code93/pdfstruc

Install dependencies

do apt install tesseract only when using ocr to do the same

use spacy download only when using nlp (pdfstruc.py script)

apt install tesseract-ocr -y
pip install -r requirements.txt
spacy download en_core_web_sm

Define Environment Variables to use the selected features only:

If all env var = false, it will only give pdf to images

export use_opencv=true
export use_nlp=true
export use_nlp_topic_modelling=true
export use_regex_for_research_paper=false
export use_regex_for_page_no=true
export use_for_ssrn=false
export use_process_again=false # begins the process anew if true

Execute Python Script for getting images named as page numbers correctly and text data of pdf

python3 pdf2jpg.py

To do the same using ocr

go to the directory containing images of pdf file and execute the script

python3 pdfocr.py

Execute Python Script for also getting summary of pdf's and using Topic Modelling to sort with relevency:

python3 pdfstruc.py

If there is any pdf file that has not been processed to jpg images in the directory where the .py script executes, the py script automatically creates a folder of the same name of the pdf file and then starts converting pdf pages to jpg images saved with appropriate page name.