Fix: Resolve dictionary key overwrites and missing pandas import in data generation pipeline

[kamilansri]

Description

This PR addresses several critical bugs in the data_generation_pipeline script that prevented successful execution and caused silent logic failures.

Changes Made

• Added Missing Import: Added import pandas as pd at the top of the file to prevent a NameError during the final dataset CSV read.
• Fixed Dictionary Key Overwrites: Refactored the book_transformations dictionary. Previously, consecutive identical keys (e.g., 'denoise_image') were overwriting each other natively in Python. These sequential steps have been bundled into lists (e.g., 'denoise_image': [{'method': 'bilateral'}, {'method': 'nlm'}]) to preserve the pipeline's operational intent. (Note: Ensure process_multiple_books is equipped to parse these list values).
• Variable Renaming: Prevented df from being assigned and overwritten three separate times in sequence. Variables are now explicitly named (regions_df, dataset_df, final_df) to improve debugging capability.
• Passed Defined Variables: Replaced the hardcoded 0.8 in mapping_bounding_boxes with the initialized similarity_threshold variable.