Google speech to text streamingマイク入力からの文字お越しの実現

前提・実現したいこと

初心者です。お世話になります。
python3.7.4
windows 10

google speech to text というapiのstreaming音声文字お越しをしたいです。
正確にはマイクからの入力を文字にしたいです。

公式ドキュメントの
（https://cloud.google.com/speech-to-text/docs/streaming-recognize?hl=ja#speech-streaming-mic-recognize-python）
「音声ストリームでのストリーミング音声認識の実行」の最中です。
こちら↓
（https://qiita.com/hamham/items/3733ac8cd9e3d7b9ccae）
を参考にしながら、C直下に公式ドキュメントをコピペしたtranscribe_streaming_mic.pyを作成し、streamを付けてコマンドラインから実行しました。

どうぞよろしくお願いいたします

発生している問題・エラーメッセージ

C:\>transcribe_streaming_mic.py stream
Traceback (most recent call last):
  File "C:\transcribe_streaming_mic.py", line 6, in <module>
    from google.cloud import speech
ImportError: No module named google.cloud

似た質問の回答のpip(sudoなし)をおこないインストールは成功しました
（https://teratail.com/questions/123983）
「Successfully installed cachetools-3.1.1 certifi-2019.9.11 chardet-3.0.4 google-api-core-1.14.3 google-auth-1.7.0 google-cloud-speech-1.2.0 googleapis-common-protos-1.6.0 grpcio-1.25.0 idna-2.8 protobuf-3.10.0 pyasn1-0.4.7 pyasn1-modules-0.2.7 requests-2.22.0 rsa-4.0 six-1.13.0 urllib3-1.25.6」

コマンド：(sudo) pip install google-cloud-speech

しかしもう一度C:>transcribe_streaming_mic.py stream　コマンドを実行してもまったく同じエラーがでます。

該当のソースコード

python
1from __future__ import division
2
3import re
4import sys
5
6from google.cloud import speech
7from google.cloud.speech import enums
8from google.cloud.speech import types
9import pyaudio
10from six.moves import queue
11
12# Audio recording parameters
13RATE = 16000
14CHUNK = int(RATE / 10)  # 100ms
15
16class MicrophoneStream(object):
17    """Opens a recording stream as a generator yielding the audio chunks."""
18    def __init__(self, rate, chunk):
19        self._rate = rate
20        self._chunk = chunk
21
22        # Create a thread-safe buffer of audio data
23        self._buff = queue.Queue()
24        self.closed = True
25
26    def __enter__(self):
27        self._audio_interface = pyaudio.PyAudio()
28        self._audio_stream = self._audio_interface.open(
29            format=pyaudio.paInt16,
30            # The API currently only supports 1-channel (mono) audio
31            # https://goo.gl/z757pE
32            channels=1, rate=self._rate,
33            input=True, frames_per_buffer=self._chunk,
34            # Run the audio stream asynchronously to fill the buffer object.
35            # This is necessary so that the input device's buffer doesn't
36            # overflow while the calling thread makes network requests, etc.
37            stream_callback=self._fill_buffer,
38        )
39
40        self.closed = False
41
42        return self
43
44    def __exit__(self, type, value, traceback):
45        self._audio_stream.stop_stream()
46        self._audio_stream.close()
47        self.closed = True
48        # Signal the generator to terminate so that the client's
49        # streaming_recognize method will not block the process termination.
50        self._buff.put(None)
51        self._audio_interface.terminate()
52
53    def _fill_buffer(self, in_data, frame_count, time_info, status_flags):
54        """Continuously collect data from the audio stream, into the buffer."""
55        self._buff.put(in_data)
56        return None, pyaudio.paContinue
57
58    def generator(self):
59        while not self.closed:
60            # Use a blocking get() to ensure there's at least one chunk of
61            # data, and stop iteration if the chunk is None, indicating the
62            # end of the audio stream.
63            chunk = self._buff.get()
64            if chunk is None:
65                return
66            data = [chunk]
67
68            # Now consume whatever other data's still buffered.
69            while True:
70                try:
71                    chunk = self._buff.get(block=False)
72                    if chunk is None:
73                        return
74                    data.append(chunk)
75                except queue.Empty:
76                    break
77
78            yield b''.join(data)
79
80def listen_print_loop(responses):
81    """Iterates through server responses and prints them.
82
83    The responses passed is a generator that will block until a response
84    is provided by the server.
85
86    Each response may contain multiple results, and each result may contain
87    multiple alternatives; for details, see https://goo.gl/tjCPAU.  Here we
88    print only the transcription for the top alternative of the top result.
89
90    In this case, responses are provided for interim results as well. If the
91    response is an interim one, print a line feed at the end of it, to allow
92    the next result to overwrite it, until the response is a final one. For the
93    final one, print a newline to preserve the finalized transcription.
94    """
95    num_chars_printed = 0
96    for response in responses:
97        if not response.results:
98            continue
99
100        # The `results` list is consecutive. For streaming, we only care about
101        # the first result being considered, since once it's `is_final`, it
102        # moves on to considering the next utterance.
103        result = response.results[0]
104        if not result.alternatives:
105            continue
106
107        # Display the transcription of the top alternative.
108        transcript = result.alternatives[0].transcript
109
110        # Display interim results, but with a carriage return at the end of the
111        # line, so subsequent lines will overwrite them.
112        #
113        # If the previous result was longer than this one, we need to print
114        # some extra spaces to overwrite the previous result
115        overwrite_chars = ' ' * (num_chars_printed - len(transcript))
116
117        if not result.is_final:
118            sys.stdout.write(transcript + overwrite_chars + '\r')
119            sys.stdout.flush()
120
121            num_chars_printed = len(transcript)
122
123        else:
124            print(transcript + overwrite_chars)
125
126            # Exit recognition if any of the transcribed phrases could be
127            # one of our keywords.
128            if re.search(r'\b(exit|quit)\b', transcript, re.I):
129                print('Exiting..')
130                break
131
132            num_chars_printed = 0
133
134def main():
135    # See http://g.co/cloud/speech/docs/languages
136    # for a list of supported languages.
137    language_code = 'en-US'  # a BCP-47 language tag
138
139    client = speech.SpeechClient()
140    config = types.RecognitionConfig(
141        encoding=enums.RecognitionConfig.AudioEncoding.LINEAR16,
142        sample_rate_hertz=RATE,
143        language_code=language_code)
144    streaming_config = types.StreamingRecognitionConfig(
145        config=config,
146        interim_results=True)
147
148    with MicrophoneStream(RATE, CHUNK) as stream:
149        audio_generator = stream.generator()
150        requests = (types.StreamingRecognizeRequest(audio_content=content)
151                    for content in audio_generator)
152
153        responses = client.streaming_recognize(streaming_config, requests)
154
155        # Now, put the transcription responses to use.
156        listen_print_loop(responses)
157
158if __name__ == '__main__':
159    main()
160
161

試したこと

pip install google-cloud-speech

補足情報（FW/ツールのバージョンなど）

公式のpythonコードが書いてあるページ
https://cloud.google.com/speech-to-text/docs/streaming-recognize?hl=ja#speech-streaming-mic-recognize-python
import errorで解決した？とされるサイト
https://qiita.com/hanlio/items/875b91e0d4931a57e86b

参考にしていたサイト
https://qiita.com/hamham/items/3733ac8cd9e3d7b9ccae

回答1件

ベストアンサー

私も初心者なのですが、インストールするライブラリを
grpc-google-cloud-speech-v1
にしてサンプルプログラムのimport部分を
from google.cloud import speech_v1 as speech from google.cloud.speech_v1 import enums from google.cloud.speech_v1 import types
に変更したところ、動きました。

参考にしているサイト
https://qiita.com/hamham/items/3733ac8cd9e3d7b9ccae
に詳しく書いてあったと思います。

投稿2019/12/04 08:49