feat(pipeline): add official client for pyannoteAI web API

[hbredin]

This is a work in progress that allows pyannoteAI users to send speaker diarization requests from pyannote.audio usual API.

from pyannote.audio import Pipeline
pipeline = Pipeline.from_pretrained('hbredin/test-api", token="{pyannoteAI-api-token}")
diarization = pipeline("/path/to/audio.wav")

The above code will:

• upload the audio file to pyannoteAI web API
• initiate a diarization job
• retrieve the job output once completed
• deserialize the output into a good old pyannote.core.Annotation instance

[hbredin]

This is now ready for review but expects this WIP package to be installed.