pdf2djvu creates DjVu files from PDF files.
It's able to extract:
- text layer,
- document outline (bookmarks),
- metadata (including XMP metadata).
pdf2djvu is developed primarily for Unix-like systems, but a Windows distribution is available too.
The following software is required to build pdf2djvu:
- C++11 compiler;
- GNU make;
- DjVuLibre (≥ 3.5.25), both the library and the command line tools;
- Poppler (≥ 0.35), including xpdf headers;
The following software is recommended:
- gettext for internationalization;
- Exiv2 and libuuid (part of util-linux or e2fsprogs) for correctly dealing with XMP metadata;
- GraphicsMagick for the
--jobs option, the compiler must support OpenMP.
To run the tests, the following software is needed:
- Python 2.7;
To correctly convert some PDF files (mostly in Chinese, Japanese or Korean), the poppler-data package must be installed.