Data Processing Instructions

1. Copy subset_1_filtered_updated_final_output.csv into the local folder

   • This is the raw output from the complete PDF extraction pipeline

2. Process the raw OCR output using process_raw_output.ipynb

   • Demonstrates REGEX application via extract_meaningful_text function
   • Pipeline:
     1. Filter pipeline errors from dirty web-scraped PDFs
     2. Apply REGEX to produce continuous training text
     3. Modify extract_meaningful_text function as needed for different outputs