feat: Add synthetic time-series data generator

This commit introduces a new generator for creating synthetic longitudinal (time-series) data, which is essential for testing the time-series capabilities of the ml-grid pipeline.

Key changes include:

SyntheticTSDataGenerator class: A new class dedicated to generating long-format time-series data. It creates a DataFrame with client_idcode and timestamp columns, where each patient has a specified number of timepoints.
Realistic Data Structure: The generator mirrors the structure of real-world longitudinal data, making it suitable for end-to-end pipeline testing. It reuses feature naming conventions from the existing SyntheticDataGenerator for consistency.
generate_synthetic_ts_data function: A convenience wrapper function is added for easy instantiation and use of the new generator.
Example Usage: The if __name__ == "__main__" block is updated to include a comprehensive example of generating, imputing, and saving a synthetic time-series dataset, alongside the existing tabular data example. This demonstrates its usage and integration with existing utility functions like mean_impute_dataframe.

Author: SamoraHunter