search

wit-ai / pywit

Python library for Wit.ai
Other
1.45k stars 358 forks source link

Add more api functions #2

Open Seraf opened 9 years ago

Seraf commented 9 years ago

Hello,

I was using pywit (https://github.com/lextoumbourou/PyWit) in my project and as you made an official python package, I was going to use it. But I miss some features to switch :

There was some useful functions like get_intents(). Is it possible to implement these please ?

Also, I think dependencies shouldn't be here in this library. Using the audio part is one of the feature of your API, but seems a bit useless to install dev dependencies to compile (as it isn't pure-python) the package by default. It should be used in a subpackage or something like that "wit.api.X, wit.text.X, wit.audio.X".

This way it will be more pythonic and more easy to distribute the package. (but it's my personal point of view)

patapizza commented 9 years ago

Hi Seraf,

Thank you for your feedback.

Can you provide me with a exhaustive list of functions you would like to see (including expected behavior)?

I see your point, yet for the sake of simplicity, we prefer to release one single Python package.

Seraf commented 9 years ago

Hi, thanks for your answer. Here is a good start of useful functions : http://pywit.readthedocs.org/en/latest/api.html At least get_intents seems to be (for me) the most useful function to know what the app can do.

Also, can you tell me if your lib can stream directly the microphone to wit ?

Thanks !

patapizza commented 9 years ago

Did you look at the methods voice_query_auto and voice_query_auto_async?

More info in the API Docs.

Seraf commented 9 years ago

Yes I saw these functions, but in the documentation it doesn't tell if it's streamed or not. I mean, does it record, then push the sound with a temporary file, or does it stream each chunk captured on the microphone ? (like https://github.com/lextoumbourou/PyWit/blob/master/examples/recorded_stream_input_example.py)

Also, does the other function seems ok for you ?

patapizza commented 9 years ago

Ok got it. Update: Just double-checked, the SDK does stream the microphone output straight to the API.

Yes, I agree that we should provide the SDK with CRUD operations on intents and entities. We will implement these in a next release.

Thank you again for the suggestion!

Seraf commented 9 years ago

Hello, any news about this @patapizza ? I would like to be able to copy intents from an account to another also.

Is there any reconsideration about splitting the voice part from the API part of the package ? It seems weird to depend on sox for my server who only need to text query and do CRUD operations.

Thanks

blandinw commented 8 years ago

@Seraf Unfortunately, we don't have the bandwidth to look at it at the moment. I totally agree that the dependency on sox should be optional. If you only need the text methods, feel free to use an HTTP client like requests directly or submit a pull request.

gulbinas commented 6 years ago

https://github.com/wit-ai/pywit/pull/101

333444555 commented 5 years ago

125 should fix a little of this, i am gonna be adding more soon.

sourabhvarshney111 commented 4 years ago

Hello everyone, Is there a method to access validate sample end point?