mirror of
https://github.com/JefferyHcool/BiliNote.git
synced 2026-06-12 19:20:00 +08:00
41f17592c2
- whisper: model.bin 截断/损坏时删目录重下重试一次，修「Unable to
open file model.bin」死循环；mlx 同样按 config.json 判完整性
- /generate_note 加就绪门禁：本地转写引擎模型没下好直接拦截，返回
reason=transcriber_model_not_ready，不让任务静默卡在首次下载
- 全局代理：新增 ProxyConfigManager（JSON 配置 + HTTP_PROXY env 兜底）
+ build_openai_client，统一注入代理到 LLM/Groq 客户端；yt-dlp 与
youtube-transcript-api 也走代理
- build_openai_client 校验 api_key 非空，空 key 给「xxx 的 API Key
未配置」而不是天书般的 Illegal header value b'Bearer '
- universal_gpt: 模型拒绝自定义 temperature（o1/o3/gpt-5 系列）时
就地去掉参数重试，不消耗重试预算
- connect_test 改用真实 chat completion 而非 /v1/models 探测
- main.py: lifespan 拆 [startup 1/5..5/5] 分段日志 + 异常清晰定位
- /sys_health 重构为结构化返回 {backend,ffmpeg,db,whisper_model}
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
322 lines
12 KiB
Python
322 lines
12 KiB
Python
# app/routers/note.py
|
||
import json
|
||
import os
|
||
import uuid
|
||
from pathlib import Path
|
||
from typing import Optional
|
||
from urllib.parse import urlparse
|
||
|
||
from fastapi import APIRouter, HTTPException, BackgroundTasks, UploadFile, File
|
||
from pydantic import BaseModel, validator, field_validator
|
||
from dataclasses import asdict
|
||
|
||
from app.db.video_task_dao import get_task_by_video
|
||
from app.enmus.exception import NoteErrorEnum
|
||
from app.enmus.note_enums import DownloadQuality
|
||
from app.exceptions.note import NoteError
|
||
from app.services.note import NoteGenerator, logger
|
||
from app.services.task_serial_executor import task_serial_executor
|
||
from app.utils.response import ResponseWrapper as R
|
||
from app.utils.url_parser import extract_video_id
|
||
from app.validators.video_url_validator import is_supported_video_url
|
||
from fastapi import APIRouter, Request, HTTPException
|
||
from fastapi.responses import StreamingResponse
|
||
import httpx
|
||
from app.enmus.task_status_enums import TaskStatus
|
||
|
||
# from app.services.downloader import download_raw_audio
|
||
# from app.services.whisperer import transcribe_audio
|
||
|
||
router = APIRouter()
|
||
|
||
|
||
class RecordRequest(BaseModel):
|
||
video_id: str
|
||
platform: str
|
||
|
||
|
||
class VideoRequest(BaseModel):
|
||
video_url: str
|
||
platform: str
|
||
quality: DownloadQuality
|
||
screenshot: Optional[bool] = False
|
||
link: Optional[bool] = False
|
||
model_name: str
|
||
provider_id: str
|
||
task_id: Optional[str] = None
|
||
format: Optional[list] = []
|
||
style: str = None
|
||
extras: Optional[str]=None
|
||
video_understanding: Optional[bool] = False
|
||
video_interval: Optional[int] = 0
|
||
grid_size: Optional[list] = []
|
||
# 客户端（如浏览器插件）已经在用户浏览器里抓到字幕，直接传给后端复用，
|
||
# 跳过 download_subtitles 和音频转写。形如：
|
||
# {"language": "zh", "full_text": "...", "segments": [{"start","end","text"}, ...]}
|
||
prefetched_transcript: Optional[dict] = None
|
||
|
||
@field_validator("video_url")
|
||
def validate_supported_url(cls, v):
|
||
url = str(v)
|
||
parsed = urlparse(url)
|
||
if parsed.scheme in ("http", "https"):
|
||
# 是网络链接，继续用原有平台校验
|
||
if not is_supported_video_url(url):
|
||
raise NoteError(code=NoteErrorEnum.PLATFORM_NOT_SUPPORTED.code,
|
||
message=NoteErrorEnum.PLATFORM_NOT_SUPPORTED.message)
|
||
|
||
return v
|
||
|
||
|
||
NOTE_OUTPUT_DIR = os.getenv("NOTE_OUTPUT_DIR", "note_results")
|
||
UPLOAD_DIR = "uploads"
|
||
|
||
|
||
def save_note_to_file(task_id: str, note):
|
||
os.makedirs(NOTE_OUTPUT_DIR, exist_ok=True)
|
||
with open(os.path.join(NOTE_OUTPUT_DIR, f"{task_id}.json"), "w", encoding="utf-8") as f:
|
||
json.dump(asdict(note), f, ensure_ascii=False, indent=2)
|
||
|
||
|
||
def _persist_prefetched_transcript(task_id: str, transcript: dict) -> None:
|
||
"""把客户端预取的字幕写到 NoteGenerator 期望的转写缓存文件里。
|
||
|
||
NoteGenerator.generate 会优先读 <task_id>_transcript.json，命中即跳过 download_subtitles
|
||
与音频转写流程。要求字段：language(可空)/full_text/segments[{start,end,text}]
|
||
"""
|
||
segments = transcript.get("segments") or []
|
||
cleaned_segments = []
|
||
for s in segments:
|
||
text = (s.get("text") or "").strip()
|
||
if not text:
|
||
continue
|
||
cleaned_segments.append({
|
||
"start": float(s.get("start", 0)),
|
||
"end": float(s.get("end", 0)),
|
||
"text": text,
|
||
})
|
||
if not cleaned_segments:
|
||
raise ValueError("prefetched_transcript 没有可用的 segments")
|
||
|
||
full_text = transcript.get("full_text") or " ".join(s["text"] for s in cleaned_segments)
|
||
payload = {
|
||
"language": transcript.get("language") or "zh",
|
||
"full_text": full_text,
|
||
"segments": cleaned_segments,
|
||
}
|
||
|
||
os.makedirs(NOTE_OUTPUT_DIR, exist_ok=True)
|
||
target = os.path.join(NOTE_OUTPUT_DIR, f"{task_id}_transcript.json")
|
||
with open(target, "w", encoding="utf-8") as f:
|
||
json.dump(payload, f, ensure_ascii=False, indent=2)
|
||
logger.info(f"已写入客户端预取字幕缓存: {target} ({len(cleaned_segments)} 段)")
|
||
|
||
|
||
def run_note_task(task_id: str, video_url: str, platform: str, quality: DownloadQuality,
|
||
link: bool = False, screenshot: bool = False, model_name: str = None, provider_id: str = None,
|
||
_format: list = None, style: str = None, extras: str = None, video_understanding: bool = False,
|
||
video_interval=0, grid_size=[]
|
||
):
|
||
|
||
if not model_name or not provider_id:
|
||
raise HTTPException(status_code=400, detail="请选择模型和提供者")
|
||
|
||
def _execute_note_task():
|
||
return NoteGenerator().generate(
|
||
video_url=video_url,
|
||
platform=platform,
|
||
quality=quality,
|
||
task_id=task_id,
|
||
model_name=model_name,
|
||
provider_id=provider_id,
|
||
link=link,
|
||
_format=_format,
|
||
style=style,
|
||
extras=extras,
|
||
screenshot=screenshot,
|
||
video_understanding=video_understanding,
|
||
video_interval=video_interval,
|
||
grid_size=grid_size,
|
||
)
|
||
|
||
logger.info(f"任务进入执行队列 (task_id={task_id})")
|
||
note = task_serial_executor.run(_execute_note_task)
|
||
logger.info(f"Note generated: {task_id}")
|
||
if not note or not note.markdown:
|
||
logger.warning(f"任务 {task_id} 执行失败，跳过保存")
|
||
return
|
||
save_note_to_file(task_id, note)
|
||
|
||
# 自动建立向量索引（用于 AI 问答），失败不影响笔记生成
|
||
try:
|
||
from app.services.vector_store import VectorStoreManager
|
||
VectorStoreManager().index_task(task_id)
|
||
except Exception as e:
|
||
logger.warning(f"向量索引失败（不影响笔记）: {e}")
|
||
|
||
|
||
@router.post('/delete_task')
|
||
def delete_task(data: RecordRequest):
|
||
try:
|
||
# TODO: 待持久化完成
|
||
# NoteGenerator().delete_note(video_id=data.video_id, platform=data.platform)
|
||
return R.success(msg='删除成功')
|
||
except Exception as e:
|
||
return R.error(msg=e)
|
||
|
||
|
||
@router.post("/upload")
|
||
async def upload(file: UploadFile = File(...)):
|
||
os.makedirs(UPLOAD_DIR, exist_ok=True)
|
||
file_location = os.path.join(UPLOAD_DIR, file.filename)
|
||
|
||
with open(file_location, "wb+") as f:
|
||
f.write(await file.read())
|
||
|
||
# 假设你静态目录挂载了 /uploads
|
||
return R.success({"url": f"/uploads/{file.filename}"})
|
||
|
||
|
||
@router.post("/generate_note")
|
||
def generate_note(data: VideoRequest, background_tasks: BackgroundTasks):
|
||
try:
|
||
# 就绪门禁：本地转写引擎（fast-whisper / mlx-whisper）必须等模型下载完才能跑视频，
|
||
# 否则任务会卡在首次下载（慢 / OOM / 截断），用户只看到一个静默失败的任务。
|
||
# 客户端已抓好字幕（prefetched_transcript）则不需要转写，跳过检查。
|
||
if not data.prefetched_transcript:
|
||
from app.services.transcriber_config_manager import TranscriberConfigManager
|
||
readiness = TranscriberConfigManager().is_model_ready()
|
||
if not readiness["ready"]:
|
||
logger.warning(f"拒绝 generate_note：{readiness['reason']}")
|
||
return R.error(
|
||
msg=readiness["reason"],
|
||
code=300102,
|
||
data={
|
||
"reason": "transcriber_model_not_ready",
|
||
"transcriber_type": readiness["transcriber_type"],
|
||
"model_size": readiness["model_size"],
|
||
"downloading": readiness["downloading"],
|
||
},
|
||
)
|
||
|
||
video_id = extract_video_id(data.video_url, data.platform)
|
||
# if not video_id:
|
||
# raise HTTPException(status_code=400, detail="无法提取视频 ID")
|
||
# existing = get_task_by_video(video_id, data.platform)
|
||
# if existing:
|
||
# return R.error(
|
||
# msg='笔记已生成，请勿重复发起',
|
||
#
|
||
# )
|
||
if data.task_id:
|
||
# 如果传了task_id，说明是重试！
|
||
task_id = data.task_id
|
||
logger.info(f"重试模式，复用已有 task_id={task_id}")
|
||
else:
|
||
# 正常新建任务
|
||
task_id = str(uuid.uuid4())
|
||
|
||
# 统一先写入 PENDING，表示已进入队列等待串行执行
|
||
NoteGenerator()._update_status(task_id, TaskStatus.PENDING)
|
||
|
||
# 客户端已经抓好字幕的话，写到转写缓存文件，NoteGenerator 的 cache-hit 逻辑会直接用上
|
||
if data.prefetched_transcript:
|
||
try:
|
||
_persist_prefetched_transcript(task_id, data.prefetched_transcript)
|
||
except Exception as e:
|
||
logger.warning(f"写入预取字幕失败 (task_id={task_id}): {e}")
|
||
|
||
background_tasks.add_task(run_note_task, task_id, data.video_url, data.platform, data.quality, data.link,
|
||
data.screenshot, data.model_name, data.provider_id, data.format, data.style,
|
||
data.extras, data.video_understanding, data.video_interval, data.grid_size)
|
||
return R.success({"task_id": task_id})
|
||
except Exception as e:
|
||
raise HTTPException(status_code=500, detail=str(e))
|
||
|
||
|
||
@router.get("/task_status/{task_id}")
|
||
def get_task_status(task_id: str):
|
||
status_path = os.path.join(NOTE_OUTPUT_DIR, f"{task_id}.status.json")
|
||
result_path = os.path.join(NOTE_OUTPUT_DIR, f"{task_id}.json")
|
||
|
||
# 优先读状态文件
|
||
if os.path.exists(status_path):
|
||
with open(status_path, "r", encoding="utf-8") as f:
|
||
status_content = json.load(f)
|
||
|
||
status = status_content.get("status")
|
||
message = status_content.get("message", "")
|
||
|
||
if status == TaskStatus.SUCCESS.value:
|
||
# 成功状态的话，继续读取最终笔记内容
|
||
if os.path.exists(result_path):
|
||
with open(result_path, "r", encoding="utf-8") as rf:
|
||
result_content = json.load(rf)
|
||
return R.success({
|
||
"status": status,
|
||
"result": result_content,
|
||
"message": message,
|
||
"task_id": task_id
|
||
})
|
||
else:
|
||
# 理论上不会出现，保险处理
|
||
return R.success({
|
||
"status": TaskStatus.PENDING.value,
|
||
"message": "任务完成，但结果文件未找到",
|
||
"task_id": task_id
|
||
})
|
||
|
||
if status == TaskStatus.FAILED.value:
|
||
return R.error(message or "任务失败", code=500)
|
||
|
||
# 处理中状态
|
||
return R.success({
|
||
"status": status,
|
||
"message": message,
|
||
"task_id": task_id
|
||
})
|
||
|
||
# 没有状态文件，但有结果
|
||
if os.path.exists(result_path):
|
||
with open(result_path, "r", encoding="utf-8") as f:
|
||
result_content = json.load(f)
|
||
return R.success({
|
||
"status": TaskStatus.SUCCESS.value,
|
||
"result": result_content,
|
||
"task_id": task_id
|
||
})
|
||
|
||
# 什么都没有，默认PENDING
|
||
return R.success({
|
||
"status": TaskStatus.PENDING.value,
|
||
"message": "任务排队中",
|
||
"task_id": task_id
|
||
})
|
||
|
||
|
||
@router.get("/image_proxy")
|
||
async def image_proxy(request: Request, url: str):
|
||
headers = {
|
||
"Referer": "https://www.bilibili.com/",
|
||
"User-Agent": request.headers.get("User-Agent", ""),
|
||
}
|
||
|
||
try:
|
||
async with httpx.AsyncClient(timeout=10.0) as client:
|
||
resp = await client.get(url, headers=headers)
|
||
|
||
if resp.status_code != 200:
|
||
raise HTTPException(status_code=resp.status_code, detail="图片获取失败")
|
||
|
||
content_type = resp.headers.get("Content-Type", "image/jpeg")
|
||
return StreamingResponse(
|
||
resp.aiter_bytes(),
|
||
media_type=content_type,
|
||
headers={
|
||
"Cache-Control": "public, max-age=86400", # 缓存一天
|
||
"Content-Type": content_type,
|
||
}
|
||
)
|
||
except Exception as e:
|
||
raise HTTPException(status_code=500, detail=str(e))
|