Search problems

[vlad-babin]

Hello!
I have issues with searching in the document.
Here is the script that performs the search and deletes the lines.

input_file_path = request.kwargs["inputFilePath"]
search_value = "Reference is made"

doc = pymupdf.open(input_file_path)

for page in doc:
    instances = page.search_for(search_value)
    for inst in instances:
        page.add_redact_annot(inst, fill=(0, 0, 0))
    page.apply_redactions()

doc.save('redacted_document.pdf')
doc.close()

The problem is that the search results provide coordinates slightly above the actual line in the file.

This issue occurs in files that had the problem described in this link before version 1.24.14.

I’m attaching the files for reference.
original.pdf
redacted.pdf

This problem appeared in version 1.25.0 (and persists in 1.25.1 as well).

I’m using version 1.25.0.

[JorjMcKie]

I found the reason for this weird behavior.
There was a behavior change in MuPDF between the PyMuPDF versions you have observed:
The font binary may (and usually does) contain the float values "ascender" and "descender" which control which portions of the glyphs appear above and below the baseline. Descender is / must be a negative value.
In addition the font object defined in the PDF may contain overriding values for the same. Since recent versions, MuPDF always accepts the PDF-defined values over the font-internal values, because this has proved to deliver a more correct extraction behavior.
In the case of your PDF however, the PDF object font definition contains a positive descender value. Therefore, the bbox computations are wrong and too much upward by the double value of the font descender.

So I suggest you open an issue - either here or directly at MuPDF's issue system, https://bugs.ghostscript.com/enter_bug.cgi.
If you decide to open a PyMuPDF issue (my recommendation), I will forward it to MuPDF. Please do not forget to attach again your original PDF again.
Thank you.