Spaces:
Running
Running
File size: 3,769 Bytes
3b13b0e |
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 |
import os
import traceback
import json
from openai import OpenAI
from pydantic import BaseModel
from typing import List
from app.utils import utils
from app.services.subtitle import extract_audio_and_create_subtitle
class Step(BaseModel):
timestamp: str
picture: str
narration: str
OST: int
new_timestamp: str
class MathReasoning(BaseModel):
result: List[Step]
def chat_with_qwen(prompt: str, system_message: str, subtitle_path: str) -> str:
"""
与通义千问AI模型进行对话
Args:
prompt (str): 用户输入的问题或提示
system_message (str): 系统提示信息，用于设定AI助手的行为。默认为"You are a helpful assistant."
subtitle_path (str): 字幕文件路径
Returns:
str: AI助手的回复内容
Raises:
Exception: 当API调用失败时抛出异常
"""
try:
client = OpenAI(
api_key="sk-a1acd853d88d41d3ae92777d7bfa2612",
base_url="https://dashscope.aliyuncs.com/compatible-mode/v1",
)
# 读取字幕文件
with open(subtitle_path, "r", encoding="utf-8") as file:
subtitle_content = file.read()
completion = client.chat.completions.create(
model="qwen-turbo-2024-11-01",
messages=[
{'role': 'system', 'content': system_message},
{'role': 'user', 'content': prompt + subtitle_content}
]
)
return completion.choices[0].message.content
except Exception as e:
error_message = f"调用千问API时发生错误：{str(e)}"
print(error_message)
print("请参考文档：https://help.aliyun.com/zh/model-studio/developer-reference/error-code")
raise Exception(error_message)
# 使用示例
if __name__ == "__main__":
try:
video_path = utils.video_dir("duanju_yuansp.mp4")
# # 判断视频是否存在
# if not os.path.exists(video_path):
# print(f"视频文件不存在：{video_path}")
# exit(1)
# 提取字幕
subtitle_path = os.path.join(utils.video_dir(""), f"duanju_yuan.srt")
extract_audio_and_create_subtitle(video_file=video_path, subtitle_file=subtitle_path)
# 分析字幕
system_message = """
你是一个视频srt字幕分析剪辑器, 输入视频的srt字幕, 分析其中的精彩且尽可能连续的片段并裁剪出来, 注意确保文字与时间戳的正确匹配。
输出需严格按照如下 json 格式:
[
{
"timestamp": "00:00:50,020-00,01:44,000",
"picture": "画面1",
"narration": "播放原声",
"OST": 0,
"new_timestamp": "00:00:00,000-00:00:54,020"
},
{
"timestamp": "01:49-02:30",
"picture": "画面2",
"narration": "播放原声",
"OST": 2,
"new_timestamp": "00:54-01:35"
},
]
"""
prompt = "字幕如下：\n"
response = chat_with_qwen(prompt, system_message, subtitle_path)
print(response)
# 保存json，注意json中是时间戳需要转换为 分:秒(现在的时间是 "timestamp": "00:00:00,020-00:00:01,660", 需要转换为 "timestamp": "00:00-01:66")
# response = json.loads(response)
# for item in response:
# item["timestamp"] = item["timestamp"].replace(":", "-")
# with open(os.path.join(utils.video_dir(""), "duanju_yuan.json"), "w", encoding="utf-8") as file:
# json.dump(response, file, ensure_ascii=False)
except Exception as e:
print(traceback.format_exc())
|