From Elvanör's Technical Wiki
- PDF is a completely different technology than for example HTML. Thus to produce quality PDF documents, use LaTeX or other applications. HTML is not well suited for PDF creation.
Tools for working with PDF files
- app-text/pdftk (PDF Toolkit) is great for extraction, concatenation and other manipulations. For instance to concatenate two PDFs:
pdftk in1.pdf in2.pdf cat output out1.pdf
To extract a single page:
pdftk A=sample.pdf cat A18 output extract.pdf
To extract several pages:
pdftk A=sample.pdf cat A3-8 output extract.pdf
- app-text/mbtpdfasm allows you to extract parts of a PDF document:
mbtPdfAsm -morigin-file.pdf -p"7;8;9;10" -ddestination-file.pdf
However there is not any stable version available on the main Portage tree.
- pdfimages used with the -i option allows to extract images from a PDF file. You can thus check if sizes seem reasonable. pdfimages is a program part of the poppler packages.
- You can open a (single page) PDF with Gimp and convert it easily to a PNG. To convert several images (PNGs, JPEGs) to PDF use imagemagick:
convert pages-*.jpg output.pdf
Converting from (X)HTML + CSS
- Converting an HTML document to a PDF one is not easy. HTML should not be chosen as a view technology if you need to output to a PDF or print your document. However, various libraries still exist to convert an HTML document to PDF.
- Be sure to choose a large font size in your HTML document. I set a base font-size on the body equal to 20px. Use relative font-sizes in your XHTML document (eg, font-size: 0.85em;).
- This PHP library (coming from a Perl port) is quite good, although it seems development has stopped. It can convert remarkably well XHTML documents with CSS properties. However, it has the following caveats:
- Make sure your CSS is perfectly valid. Sometimes you will need to specify a clear property explicitely. With a bit of effort, your output document will be very close to the original Firefox rendering.
- You can use either the FPDF or PDFLIB libraries to output to PDF. I tried both but did not see any noticeable differences (although the pdflib package must be installed separately in Gentoo).
- Home page for this project. Several other projects exist in PHP but I don't know if they handle CSS 2.1 correctly.
- html2pdf does not yet support (PNG) transparency. You should also use high quality (very large) images in your HTML document, giving its actual size via the CSS properties.
- CSS properties with !important set do not seem to work.
- Position: absolute does not work well (top: 0px and bottom: 0px not supported).
- In html2ps.config, you can define additional media (and then use this media type via the media parameter). This allows to define custom height and width for the generated PDF.
- Pisa is a Python library. At this stage, it seems to be lacking a lot of CSS 2.1 properties. To keep an eye on though.
- Flying Saucer is an open-source Java XHTML + CSS to PDF renderer. It seems to be very good already (supports transparent PNG images), and development is active. This is probably one of the best free solution out there, and its progress should be watched regularly.
- However it currently (R8) has problems with CSS properties (top: 0px and bottom: 0px not supported).
- The best converter available today. Really impressive: did not find a single problem so far. But it's not open-source.