Enhance bertweet and sentiment_data

[dino65-dev]

I'm interested in GSoC and would like to work on this as a pre-GSoC contribution. As a student passionate about open source development, I'm eager to demonstrate my skills and get familiar with the project's workflow before the official GSoC period begins.

Changes want to make into :

Bertweet_model.py

1. Error Handling: Added comprehensive error handling during model initialization and inference.

2. Documentation: Expanded docstrings with detailed information on parameters, return values, and exceptions.

3. Type Hints: Added comprehensive type annotations following PEP-484 for better IDE support.

4. Caching Mechanism: Implemented lru_cache for tokenization to improve performance for repeated texts.

5. Batch Processing: Added a dedicated batch_process method to handle multiple texts efficiently.

6. Evaluation Capability: Added an evaluate method to assess model performance against ground truth.

7. Logging System: Replaced print statements with proper logging for better debug information.

8. Model Persistence: Added methods to save and load models for reuse.

9. Progress Tracking: Integrated tqdm for progress visualization during batch processing.

10. Improved Initialization: Better organization of initialization code and class structure.

11. Device Management: Automatic device selection (CUDA if available).

12. Graceful Failure Handling: The model now returns default values instead of crashing on errors.

13. Expanded Testing Code: More comprehensive examples in the __main__ section.

14. Class/Module Organization: Better separation of concerns with helper methods.

Sentiment_data.py

1. Improved Error Handling: Added comprehensive exception handling and validation of inputs.

2. Logging System: Replaced print statements with proper logging for better monitoring and debugging.

3. Type Annotations: Added comprehensive type hints for better code editor support and documentation.

4. Result Caching: Added lru_cache to improve performance for repeated analysis of the same text.

5. Batch Processing: Enhanced batch processing capabilities with progress tracking.

6. More Detailed Results: Added options to include probabilities for all sentiment classes in results.

7. Empty Input Handling: Now properly handles empty text inputs.

8. Improved Documentation: Added comprehensive docstrings for all methods.

9. Model Information: Added method to retrieve information about the loaded model.

10. Cache Management: Added methods to clear and manage the sentiment analysis cache.

11. Processing Time Tracking: Added timing information to see how long analysis took.

12. Sample Analysis: Added utility method to quickly verify model functionality.

13. Expanded Test Code: The __main__ section now includes more comprehensive examples.

14. Pretty Printing: Added better formatting for demo output.

15. Error State Results: Ensures results always include label and confidence, even in error cases.