quick and dirty script to repair epub that is not properly tagged as EPUB mimetype

repack_epub.py

# Quick script to repair broken epub
# Need libmagic and python-magic to work
# Use case:

import argparse
import sys
from pathlib import Path
from typing import List
from zipfile import ZIP_DEFLATED, ZIP_STORED, ZipFile

import magic

ZIP_MAGIC = b"PK\x03\x04"

parser = argparse.ArgumentParser(description="Repair broken epubs")
parser.add_argument("path", type=str, help="Path to directory")
parser.add_argument("-r", "--recursive", action="store_true", help="Recursive")


args = parser.parse_args()

path = Path(args.path)
recursive = bool(args.recursive)

collect_all_epubs: List[Path] = []
if recursive:
    for p in path.rglob("*.epub"):
        collect_all_epubs.append(p)
else:
    for p in path.glob("*.epub"):
        collect_all_epubs.append(p)


to_be_fixed_epubs: List[Path] = []
for epub in collect_all_epubs:
    with epub.open("rb") as fp:
        read_meta = fp.read(128)
    if not read_meta.startswith(ZIP_MAGIC):
        continue
    mimetypes = magic.from_buffer(read_meta, mime=True)
    if not mimetypes:
        continue
    if "epub+zip" not in mimetypes:
        print(f"Found broken epub: {epub}")
        to_be_fixed_epubs.append(epub)

print("===============================================")
print(f"Found {len(to_be_fixed_epubs)} broken epubs")
continue_it = input("Continue? [y/n]: ").lower()
if continue_it != "y":
    sys.exit(0)

for epub in to_be_fixed_epubs:
    print(f"Fixing {epub}")
    # We fix it by resaving with proper mimetype
    save_target = epub.parent / f"{epub.stem}.temp.epub"
    new_epub = ZipFile(save_target, "w", compression=ZIP_DEFLATED)
    new_epub.writestr("mimetype", "application/epub+zip", compress_type=ZIP_STORED)

    with ZipFile(epub, "r") as original_epub:
        # Iterate through all files
        for path in original_epub.infolist():
            if "mimetype" in path.filename:
                continue
            new_epub.writestr(path, original_epub.read(path))

    new_epub.close()
    epub.unlink(missing_ok=True)
    save_target.rename(epub)
    print(f" | Fixed {epub}")