mirror of
https://github.com/JefferyHcool/BiliNote.git
synced 2026-06-12 03:00:09 +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>
71 lines
2.6 KiB
Python
71 lines
2.6 KiB
Python
from abc import ABC
|
||
import os
|
||
|
||
from app.decorators.timeit import timeit
|
||
from app.models.transcriber_model import TranscriptResult, TranscriptSegment
|
||
from app.services.provider import ProviderService
|
||
from app.transcriber.base import Transcriber
|
||
from app.utils.openai_client import build_openai_client
|
||
import ffmpeg
|
||
import tempfile
|
||
from dotenv import load_dotenv
|
||
load_dotenv()
|
||
MAX_SIZE_MB = 18
|
||
MAX_SIZE_BYTES = MAX_SIZE_MB * 1024 * 1024
|
||
def compress_audio(input_path: str, target_bitrate='64k') -> str:
|
||
output_fd, output_path = tempfile.mkstemp(suffix=".mp3") # 临时输出文件
|
||
os.close(output_fd) # 关闭文件描述符，ffmpeg 会用路径操作
|
||
ffmpeg.input(input_path).output(output_path, audio_bitrate=target_bitrate).run(quiet=True, overwrite_output=True)
|
||
return output_path
|
||
|
||
class GroqTranscriber(Transcriber, ABC):
|
||
|
||
|
||
@timeit
|
||
def transcript(self, file_path: str) -> TranscriptResult:
|
||
file_size = os.path.getsize(file_path)
|
||
if file_size > MAX_SIZE_BYTES:
|
||
print(f"文件超过 {MAX_SIZE_MB}MB，开始压缩（当前 {round(file_size / (1024 * 1024), 2)}MB）...")
|
||
file_path = compress_audio(file_path)
|
||
print(f"压缩完成，临时路径：{file_path}")
|
||
provider = ProviderService.get_provider_by_id('groq')
|
||
|
||
if not provider:
|
||
raise Exception("Groq 供应商未配置,请配置以后使用。")
|
||
# build_openai_client 会校验 api_key 非空（空 key 会抛天书般的
|
||
# `Illegal header value b'Bearer '`），并自动注入全局代理
|
||
client = build_openai_client(
|
||
api_key=provider.get('api_key'),
|
||
base_url=provider.get('base_url'),
|
||
key_label="Groq 转写引擎的 API Key",
|
||
)
|
||
filename = file_path
|
||
|
||
with open(filename, "rb") as file:
|
||
transcription = client.audio.transcriptions.create(
|
||
file=(filename, file.read()),
|
||
model=os.getenv('GROQ_TRANSCRIBER_MODEL'),
|
||
response_format="verbose_json",
|
||
)
|
||
print(transcription.text)
|
||
print(transcription)
|
||
segments = []
|
||
full_text = ""
|
||
|
||
for seg in transcription.segments:
|
||
text = seg.text.strip()
|
||
full_text += text + " "
|
||
segments.append(TranscriptSegment(
|
||
start=seg.start,
|
||
end=seg.end,
|
||
text=text
|
||
))
|
||
|
||
result = TranscriptResult(
|
||
language=transcription.language,
|
||
full_text=full_text.strip(),
|
||
segments=segments,
|
||
raw=transcription.to_dict()
|
||
)
|
||
return result
|