Hacker News Reader

Comment by tough | Hacker News Reader

I had edited prior message but would be great if supported PDF rendering as good as docx! (for research mostly, not a lawyer)

View on HN

One thing it does is cast the PDF to grayscale to render for speed (lawyers rarely care about the color of a legal doc in PDF), so perhaps I'll make that an adjustable setting which will trade off speed for clarity. Otherwise, it uses PDFium for PDF rendering so it should come out just as well as what you see in a Chromium-based browser.

might be an issue with DPI / Retina screens on macOS?

it's really more blurry than any pdf on chrome https://imgur.com/a/AElOuaA

Oh, yes, that is also a cheat code to render them quicker -- it downsamples the render to save time. That will improve in future versions as Tritium uses spare cycles to increase the resolution.

I also noticed when i opened a folder with hundreds of big pdf's it choked on it

seems you're doing work on all files when loading, you should be more like lazy-loading?

^ yes it does greedily ingest the PDFs to allow for instantaneous search. What it needs (and doesn't yet do) is at least multi-thread that parsing step to get a 2-4x speedup.

If this is Electron, try pdf.js - really good rendering, you can create a text layer (for text selection, etc). Probably the best result per effort you can get

care to link to repo? alwayss good to have one of these at hand

https://github.com/mozilla/pdf.js#readme (Apache 2)

Reply on news.ycombinator.com