[angelphanth]

This pull request updates the NodeMapper tutorial, core implementation, and tests to improve usability and flexibility when finding similar text nodes. The main changes include enhancing the get_similar method to support returning the top-k matches, updating tutorial data and outputs for clarity, switching from print to logging for model loading, and improving test coverage for the new features.

Enhancements to NodeMapper functionality:

• Added a top_k parameter to the get_similar method in hugger/mapper.py, allowing users to specify the number of top similar items to return. This includes input validation and updated docstrings/examples. [1] [2] [3] [4] [5]
• Changed the default similarity threshold for get_similar and get_match to 0, making it easier to retrieve results without filtering by a minimum score. [1] [2] [3]

Usability and output improvements:

• Replaced print statements with logging for model and tokenizer loading in hugger/mapper.py for better integration in various environments. [1] [2]

Tutorial and example updates:

• Expanded the example dataset in both nodemapper-tutorial.ipynb and nodemapper-tutorial.py to use longer, more realistic sentences and additional entries, improving the clarity and usefulness of the tutorial. [1] [2]
• Updated tutorial outputs and example queries to reflect the new dataset, demonstrate the new top_k parameter, and use more descriptive queries and thresholds. [1] [2] [3] [4]

Testing improvements:

• Refactored and expanded tests in tests/test_mapper.py to cover the new top_k functionality, renamed test functions for clarity, and adjusted assertions to match the updated logic and data. [1] [2]

Miscellaneous:

• Updated notebook metadata, kernel name, and Python version in the tutorial notebook for reproducibility and environment clarity. [1] [2]

These changes collectively make the NodeMapper more flexible and user-friendly, with improved documentation, examples, and test coverage.