PDF text extraction using PDF.js

[RyotaUshio]

Hi, thank you so much for making this awesome plugin!

The original implementation uses the Text Extractor plugin for this purpose, but I found sometimes it didn't work as expected, especially for handwritten notes exported from GoodNotes.

These notes already have embedded text generated by the great OCR ability of GoodNotes, and this text is searchable by the built-in search of Obsidian's PDF viewer, as shown below.

However, Omnisearch can't find it.

So I modified the PDF text extraction part so that it simply uses PDF.js.
It can now locate the handwritten text as below.

What do you think of this approach?

[scambier]

First, thanks for your PR :)

Omnisearch itself doesn't manage text extraction. This work has been offloaded to Text Extractor (that I also maintain), by design.

Now, PDFjs is indeed the library that gives the best results when extracting text, and it was actually my first choice when I started working on text extraction a year ago. It works super great, but there's a catch: it doesn't scale at all. Give it more than a dozen PDFs to extract back-to-back, and it will hard crash Obsidian.

I tried your PR on my test vault, and unfortunately this problem hasn't been solved since last year, so for now, I'll stay with the "half working but not crashing" library used in Text Extractor. Sorry :/

[RyotaUshio]

Thank you for the clarification. I see, that's a sad news...