Recursive pipeline

[saif-ellafi]

Description

An alternative pipeline that provides annotators with knowledge about their last-state pipeline. This allows some annotator machine learning models re-utilize current pipelines among different datasets, avoiding to have to create inner-pipelines to just run a few NLP tasks within an annotator.

Recursive Pipelines are backwards compatible and are an optional alternative which should boost training performance in some scenarios. As of this writing, Recursive pipelines only help to the CRF NER approach training stage.

Motivation and Context

Avoid unnecessary waste of time in re-creating inner pipelines just to perform nlp tasks within an annotator. Facilitates the annotator with a tool to re-utilize already trained pipeline instead.

How Has This Been Tested?

Screenshots (if appropriate):

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.