Feat/implement mts dialog based clinical summary evaluation

[chakravarthik27]

trafficstars

This pull request introduces several changes to the langtest module, primarily focusing on enhancing functionality and improving code structure. The most important changes include the addition of dialogue-related columns, the introduction of a new SummaryEval class for evaluating clinical summaries, and the correction of a recurring typo in class names.

Enhancements to functionality:

• Added dialogue column to various DataFrame manipulations in langtest/langtest.py to support new data types. [1] [2]
• Introduced SummaryEval class in langtest/metrics/llm_eval.py for evaluating clinical summaries generated from doctor-patient dialogues.

Code structure improvements:

• Corrected the typo "BaseClincial" to "BaseClinical" throughout langtest/transform/clinical.py to ensure consistency and clarity. [1] [2]
• Added new imports and type hints to improve code readability and type safety in langtest/modelhandler/llm_modelhandler.py and langtest/modelhandler/modelhandler.py. [1] [2]
• Updated the tqdm import in langtest/transform/base.py to use tqdm.auto for better compatibility.