Trouble with searching

[vlad-babin]

Description of the bug

Hello!
I have issues with searching in the document.

I searched for it to redact the word "tour," and it is not.
It is redacting some, but not all, and weird blocks, such as "Roa" in Roadshow and only "s" in Tours on the cover page.

I’m attaching the file for reference.
6.0-8 Deal Roadshow SiteTour by Finsight (2).pdf

How to reproduce the bug

Enter the word “tour” in the search bar.

PyMuPDF version

1.25.1

Operating system

MacOS

Python version

3.12

[JorjMcKie]

Found the problem! Not sure whether this is caused by the base library, so I will submit an issue there and report back here.
In essence, your PDF uses tags (i.e. has a "Structure Tree") which confuses standard text search and extraction for some reason.
Removing these tags establishes the normal situation again.
Taking into account that you want to modify the PDF (redactions), removing the tags is required anyway and will be a new option in one of the next versions.
Here's how to do this:

doc = open("6.0-8.Deal.Roadshow.SiteTour.by.Finsight.2.pdf")
doc.xref_set_key(doc.pdf_catalog(), "StructTreeRoot", "null")  # remove the tags info

# from here on, everything works as expected

[JorjMcKie]

MuPDF bugzilla item: https://bugs.ghostscript.com/show_bug.cgi?id=708323

[JorjMcKie]

Update:

The MuPDF team has identified the root cause: This file contains errors in its Structure Tree definition.
They have developed a fix which will detect and circumvent the problem.
The fix will become available in one of the next PyMuPDF versions - in which case my above circumvention will become obsolete.

I am going to reclassify this issue as "upstream enhancement".

[julian-smith-artifex-com]

Fixed in PyMuPDF-1.26.0.