Version 0.1.9

[AnFreTh]

This pull request introduces significant updates to the stream_topic project, including the addition of a new neural topic model, updates to documentation, and various codebase improvements. The most important changes are summarized below.

New Features:

• Added NSTM Model: Introduced the Neural Topic Model via Optimal Transport (NSTM) with its base class NSTMBase in stream_topic/models/neural_base_models/nstm_base.py and included it in the __init__.py and models.rst files. [1] [2] [3] [4]

Documentation Updates:

• README Enhancements: Updated README.md to include a new GIF for the header, removed an old table of GIFs, and added a reference to the NSTM model. [1] [2] [3]
• API Documentation: Added NSTM to the API documentation in docs/api/models/models.rst.
• Notebook Kernel Updates: Updated the kernel specifications in docs/notebooks/examples.ipynb and docs/notebooks/quickstart.ipynb to use the stream_topic_venv environment. [1] [2]

Codebase Improvements:

• Removed octis Dependency: Eliminated the octis library from the dependencies in docs/conf.py and replaced its usage with custom dataset handling in topic_extraction.py. [1] [2] [3] [4] [5] [6]
• Updated Preprocessing Steps: Added preprocessing configurations for the NSTM model in default_preprocessing_steps.json.

Version Update:

• Version Bump: Incremented the version from 0.1.8 to 0.1.9 in stream_topic/__version__.py.

Miscellaneous:

• Parameter Change in CEDC Model: Changed the default expansion_corpus parameter from "octis" to "brown" in the CEDC model and its documentation. [1] [2]

These changes collectively enhance the functionality, documentation, and maintainability of the stream_topic project.