speech.py 文件源码

def recognize(self, path, lang="zh-CN"):
        if isinstance(path, str):
            file = open(path, 'rb')
        else:
            return ["ERROR!", "File must by a path string."]
        if lang not in self.lang_list:
            return ["ERROR!", "Invalid language."]

        audio = pydub.AudioSegment.from_file(file)
        audio = audio.set_frame_rate(16000)
        audio.export("%s.wav" % path, format="wav")
        header = {
            "Authorization": "Bearer %s" % self.access_token,
            "Content-Type": "audio/wav; samplerate=16000"
        }
        d = {
            "version": "3.0",
            "requestid": str(uuid.uuid1()),
            "appID": "D4D52672-91D7-4C74-8AD8-42B1D98141A5",
            "format": "json",
            "locale": lang,
            "device.os": "Telegram",
            "scenarios": "ulm",
            "instanceid": uuid.uuid3(uuid.NAMESPACE_DNS, 'com.1a23.eh_telegram_master'),
            "maxnbest": 5
        }
        with open("%s.wav" % path, 'rb') as f:
            r = requests.post("https://speech.platform.bing.com/recognize", params=d, data=f.read(), headers=header)
        os.remove("%s.wav" % path)
        try:
            rjson = r.json()
        except:
            return ["ERROR!", r.text]
        if r.status_code == 200:
            return [i['name'] for i in rjson['results']]
        else:
            return ["ERROR!", r.text]