Adding audio input support in LightPipeline

[danilojsl]

Description

This change adds the required methods in Scala and Python to allow sending audio inputs as a collection of floats to fullAnnotate method in LightPipeline.

Motivation and Context

Currently Wav2Vec2ForCTC is the only available as a Spark transformer, users that rely on LightPipline for their processes cannot use this new annotator. By enabling this usage in LightPipeline users will be able to use it.

How Has This Been Tested?

Screenshots (if appropriate):

• Unit tests
• Local tests
• Google Colab Notebook

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• Code improvements with no or little impact
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

Checklist:

• My code follows the code style of this project.
• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I have read the CONTRIBUTING page.
• I have added tests to cover my changes.
• All new and existing tests passed.

[maziyarpanahi]

@danilojsl Thanks for this feature. Something seems to be different in LightPipeline that resulted in this test failing:

• unit test:

  spark-nlp/src/test/scala/com/johnsnowlabs/nlp/annotators/RecursiveTestSpec.scala

  Line 185 in 2428136

  "Lazy Recursive Model" should "work well in LightPipeline" taggedAs FastTest in {

• error: https://github.com/JohnSnowLabs/spark-nlp/actions/runs/3192853478/jobs/5210797501#step:7:5198

[danilojsl]

@maziyarpanahi, yes I added the commit that solves that error