视/音频转换代码

import requests
import re
import json
import io
headers={
    "User-Agent":"Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/69.0.3497.100 Safari/537.36",
    "referer": "https://message.bilibili.com/",#这段代码的意思是你从哪儿获得这个网址的(换一句话讲,谁推荐你去访问这个网址的).有了它就能够正常访问.
}
def extract_content(text, symbol):
    pattern = r'%s(.*?)%s' % (symbol, symbol)
    result = re.findall(pattern, text)
    return result
def get_url_html(url):
    req=requests.get(url,headers=headers)
    htmltext=req.text
    get_json(htmltext)
def get_json(htmltxt):
    r=re.findall(r'<script>window.__playinfo__=(.*?)</script>',htmltxt)[0]
    js=json.loads(r)
    audiourl=js["data"]["dash"]["audio"][0]["base_url"]
    videourl=js["data"]["dash"]["video"][0]["base_url"]
    download(audiourl)
def download(audiourl):
    res=requests.get(url=audiourl,headers=headers)
    print('爬取中，等待....')
    with open(mp3,"wb") as f:
        f.write(res.content)
    print('爬取完毕！')
if __name__=="__main__":
    """
    爬取音频，获取视频网址，下载其中音频到本地
    """
    for i in range(1):
       
        text="【"+str(i+1)+"】"+"https://www.bilibili.com/video/BV1Aq421A7Cb/?spm_id_from=333.337.search-card.all.click&vd_source=42e8b3d0f4cd4b9549473a905d860f79"
        symbol = ["【","】"]
        result = (extract_content(text, symbol))[0]
        print('文件名为:'+result)  # 输出: ['World']
        url_regex = r'http[s]?://(?:[a-zA-Z]|[0-9]|[$-_@.&amp;+]|[!*\(\),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+'
        urls = (re.findall(url_regex, text))[0]
        print('网址:'+urls)
        mp3="./output_path/"+result+'.wav'
        get_url_html(urls)




from ffmpy import FFmpeg
import os
def audio_transfor(audio_path: str, output_dir: str):
    ext = os.path.basename(audio_path).strip().split('.')[-1]
    if ext != 'mp3':
        raise Exception('format is not mp3')
    result = os.path.join(output_dir, '{}.{}'.format(os.path.basename(audio_path).strip().split('.')[0], 'wav'))
    filter_cmd = '-f wav -ac 1 -ar 16000'
    ff = FFmpeg(
        inputs={
            audio_path: None}, outputs={
            result: filter_cmd})
    print(ff.cmd)
    ff.run()
    return result
def handle(mp3_audio_folder: str, mav_output_folder: str):
    """
    mp3转wav
    @param mp3_audio_folder MP3文件所在文件夹
    @param mav_output_folder 保存文件夹路径
    """
    for x in os.listdir(mp3_audio_folder):
        audio_transfor(os.path.join(mp3_audio_folder, x), mav_output_folder)
        os.remove(os.path.join(mp3_audio_folder, x))




import shutil 
import os
def move_folder(movabs_path,rawabs_path):
    """
    文件/文件夹复制
    @param movabs_path 目标文件的根目录
    @param rawabs_path 移动文件的根目录
    """
    # movabs_path = '/ML-A100/team/infra/project_human_data/wangningbo/'
    # rawabs_path = '/ML-A100/team/infra/project_human_data/tujiexu/bak'
    for item in os.listdir(movabs_path):
        if item in [".vscode","Anaconda"]:
            continue
        s = os.path.join(movabs_path, item)
        d = os.path.join(rawabs_path, item)
        if os.path.isdir(s):
            print(s)
            shutil.copytree(s, d, True, None)
        else:
            print(d)
            shutil.copy2(s, d)


import subprocess
import os
def cut_audio(audio_folder,output_folder):
    """
    音频截取
    @param pkl_folder pkl文件所在文件夹路径
    @param txt_folder 要保存的txt文件的文件夹路径
    """
    files = os.listdir(audio_folder)
    for file in files:
        audio_path = os.path.join(audio_folder,file)
        output_path = os.path.join(output_folder,file)
        # 截取从第30秒到结束的音频部分
        command = ['ffmpeg', '-i', audio_path, '-ss', '27', '-acodec', 'copy', output_path]
        subprocess.call(command)




import pickle
import os
def pkl2txt(pkl_folder,txt_folder):
    """
    SD的pkl文件转txt
    @param pkl_folder pkl文件所在文件夹路径
    @param txt_folder 要保存的txt文件的文件夹路径
    """
    files=os.listdir(pkl_folder)
    for file in files:
        file_path=os.path.join(pkl_folder,file)
        with open(file_path,"rb") as f:
            data=pickle.load(f)
            with open(txt_folder+file.replace("pkl","txt"),'a',encoding="UTF-8") as f:
                for item in data:
                    f.write("%s [%.3fs -> %.3fs] \n" % (item[0],item[1],item[2]))
            f.close()



from moviepy.editor import *
def v2a(filder_path,audio_path):
    """
    mp4转mp3或wav，视频转音频
    @param filder_path 视频所在文件夹路径
    @param audio_path 要保存音频的文件夹路径
    """
    import os
    files=os.listdir(filder_path)
    for file in files:
        path=os.path.join(filder_path,file)
        video = VideoFileClip(path)
        audio = video.audio
        audio.write_audiofile(os.path.join(audio_path,file.split(".")[0]+".wav"))