This project explores various tools for converting pdf to html/css. The evaluation is based on the following perspectives:

• Pixel perfect;
• Accessibility;
• Link integration;
• Foot print;

How to run

# Clean up the workspace
just clean

# Install
brew install poppler
yarn

# Run the experiments with node-poppler
just run-node-poppler

node-poppler

https://www.npmjs.com/package/node-poppler

Learnings

• This library depends on external binaries;
• It can export pdf to html
  • With option complexOutput, it can preserve the original layout with css 👍;
  • Fonts are not extracted, so the layout will look different than the original pdf 👎;
  • It can extract links. But it will render the links with some default styling and the page will look different than the original pdf 👍👎;
  • It provides an option to "use data URLs instead of external images in HTML", but enabling this option will end up crashing 👎;
• It can export pdf to svg
  • The output looks pixel perfect 👍;
  • Each font symbol is presented with glyph tag (but it tells nothing about the actual character) 👎;
  • The texts are not accessible to screen readers 👎;
  • No link integration 👎;
  • Footprint is huge (246KB vs 1.8MB) 👎;

pdf2htmlEX

https://github.com/pdf2htmlEX/pdf2htmlEX

Learnings