Editing the text in PDF

[priyamharsh14]

I was looking for a way to edit or replace some text inside a PDF file. I read some suggestions on #257, but I am unable to perform it. When I extracted xref stream, there were some Tj and TJ, and the text data was inside <> brackets, but when I tried to decode it (hex to string), the decoded text is not valid.

Example,
<000e001c00140003001b0019001a0017001b001b0017001300130014> Tj

I am unable to find the text data stream which I want to replace.

[JorjMcKie]

It cannot work this way!
The hex chars you see there are (in the vast majority of cases) the result of a quite complex translation process, which involves a table (called "CMAP") of the used font.
So you would have to employ redaction annotations for this. The basic approach being:

1. Locate the text, say "needle".
2. Add a redaction annotation with the hit bbox as parameter and apply annotations. This will remove "needle".
3. Insert "newtext" at the location of "needle".

There are a number of additional complexities coming in:

• Use of the same font? If yes, consider that it may be a subset font, which may not have all characters contained in "newtext".
• How long is "newtext"? If shorter or longer than "needle", the result will look ugly.
• "needle" could have been (partly) invisible or transparent. This information cannot be extracted right now. The replacement "newtext" will be definitely visible! The text color however can be extracted and reused for "newtext".
• ...

[priyamharsh14]

Okay. Thank you for your detailed answer.

[JorjMcKie]

then you did not correctly locate the text, or you did not invoke page.apply_redactions()

[priyamharsh14]

Yes, I forgot to apply the redactions. My bad.