A script to extract the alt text from images in an ePub

extractAltText.py

from ebooklib import epub
import io
from PIL import Image
from bs4 import BeautifulSoup

# Open the EPUB file
book = epub.read_epub('frugal.epub')

image_alts = {}
raw_images = {}

# Iterate over each item in the book, extracting the alt and the raw image
# files
for item in book.get_items():

    if item.get_type() == 9: # xhtml file
        soup = BeautifulSoup(item.get_content(), 'lxml')
        images = soup.find_all('img')
        for image in images:
            alt = image.get('alt', None)
            src = image.get('src', None)
            src = src[10:] # strip ../Images/ prefix
            image_alts[src] = alt

    if item.get_type() == 1: # image
        img = Image.open(io.BytesIO(item.get_content()))
        name = item.get_name()[7:] # strip Images/ prefix
        raw_images[name] = img

# For Frugal Only
# # Chapter Titles
# sorted_chapter_titles = {}
# for key in image_alts:
#     if key[:2] == 'ch':
#         number = int(key[2:][:-4])
#         sorted_chapter_titles[number] = image_alts[key]

# sorted_chapter_titles = dict(sorted(sorted_chapter_titles.items()))
# for k in sorted_chapter_titles:
#     print(str(k) + ' - ' + sorted_chapter_titles[k])
#     print()

# print('Epilogue - ' + image_alts['epi.jpg'])

# # FAQs
# print()
# print('FAQs')
# print()
# sorted_faqs = {}
# for key in image_alts:
#     if key[:3] == 'faq':
#         number = int(key[3:][:-4])
#         sorted_faqs[number] = image_alts[key]

# sorted_faqs = dict(sorted(sorted_faqs.items()))
# for k in sorted_faqs:
#     print(str(k) + ' - ' + sorted_faqs[k])
#     print()

# # Margins
# print()
# print('Margins')
# print()
# sorted_margins = {}
# for key in image_alts:
#     if key[:7] == 'margins':
#         number = int(key[7:][:-4])
#         sorted_margins[number] = image_alts[key]

# sorted_margins = dict(sorted(sorted_margins.items()))
# for k in sorted_margins:
#     print(str(k) + ' - ' + sorted_margins[k])
#     print()

for key in image_alts:
    print(key)
    print(image_alts[key])
    print()