Python之爬虫-全民k歌

import re
import os
import requests
from aip import AipSpeech
from pydub import AudioSegment

APP_ID = ''
API_KEY = ''
SECRET_KEY = ''


def save_mp4(filename, url):
    response = requests.get(url)
    data = response.text

    mp4_url = re.findall('"playurl":"(.*?)"', data)

    mp4_response = requests.get(mp4_url[0])
    mp4_data = mp4_response.content

    with open(filename, 'wb') as fw:
        fw.write(mp4_data)


def cut_speech(filename):
    file_name = filename
    sound = AudioSegment.from_wav(filename)
    start_time = "0:00"
    stop_time = "0:42"
    print("time:", start_time, "~", stop_time)
    start_time = (int(start_time.split(':')[0]) * 60 + int(start_time.split(':')[1])) * 1000
    stop_time = (int(stop_time.split(':')[0]) * 60 + int(stop_time.split(':')[1])) * 1000
    print("ms:", start_time, "~", stop_time)
    word = sound[start_time:stop_time]
    save_name = "word" + file_name[6:]
    print(save_name)

    word.export(save_name, format="mp3", tags={'artist': 'AppLeU0', 'album': save_name[:-4]})


def automatic_speech_recognition(filename):
    """语音识别"""
    # 调用语音识别接口
    client = AipSpeech(APP_ID, API_KEY, SECRET_KEY)

    # 把.m4a格式的文本转换成.pcm的二进制音频序列文本
    # f''是f-string格式化字符串
    os.system(f'ffmpeg -y  -i {filename}  -acodec pcm_s16le -f s16le -ac 1 -ar 16000 {filename}.pcm')

    # 打开.pcm格式的二进制音频序列文本
    with open(f'{filename}.pcm', 'rb') as f:
        data = f.read()

    # 语音识别。参数为：音频流,音频格式/pcm格式/wav格式,音频采样率支持8000-16000,{语言id，1536可以识别简单英文、1537不能识别简单英文}
    asr_result = client.asr(data, 'pcm', 16000, {'dev_pid': 1536})
    print('\nasr_result:{}'.format(asr_result))

    if asr_result.get('result') is None:
        return '', False
    else:
        return asr_result, True


if __name__ == '__main__':
    name = 'test'
    url = 'https://kg2.qq.com/node/play?s=GJH9sHGzgexrKGh-&shareuid=6a9a958c252a378836&topsource=a0_pn201001006_z1_u878907032_l1_t1558015859__'
    filename = f'{name}.m4a'
    save_mp4(filename, url)
    os.system(f'ffmpeg -y  -i {filename}  -acodec pcm_s16le -f s16le -ac 1 -ar 16000 {filename}.pcm')
    filename = f'{filename}.pcm'
    cut_speech(filename)
    # print('done...')
    # asr_result = automatic_speech_recognition(filename)
    # print(asr_result)