Question: using ipysegment for OCR text?

[joseberlines]

Is it possible to select part of a pdf doc in order to ocr it with whatever other library?
thx.

Use case: Official letters where you want to select and grab a part of the text to pass it to the clipboard.

[ianhi]

Unfortunately that's probably not easy. this library has the assumption that your image can represented as a numpy array baked in pretty deep. So if you can convert your PDF to to a with rasterized image then it could work, but it would no longer be a pdf.

[joseberlines]

It's not important that it remains a pdf. So to be clear the pdf has to be converted in a numpy "rastered image"?

[ianhi]

I think I may have been out of my depth when I used words like rastered what I mean is, get it into a numpy array (i.e. something that you could call plt.imshow on with matplotlib) and it should work.

[ianhi]

Maybe you could use this https://github.com/Belval/pdf2image to get the array first.