How do I copy/paste from a searchable pdf?

jude7 · 12-04-2023, 04:04 PM

I am running Mint 18.2, Cinnamon 3.4.3, and all I want to do is copy/paste from a searchable pdf. I have looked in LibreOffice writer, and xreader, and GIMP, but I don't see any way to do this....

Seems like this should be easy, but I'm not seeing it.....

rkelsen · 12-04-2023, 04:53 PM

Try using LibreOffice Draw.

goldennuggets · 12-04-2023, 05:17 PM

Did you try Atril?
What about opening the PDF in your web browser and copying the text from there?

jude7 · 12-04-2023, 05:21 PM

I'll try it.....

jude7 · 12-04-2023, 05:41 PM

Ok, It doesn't really work: When I highlight the text in the pdf, only about 50% actually gets highlighted, and that is all that will "paste", so I end up with mainly letters with spaces between them, not words.

I should add that the searchable pdf was made from an online utility to convert the pdf to a supposedly searchable form, and it seemed to work fine.

maybe I need a better conversion program, because it looks as if the converter didn't recognize all the letters, but there is nothing wrong with the original text - it's clear and sharp.

goldennuggets · 12-04-2023, 05:43 PM

Quote:

Originally Posted by jude7

Ok, It doesn't really work: When I highlight the text in the pdf, only about 50% actually gets highlighted, and that is all that will "paste", so I end up with mainly letters with spaces between them, not words.

I should add that the searchable pdf was made from an online utility to convert the pdf to a supposedly searchable form, and it seemed to work fine.

maybe I need a better conversion program, because it looks as if the converter didn't recognize all the letters, but there is nothing wrong with the original text - it's clear and sharp.

It does sound as if the pdf document you're attempting to copy from is corrupted in some way or wasn't formatted properly. You're on the right track to go back to the source.

pan64 · 12-05-2023, 02:06 AM

I don't know what kind of pdf is it, but sometimes they contain images of pages. In that case you cannot mark/select text. You can use an OCR to convert it to text (in that case).

goldennuggets · 12-05-2023, 06:44 AM

You might also be able to upload it to an AI service, even chatgpt or perplexity and ask it to provide you with the text.

business_kid · 12-05-2023, 08:29 AM

Have you tried pdftotext?

teckk · 12-06-2023, 08:54 PM

pdftotext works fair. If you have a two colum pdf it's worthless, or one with images. Pdf is just not a format made for that. Pdf should be an end product not a source.

Depends on how the pdf is made.

Example:

cat file1.txt

Code:

Ok, It doesn't really work: When I highlight the text in the pdf, 
only about 50% actually gets highlighted, and that is all that will 
"paste", so I end up with mainly letters with spaces between them, 
not words.

I should add that the searchable pdf was made from an online utility 
to convert the pdf to a supposedly searchable form, and it seemed to 
work fine.

maybe I need a better conversion program, because it looks as if the 
converter didn't recognize all the letters, but there is nothing wrong 
with the original text - it's clear and sharp.

Code:

libreoffice --convert-to "pdf" file1.txt

toc.txt

Code:

[/Page 1 /View [/XYZ null null null] /Title (My Bookmark) /OUT pdfmark

Code:

gs -o file2.pdf -sDEVICE=pdfwrite toc.txt -f file1.pdf

Open file1.pdf and file2.pdf in pdf viewer. All the text is copy-able.

jefro · 12-12-2023, 03:25 PM

Can we assume the author of the work has not enabled cut?

jude7 · 12-12-2023, 03:58 PM

I'm not overly sophisticated on that sort of thing - wish I was - but since it's a screen-cap of as doc saved by me as a pdf, I would imagine the answer must be no. It's likely a pdf that's not going to offer any help.

goldennuggets · 12-12-2023, 04:07 PM

Quote:

Originally Posted by jude7

I'm not overly sophisticated on that sort of thing - wish I was - but since it's a screen-cap of as doc saved by me as a pdf, I would imagine the answer must be no. It's likely a pdf that's not going to offer any help.

You're going to need to use an OCR service or AI to read the text first and convert it to actual text. Then you'll be golden.

jude7 · 12-12-2023, 04:11 PM

OK, thanks.

Not the news I wanted, but at least I won't waste more time looking for something that's not available.

goldennuggets · 12-12-2023, 04:12 PM

Quote:

Originally Posted by jude7

OK, thanks.

Not the news I wanted, but at least I won't waste more time looking for something that's not available.

You're welcome. There's a lot of free options out there like this one: https://www.iloveocr.com