Page.show_pdf_page() function alternative/work around

[TimHuynh0905]

Hi there,

I am trying to find drawings inside a specific area (clip) of a page in a similar manner to Page.get_text(). However, both Page.get_drawings() and Page.get_cdrawings() do not provide that functionality out of the box.

I wonder if there is a way to work around this issue. If possible, why don't you add a clip parameter inside the Page.get_drawings() and Page.get_cdrawings()?

One way I considered using to work around this problem is to use Document.show_pdf_page(..., clip = ). But as being said in the documentation, all resources "will be imported ... even if they are not contained in the visible area given by clip." Thus, Page.get_cdrawings() still won't work the way I want.

References:

• https://pymupdf.readthedocs.io/en/latest/page.html#Page.get_cdrawings
• https://pymupdf.readthedocs.io/en/latest/page.html#Page.show_pdf_page

Thank you for your help!

Cheers.

[JorjMcKie]

The simplest thing to do surely is

paths = [p for p in page.get_drawings() if p["rect"] in clip]

Done.

[JorjMcKie]

If you however had in mind to include the subset of a drawing that is partly contained in the clip (and some parts outside) ...
This is not possible to implement.
So please let me know your intention.

[TimHuynh0905]

Thank you so much for your responses!

In my current project, my intention is to extract an area out of a page and export it as a stand-alone pdf file. I want to do some analyses specifically on this area only. Apparently, the show_pdf_page() function also imports all resources outside of the visible clip area, thus get_drawing() outputs things I don't need. Do you have any recommendation to work around it?

Thank you!

[JorjMcKie]

If you set a page's /CropBox to the area of interest, you will at least only get stuff displayed from this area.

Other than mentioned above I have no advice. This is BTW the reason why I detest introducing a clip parameter: (1) the above one-line is too trivial to make a big deal out of it, and (2) it does not work like the clip parameter of other methods (pixmaps, text extractions,...): there, a precise cut-out of the clip happens, whereas a drawing cannot be partially contained - it is either in or out completely.

Well, it seems that if you make an SVG image from a page with modified cropbox, the results may better. Haven't checked how far this algorithm goes though.
At least once you have the SVG, you could convert it back to a PDF ...
Tell me your findings please.