python
复制代码
import re
import json
from urllib.parse import urlparse, parse_qs
import requests
import dashscope
from dashscope.audio.asr import Transcription
"""
Each file needs to export a function named `handler`. This function is the entrance to the Tool.
Parameters:
args: parameters of the entry function.
args.input - input parameters, you can get test input value by args.input.xxx.
args.logger - logger instance used to print logs, injected by runtime.
Remember to fill in input/output in Metadata, it helps LLM to recognize and use tool.
Return:
The return data of the function, which should match the declared output parameters.
"""
def get_video_id(text_or_url):
try:
# 1.提取视频地址
# 格式1：4.66 B@t.EH 08/24 VYM:/ 有些私活真裁员降薪 https://v.douyin.com/i5WJFJWv/ 复制
# 格式2：https://www.douyin.com/user/self?from_tab_name=main&modal_id=7452662497322077490
# 格式3：https://www.douyin.com/video/7452662497322077490?modeFrom=userPost&secUid=MS4wLjABAAAACdnpzvOcEtvmZ8h9
match_list = re.findall(r'https://(?:[a-zA-Z]|[0-9]|[$-_@.&+]|[!*\\(\\),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+',
text_or_url)
if not match_list:
return False, "短视频地址格式错误"
video_url = match_list[0]
# 2.处理 短连接
if video_url.startswith("https://v.douyin.com"):
res = requests.get(video_url, allow_redirects=False)
video_url = res.headers["Location"]
parsed_url = urlparse(video_url)
query_params = parse_qs(parsed_url.query)
# 3.地址GET参数处的modal_id读取
modal_id_list = query_params.get("modal_id")
if modal_id_list:
video_id = modal_id_list[0]
else:
path_list = parsed_url.path.strip("/").split("/")
video_id = path_list[-1]
return True, video_id
except Exception as e:
return False, "短视频ID提取失败"
def get_video_info(video_id):
try:
# 拼接生成m端地址
res = requests.get(
url=f"https://m.douyin.com/share/video/{video_id}",
headers={
"user-agent": "Mozilla/5.0 (iPhone; CPU iPhone OS 16_6 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Mobile/15E148 Safari/604.1"
}
)
# 正则提取 视频字典信息
match_list = re.findall("window._ROUTER_DATA = (.*)</script>", res.text)
data_dict = json.loads(match_list[0])
video_info = data_dict['loaderData']["video_(id)/page"]["videoInfoRes"]['item_list'][0]
# 视频信息
nickname = video_info["author"]['nickname']
digg_count = video_info["statistics"]['digg_count']
comment_count = video_info["statistics"]['comment_count']
desc = video_info["desc"]
video_url = video_info["video"]['play_addr']["url_list"][0]
return True, (nickname, desc, video_url, digg_count, comment_count)
except Exception as e:
return False, "视频信息提取失败"
def get_text_by_ali(api_key, video_url):
try:
# 1.阿里云百炼账号，创建并获取api_key
dashscope.api_key = api_key#注册后输入
# 2.调用模型提取文案
# 文档：https://help.aliyun.com/zh/model-studio/developer-reference/paraformer-recorded-speech-recognition-python-api?spm=a2c4g.11186623.help-menu-2400256.d_3_3_7_3_3_1.7ebf7f01D3wF0M
# https://help.aliyun.com/zh/model-studio/developer-reference/paraformer-recorded-speech-recognition-python-api?spm=0.0.0.i1
response = Transcription.call(
model='paraformer-v1',
file_urls=[video_url],
language_hints=['zh', 'en'] # "language_hints"只支持paraformer-v2和paraformer-realtime-v2模型
)
# 3.读取结果 {"transcription_url":"提取文案结果的URL地址",...}
status_code = response.get('status_code')
if status_code != 200:
return False, response.get('message', "通义千问提取失败")
res_dict = response["output"]["results"][0]
transcription_url = res_dict["transcription_url"]
# 4.获取文案
res = requests.get(transcription_url)
res_dict = res.json()
text = res_dict["transcripts"][0]["text"]
return True, text
except Exception as e:
return False, "文案提取异常" + str(e)
def main():
douyin_url = input.url
ali_bailian_api_key = input.ali_bailian_api_key
if not douyin_url.startswith("http"):
douyin_url = re.search("https?://[^\\s]+", douyin_url).group()
# 1.获取短视频ID
status, video_id = get_video_id(douyin_url)
# args.logger.info(video_id)
if not status:
return {"status": False, "error": video_id}
# 2.获取视频信息（标题、作者、视频地址）
status, video_info = get_video_info(video_id)
if not status:
return {"status": False, "error": str(video_info)}
nickname, desc, video_url, digg_count, comment_count = video_info
# 3.视频地址转换
# https://aweme.snssdk.com/aweme/v1/playwm/?video_id=v0300f
# https://aweme.snssdk.com/aweme/v1/play/?video_id=v0300f
video_url = video_url.replace("playwm", "play")
# 4.调用 阿里云模型，提取视频文案
# args.logger.info(video_url)
status, text = get_text_by_ali(ali_bailian_api_key, video_url)
if not status:
return {"status": False, "error": text}
return {"status": True, "data": {"nickname": nickname, "desc": desc, "text": text, "digg_count": digg_count,
"comment_count": comment_count}}