It's definitely far easier to emit a controlled, useful subset of PDF than it is to parse PDF documents. I wrote a small PDF library for the Decker ecosystem that just focuses on bitmaps and page layout; roughly 4kb and 135 LoC.
I’m working on one rn. It takes arbitrary PDFs and builds composable dynamic pandoc pipelines to match the source byte for byte output. It’s very very complex. But if I can get it finished it will fuck over Adobe so worth it.
While not quite as small as 3kb, I recently found this incredible library called html-to-image that's only 300kb. It clones whatever subtree of your document you want to a <foreignObject> inside an svg which then allows it to output canvas, png, svg, pdf, blob, jpeg, etc. Even more impressively is that it handles custom fonts, pseudo-elements, computed styles and more.
Only supports ASCII characters, which is part of the trick here. As soon as you need more Unicode (even just typographic quote characters and such), you’ll need significantly more logic. Also no bold, italics, etc.
Back in the day I needed PDF export for some client thing. I can't remember if I was using pdfjs or jspdf. I do however remember that it was many thousands of lines of code, and yet, I had to lay out the lines of text on the page manually.
My page layout code was like 50 lines of code. And I remember thinking... OK they already wrote 8,000 lines of code... They couldn't have added 50 more?!
400 lines though. Respect. I will take a proper look at this when I recover from burnout :)
Great exercize, but for most use cases - people will continue reaching for jsPDF.
I think if you have a markdown->PDF function included, where I can send in markdown and get PDF, that would solve quite many needs, and would be useful.
Was Typst falling short in any particular area that made you not want to use it? (If it was on your radar at all). I think it would work for your use case and could also run client side if needed.
Nice work! I'm curious though, what was your use case for needing a smaller library? Since you're running this on a server, what difference does an extra 226KB make?
I still have a tiny DOS binary (x86 Asm) that I wrote decades ago for turning plaintext ASCII files into PDFs, for those annoying use-cases where the former isn't accepted but the latter is. It's only a few hundred bytes, with the majority being data to be copied verbatim into the output file.
docs/demos: https://beyondloom.com/decker/pdf.html
browsable source: https://github.com/JohnEarnest/Decker/blob/main/examples/dec...
https://github.com/bubkoo/html-to-image
It's probably the most impressive and seamless experience I've had with converting HTML to pdfs/images so I just wanted to sing its praises here
My page layout code was like 50 lines of code. And I remember thinking... OK they already wrote 8,000 lines of code... They couldn't have added 50 more?!
400 lines though. Respect. I will take a proper look at this when I recover from burnout :)
I think if you have a markdown->PDF function included, where I can send in markdown and get PDF, that would solve quite many needs, and would be useful.
https://github.com/Lulzx/tinypdf/commit/961e6b602f19e125f210...
Here's the TS library: https://github.com/Myriad-Dreamin/typst.ts
https://developers.cloudflare.com/workers/platform/limits/
That said, thank you for fighting library bloat! Small libraries tend to perform better (not just on Workers but in general).