Adds Parameter use_enhanced and model to GoogleCloudSpeech (Fix #734)

[HideyoshiNakazone]

Adds the parameters use_enhanced and model to the recognize_google_cloud method for more customizable options for the user and better results in specific cases

[HideyoshiNakazone]

Hello @Uberi and @ftnext, i was wondering if it's possible for someone to review my merge request.

Thank you very much,
Vitor Hideyoshi.

[HideyoshiNakazone]

Hello @ftnext, is there any interest in this feature? It doesn't break any of GoogleCloudSpeech python api, only extends it. I'm currently already using this implementation in the company i work in, but would love to have this feature merged.
If there is anything blocking the merge please tell me :)

[Uberi]

Hi @HideyoshiNakazone!

Looks good overall, but would it be possible to document these parameters in the docs for that function? If so, happy to merge this!

[HideyoshiNakazone]

@Uberi, thanks a lot! I added the parameters to the Docstring of the method Recognizer.recognize_google_cloud and added them to the library reference file.
If there is any other places you'd like me to add documentation i'll be happy to :)

[ftnext]

@HideyoshiNakazone Thank you very much for this pull request! I'm very sorry to respond too late.
@Uberi Thanks your comment!

In my opinion, it seems to be better to introduce keyword arguments (a.k.a. **kwargs)
https://docs.python.org/3/tutorial/controlflow.html#keyword-arguments

Certainly, adding use_enhanced and model as arguments would implement this feature.
However, if there are additional arguments to be added in the future, there is a concern that they could be added again (not easy to extend).

I think it would be preferable for Cloud Speech API-specific arguments to be specified as variant keyword arguments.

def recognize_google_cloud(self, audio_data, credentials_json=None, language="en-US", preferred_phrases=None, show_all=False, **api_params):
    """
    If ``preferred_phrases`` is an iterable of phrase strings, ...

    api_params: Cloud Speech API-specific parameters as dict (optional)

        The ``use_enhanced`` is a boolean option ...

        Furthermore, you can use the option ``model`` to set your desired model,

    Returns the most likely transcription if ``show_all`` is False (the default).
    """

    config = {
        'encoding': speech.RecognitionConfig.AudioEncoding.FLAC,
        'sample_rate_hertz': audio_data.sample_rate,
        'language_code': language,
        **api_params,
    }

(It seems that preferred_phrases might be included in api_params too, but this is another issue)

[HideyoshiNakazone]

Hi @ftnext and @Uberi ,

I know it has been a while, but could you please take another look at my pull request? I believe it now meets the requirements discussed in this thread, and it would be great if we could proceed with merging it.

Thanks for your time!

[ftnext]

Thanks!
May I make a simple correction?

[ftnext]

Thanks!