Remove everything except images

[nsklei]

I am looking for a solution to remove all elements from a PDF (text, paths ...) except images.

My current solution extracts all images from the document and inserts them inside a new document.
Is there any simpler way to solve this problem, because extracting and inserting the images feels like a workaround?

[JorjMcKie]

Sure there is a canonical way: Use redaction annotations!

1. add a redaction for the full page: page.add_redact_annot(page.rect)
2. apply redactions using an option that ignores images for the removal: page.apply_redactions(images=fitz.PDF_REDACT_IMAGE_NONE).
3. Save the file with sufficient garbage collection and compression: doc.ez_save("redacted.pdf").

Since the recent version, all 3 categories, images, graphics and text can be selectively kept or removed.

[nsklei]

I already tried that, but it did not work completely.
Most of the text elements were removed, but some of them and also the path elements are still present.

import fitz

doc = fitz.open("Testfile.pdf")

for page in doc:
    page.add_redact_annot(page.rect)
    page.apply_redactions(images=fitz.PDF_REDACT_IMAGE_NONE)

doc.ez_save("Testfile_redacted.pdf")
doc.close()

Testfile.pdf

[JorjMcKie]

That happen when the PDF was sloppily created: with text or drawings not completely inside page.rect.
For drawings, you can / should use the option that erase even when only parts are overlapping.
I believe fitz.PDF_REDACT_LINE_ART_REMOVE_IF_TOUCHED (=2).
For stubborn text, simply increase the redaction rectangle like so page.rect + (-20, -20, 20, 20) which is a rect 20 points larger in every direction.

[nsklei]

Yep, that worked. Thank you :)