No wit.init() about pywit

[rmeertens]

Cool. Any chance of getting an offline Wit interpreter in the future? (downloading your trained intents, and saving everything for upload later)

[rmeertens]

Could it be that the version on pip is NOT the official wit-ai library?
I was able to build it from source, although it took more steps than I liked (had to manually download and rename the lib file and move it to the right folder).
Maybe it would be a good idea to create an install script or create a copy-paste commands procedure on the website specifically for ubuntu etc.

[blandinw]

Hey Roland, it looks like the version on PyPi is the same as on the repo (https://pypi.python.org/pypi/wit/1.1).
I'm not knowledgeable in Python packaging, especially modules written in C.
We'll gladly accept a pull request if you find the time to look at it.

[rmeertens]

Are you sure? What do you see if you run the following in python:
import wit
dir(wit)

For me the whole init function is missing (also missing in the unofficial library, hence my fast conclusion)

[andreimarcu]

Hi @rmeertens, is the file you're working on named wit.py by any chance?
I just installed pywit from pip and I haven't had that issue.

[walchko]

Make sure you installed wit using:

pip install wit

and not

pip install pywit

The confusing part is the official directions say Let’s install the pywit package! Which is the wrong package and doesn't have a wit.init() ... it confused me until I noticed the command line to the right of this sentence said pip install wit and not pywit.

Since pywit is still in pip, I strongly suggest you remove reference to it in the documentation so people don't get confused (like me).

[blandinw]

We moved away from audio recording and streaming because:

• many people only needed access to the HTTP API, and audio recording did not make sense for server-side use cases
• dependent on platform, choice best left to developers
• forced us to maintain native bindings as opposed to a pure Pythonic library

[rmeertens]

So... how to turn speech into actionable data?

[blandinw]

We now expect the developer to pick the best audio recording library for their platform and stream it to POST /speech (https://wit.ai/docs/http/20141022#get-intent-via-speech-link).

[Sultan91]

Could update info about speech recognition, is post audio file function is working?