path) or os.path.getsize(output_path) < 1000:
logger.warning("🛑 Audio generation failed or file is too small.")
audio_key = None
else:
logger.warning("🛑 Invalid voice")
return QueryResponse(answer=answer, audio_key=audio_key, sources=[], key_points=key_points)
except Exception as e:
logger.error(f"🔥 Internal error: {str(e)}")
import traceback
logger.error("".join(traceback.format_exception(None, e, e.__traceback__)))
raise HTTPException(status_code=500, detail=str(e))
@app.get("/get-audio/{key}")
async def get_audio(key: str, request: Request):
audio_path = os.path.join(AUDIO_DIR, f"{key}.mp3")
if not os.path.exists(audio_path):
raise HTTPException(status_code=404, detail="Audio not found")
if request.method == "HEAD":
return StreamingResponse(iter([]), status_code=200)
def iterfile():
with open(audio_path, mode="rb") as file:
yield from file
return StreamingResponse(iterfile(), media_type="audio/mpeg")
|