vosk文件转写

雨点打透心脏的1/2处 2022-09-02 13:53 7阅读 0赞

> vosk介绍以及安装，参考地址：[https://blog.csdn.net/qq\_35385687/article/details/119209189?spm=1001.2014.3001.5501][https_blog.csdn.net_qq_35385687_article_details_119209189_spm_1001.2014.3001.5501]  
> vosk实时语音识别，参考地址：[https://blog.csdn.net/qq\_35385687/article/details/119357825?spm=1001.2014.3001.5501][https_blog.csdn.net_qq_35385687_article_details_119357825_spm_1001.2014.3001.5501]

### 文章目录 ###

*   *  命令行模式
     *  文件上传方式

## 命令行模式 ##

#!/usr/bin/env python3
    
    import os
    import wave
    from vosk import Model, KaldiRecognizer, SetLogLevel
    
    SetLogLevel(0)
    
    if not os.path.exists("model"):
        print(
            "Please download the model from https://alphacephei.com/vosk/models and unpack as 'model' in the current folder.")
        exit(1)
    # sys.argv[1]
    wf = wave.open("./test.wav", "rb")
    if wf.getnchannels() != 1 or wf.getsampwidth() != 2 or wf.getcomptype() != "NONE":
        print("Audio file must be WAV format mono PCM.")
        exit(1)
    
    model = Model("model")
    rec = KaldiRecognizer(model, wf.getframerate())
    rec.SetWords(False)
    
    while True:
        data = wf.readframes(4000)
        if len(data) == 0:
            break
        if rec.AcceptWaveform(data):
            print("r" + rec.Result())
        else:
            print("p" + rec.PartialResult())
    
    print("f" + rec.FinalResult())

## 文件上传方式 ##

#!/usr/bin/env python
    # -*- coding:utf-8 -*-
    import os
    import wave
    from flask import Flask, request
    from vosk import Model, KaldiRecognizer, SetLogLevel
    
    app = Flask(__name__)
    download_floder = './upload/'
    SetLogLevel(0)
    
    if not os.path.exists("model"):
        print("请从'https://alphacephei.com/vosk/models'下载模型并解压到'model'文件夹")
        exit(1)
    
    
    # 文件类型检查
    def allow_file(filename):
        allow_list = ['wav', 'pcm']
        suffix = filename.split('.')
        return len(suffix) > 1 and suffix[1] in allow_list
    
    
    @app.route('/upload', methods=['POST'])
    def upload():
        f = request.files['file']
        if f and allow_file(f.filename):
            filePath = os.path.join(download_floder, f.filename)
            print(filePath)
            f.save(filePath)
    
            if f.filename.split('.')[1] == 'pcm':
                filePath = download_floder + f.filename.replace('pcm', 'wav')
                with open(filePath, 'rb') as pcmfile:
                    pcmdata = pcmfile.read()
                with wave.open(filePath, 'wb') as wavfile:
                    wavfile.setparams((1, 2, 16000, 0, 'NONE', 'NONE'))
                    wavfile.writeframes(pcmdata)
    
            print(filePath)
            wf = wave.open(filePath, "rb")
            if wf.getnchannels() != 1 or wf.getsampwidth() != 2 or wf.getcomptype() != "NONE":
                print("Audio file must be WAV format mono PCM.")
                exit(1)
            model = Model("model")
            rec = KaldiRecognizer(model, wf.getframerate())
            rec.SetWords(False)
    
            while True:
                data = wf.readframes(4000)
                if len(data) == 0:
                    break
                if rec.AcceptWaveform(data):
                    print(rec.Result())
                else:
                    print(rec.PartialResult())
    
            return rec.FinalResult()
    
    
    if __name__ == '__main__':
        if not os.path.exists(download_floder):
            os.makedirs(download_floder)
        app.run(debug=True, host='0.0.0.0')

[https_blog.csdn.net_qq_35385687_article_details_119209189_spm_1001.2014.3001.5501]: https://blog.csdn.net/qq_35385687/article/details/119209189?spm=1001.2014.3001.5501
[https_blog.csdn.net_qq_35385687_article_details_119357825_spm_1001.2014.3001.5501]: https://blog.csdn.net/qq_35385687/article/details/119357825?spm=1001.2014.3001.5501