All formats

caj2pdf

LanguagePython 3
LicenseGLWTPL ("Good Luck With That Public License", permissive but not OSI-approved)
Repositorygithub.com/caj2pdf/caj2pdf
DocumentationREADME and wiki (Chinese)
Depends onPyPDF2, imagesize, the mutool program from MuPDF; for HN files also libjbigdec and libjbig2codec (built with poppler or jbig2dec)
Read / writereads CAJ, KDH, PDF-in-CAJ, partially HN and C8 / writes PDF only

caj2pdf is the project that first reverse-engineered the CAJ format. It reads the header and table of contents, extracts the embedded PDF from CAJ files, de-obfuscates KDH files, repairs the result with mutool clean and adds bookmarks. For HN files it decodes page images and builds an image PDF, which requires compiling two helper libraries (prebuilt DLLs are included for Windows).

The project describes itself as working on some files only: unknown variants exit with Unknown file type., and TEB files aren’t supported. It’s not on PyPI; you run it from a checkout. It pins the old PyPDF2 2.2.0. Development has slowed; for new projects, caj2pdf-core (Rust) converts more variants.

Installation#

Shell
git clone https://github.com/caj2pdf/caj2pdf.git
cd caj2pdf
pip install -r requirements.txt     # PyPDF2==2.2.0, imagesize

# mutool (MuPDF) must be on PATH
sudo apt install mupdf-tools         # Debian / Ubuntu
brew install mupdf-tools             # macOS

Basic usage: convert CAJ to PDF#

From the command line:

Shell
python caj2pdf show paper.caj                  # type, page count, number of bookmarks
python caj2pdf convert paper.caj -o paper.pdf

From Python, import CAJParser from the checkout (run your script from the caj2pdf folder or add it to sys.path):

Python
import sys
sys.path.insert(0, "/path/to/caj2pdf")

from cajparser import CAJParser

caj = CAJParser("paper.caj")
print("format:", caj.format)            # "CAJ", "HN", "C8", "KDH", "PDF" or "TEB"
if caj.format in ("CAJ", "HN", "C8"):
    print("pages:", caj.page_num)
caj.convert("paper.pdf")

CAJParser raises SystemExit("Unknown file type.") for files it doesn’t recognize; catch SystemExit when you process many files.

How to detect the CAJ variant#

The variant decides whether conversion will keep text. Check it before converting:

Python
from cajparser import CAJParser

try:
    caj = CAJParser("paper.caj")
except SystemExit as e:
    print("not supported:", e)
else:
    if caj.format in ("CAJ", "KDH", "PDF"):
        print("contains a PDF: text and bookmarks are kept")
    elif caj.format in ("HN", "C8"):
        print("page images: needs the jbig libraries, output is image-only")
    else:
        print("cannot convert:", caj.format)

How to read the table of contents#

get_toc() returns a list of dicts with the title (UTF-8 bytes), target page and level:

Python
from cajparser import CAJParser

caj = CAJParser("thesis.caj")
for entry in caj.get_toc():
    indent = "  " * (entry["level"] - 1)
    print(f'{indent}{entry["title"].decode("utf-8")}  p.{entry["page"]}')

How to add bookmarks to a PDF printed from CAJViewer#

When conversion fails, print the CAJ to PDF in CAJViewer, then copy the bookmarks from the CAJ file into that PDF:

Shell
python caj2pdf outlines thesis.caj -o printed.pdf

Tips#