Added timeseries normalization methods based on User computed statistics

[mikeyautorino2]

Key changes:

1. Added normalization functionality with three methods:

• z-score: (value - mean) / std
• min-max: (value - min) / (max - min)
• robust: (value - median) / MAD
• Added support for two normalization axes:
• global: statistics computed across all values
• per_feature: statistics computed for each feature separately
• Implemented fit() to calculate normalization statistics from provided samples.
• Integrated _apply_normalization() into process() when normalize=True.
• Added docstrings for clarity and maintainability.

Why:
These changes make TimeseriesProcessor capable of producing standardized outputs, improving downstream model performance and ensuring consistency when working with time series data from varying sources.
Testing:
Verified fit() correctly computes normalization statistics for both global and per-feature cases.
Wrote unit tests in test_normalization.py and test_advanced_normalization.py for edge cases