👽 Update MelusinePipeline for sklearn compatibility

[HugoPerrier]

Python version and environment updates:

• Updated the minimum required Python version to 3.10 in pyproject.toml and removed references to Python 3.8 and 3.9 from both pyproject.toml and tox.ini. Test environments for Python 3.10 and 3.12 are now specified. [1] [2] [3]

Pipeline compatibility and type simplification:

• Added a no-op fit method and a __sklearn_is_fitted__ property to MelusinePipeline in melusine/pipeline.py to ensure compatibility with scikit-learn >=1.2, which requires pipelines to be fitted before calling transform.
• Simplified type annotations throughout melusine/pipeline.py by removing explicit type hints from variable assignments, making the code more concise and compatible with Python 3.10+. [1] [2] [3] [4] [5] [6]
• Changed default arguments and return type annotations in pipeline configuration methods for better Python 3.10+ compatibility. [1] [2]

Regex and segmentation improvements:

• Improved email segmentation regex patterns in melusine/processors.py to better handle cases where meta keywords are followed by a newline instead of a colon, increasing robustness in email parsing. [1] [2]

Minor bugfixes:

• Fixed a minor typo in a regex pattern for matching addresses in melusine/processors.py ({,3} to {0,3} for clarity and correctness).