Extracting text from PDF?

Question

There is a PDF file, which contains mainly text. However some parts of these texts are impossible to select and copy:

And when I select this:

So:

ult. Optimist

becomes

ult. Op mist

What is causing this, and can I somehow overcome on this limitation?

I'd suggest that this behaviour is likely to depend on the software you're running: i.e. the reader plus any underlying support libraries as determined by the operating system. — Mark Morgan Lloyd, Commented Jun 18, 2023 at 7:02

RedGrittyBrick · Accepted Answer · 2023-06-17 20:09:36Z

4

It may be because the program that wrote the PDF has "flattened" the "ti" ligature into a graphics line-draw/fill object.

It might be that the "ti" ligature encoding is not recognised by the target into which you pasted the text.

It looks like the first is more likely the case if your image represents the text selection in a PDF viewer itself.

There are many other difficult problems in extracting plain text from all possible PDFs.

So there might not be any easy solution for this.

answered Jun 17, 2023 at 20:03

RedGrittyBrick

83.6k20 gold badges139 silver badges212 bronze badges

1

This. Funny enough, the built-in OCR of macOS handles this image fine... but OCR usually is not what you want. Encoding of ligatures and of eastern European characters often is a weak point. For professional purposes, there are some libraries (like PDFlib TET) which usually handle this a bit better.
– jvb
Commented Jun 17, 2023 at 20:22

Add a comment |

1 Answer 1