Text selection in cloze generation not working as expected

[ghost]

Suppose I want to generate a close from the following page

I got

A b ottom-up parse corresp onds to the construction of a parse tree for an input string b eginning at the lea v es (the b ottom) and w orking up to w ards the ro ot (the`

instead. Is there any way to fix this? I'm already using the alternative renderer.

[fonol]

Hi, unfortunately, that is how text comes out of pdf.js for some PDFs, and there is not much I can do about it. Internally, the characters in the PDF are just absolute positioned glyphs, and pdf.js tries to correctly assign/segment them into words. That works ok for most PDFs, but for some, the results are not that great. https://github.com/mozilla/pdf.js/issues/12509 https://github.com/mozilla/pdf.js/issues/9998 https://github.com/mozilla/pdf.js/issues/6705

[ghost]

Thanks for your efforts anyway! Your project really helped me a ton:P