Ensure:
 - no words are correctly not matched
 - some words are matched

Calls pulled out relate to setup and working of Whisper:
 - _whispermodel()
 - _batched_inference_pipeline()
 - _transcribe()

Defaults defined: model, device, compute type, beamsize, batchsize, pipeline type

Tests:
 - basic init
 - init with no media
 - run() with no words (early exit 0 Features)
 - run() with mocked transcribe

NOTE: these are unit tests and do not exercise Whisper

Uses unittest.mock.path to patch out call to matplotlib.pyplot.savefig()

TODO: better, more-granular tests of the bits of functionality here

mocks out calls to ffmpeg for obvious reasons

Includes:
 - MockInterval
 - MockFeature
 - MockSource

Gets utils coverage up to 100% (for now)

@see: 0a08af4

Covers:
 - to_json
 - move_end
 - update_duration
 - unhappy path exceptions (mostly init)

Gets test coverage for this module to 100%! (for now)

TestLoudAudioFeatureExtractor -- init only

TestSourceMedia is a mockinobject so

TODO: rename

Initialset of tests for LaughterFeatureExtractor and
RandomFeatureExtractor

This covers:
 - Source
 - SourceMedia
 - Interval