Extract and replace test in PDF

[adrian5FDF]

Hi all, 

I have  a pdf generated in appl Pages. The pdf has several data tables / text boxes per page.

Wondering if the following is possible:

Detect all text content on page, whether in tables or text boxes, etc. 

Extract text and replace w/ other text, thus generating the same doc w/ entirely new text. 

I've looked at some of the OCR and text extraction samples/docs. Extracting the data seems pretty strait forwards. But can I replace it? The use case is translating the doc to another language.

[Joel Geraci]

Adobe doesn't have a Document Services API to do this. That said, it's not really something you'll want to do. Text in a PDF is generally laid down with precise coordinates and isn't able to be replaced without causing overlaps.

If you want to create new PDF files with different text, I suggest looking at the Document Generation API which will allow you to start with a Word template plus some JSON and output a PDF where the JSON is merged into tagged "fields" in the document.

[adrian5FDF]

Hi Joel, thanks for the feedback.

That makes sense. Looks like using a template is the way to go. We don't
use Word though, and also manually tagging our complex documents would be
really time consuming.

I wonder if something like this would work:
> extract text using the extract API, thus getting content and coordinates
> populate a new blank PDF with the data, using the coordinate info to
insert text in blank document
( I would use another PDF js library for this part, if needed)

I downloaded the sample json output from the Extract API, and see a bunch
of coordinate data.
Could I use that to position text in a new document?

[Joel Geraci]

Do you control the authoring process of these documents before they get converted to PDF or is this situation where you need to take what you are given?

[adrian5FDF]

Yes, I create them in Apple Pages.

[Joel Geraci]

So you have the source files? Why do you need the Extract service then? I'm missing something.

[adrian5FDF]

The pdf's change depending on our clients' needs — we use tables & formulas to show/hide different options, and we are always making customiztion to them on the fly - adding notes, etc.. And they change over time. I want to completely automate translating new features in the file. 

The end goal is to translate the pdf's. The translation part I have covered (google translate api).