pdf_extract_images.py

import os
from PIL import Image
import fitz # pip install PyMuPDF
import io

def extimg(location, pdffile):
    pf = fitz.open(os.path.join(location, pdffile))
    for pi in range(len(pf)):
        iml = pf[pi].getImageList()
        if not iml:
            continue
        for i in range(len(iml)):
            im = pf.extractImage(iml[i][0]) #image index of page, xml ref string is in [0]
            image_bytes = im["image"]
            # get the image extension
            image_ext = im["ext"]
            # load it to PIL
            image = Image.open(io.BytesIO(image_bytes))
            # save it to local disk
            fn = os.path.join('/tmp/images',location, pdffile)
            fn = f'{"".join(fn.split(".")[:-1])}/{pi+1}_{i}.{image_ext}'
            print(fn)
            os.makedirs(os.path.dirname(fn),exist_ok=True)
            with open(fn, "wb") as w:
                image.save(w)

def map_files(path = '.'):
    for root, dirs, files in os.walk(path):
        for name in files:
            if '.pdf' in name:
                extimg(root,name)
        for name in dirs:
           if name == 'venv' or name =='images':
               continue
           else:
               map_files(os.path.join(path,root,name))

map_files()