mobi (Python)

mobi is a fork of KindleUnpack, the standard tool for taking Kindle files apart, without its GUI and packaged for PyPI. Its one public function, mobi.extract(), unpacks a book into a temporary folder and returns the path to the most useful result:

• an EPUB rebuilt from the KF8 part, for AZW3 and hybrid .mobi files;
• an HTML file (book.html) with images, for MOBI 6 files;
• a PDF for Print Replica (.azw4) books.

It decodes PalmDOC and HUFF/CDIC text and reconstructs the KF8 source with its CSS, fonts and table of contents. It doesn’t decrypt DRM books. The GPL-3.0 license applies to programs that you distribute with it.

Installation#

pip install mobi

The package also installs the mobiunpack command, which is the original KindleUnpack command line.

Basic usage: convert MOBI or AZW3 to EPUB#

import shutil
import mobi

tempdir, filepath = mobi.extract("book.azw3")
try:
    print("unpacked to", filepath)  # .epub, .html or .pdf
    if filepath.endswith(".epub"):
        shutil.copy(filepath, "book.epub")
finally:
    shutil.rmtree(tempdir)  # extract() leaves the folder for you to delete

How to extract text from a MOBI file#

Unpack, then read the HTML or EPUB that extract returns:

import shutil
import zipfile
import mobi
from bs4 import BeautifulSoup  # pip install beautifulsoup4

def mobi_text(path):
    tempdir, filepath = mobi.extract(path)
    try:
        if filepath.endswith(".html"):  # MOBI 6
            with open(filepath, "rb") as f:
                return BeautifulSoup(f.read(), "html.parser").get_text("\n", strip=True)
        if filepath.endswith(".epub"):  # KF8
            parts = []
            with zipfile.ZipFile(filepath) as z:
                for name in sorted(z.namelist()):
                    if name.endswith((".xhtml", ".html")):
                        parts.append(BeautifulSoup(z.read(name), "html.parser").get_text("\n", strip=True))
            return "\n\n".join(parts)
        return ""  # PDF from a Print Replica book
    finally:
        shutil.rmtree(tempdir)

print(mobi_text("book.mobi")[:500])

For the exact chapter order of the EPUB, read it with EbookLib and follow the spine.

How to unpack a book from a file object#

extract also accepts an open binary file (it copies it to the temporary folder first):

with open("book.mobi", "rb") as f:
    tempdir, filepath = mobi.extract(f)

How to get all unpacked parts (images, OPF, both MOBI 6 and KF8)#

Call KindleUnpack directly to choose the output folder and the EPUB version:

from mobi.kindleunpack import unpackBook

unpackBook("book.mobi", "out", epubver="3")  # epubver: "2", "3", "A" (automatic) or "F"
# out/mobi7/ : MOBI 6 HTML and images
# out/mobi8/ : KF8 source and the rebuilt EPUB

The same from the command line: mobiunpack --epub_version=3 book.mobi out.

Tips#

• extract raises an exception for DRM-protected or damaged books, and ValueError when it finds no EPUB, HTML or PDF output; wrap it in try when processing many files.
• Each call writes the whole unpacked book to disk. If extract raises, its temporary folder (named mobiex... in the system temp directory) stays behind; clean those up in batch jobs.