Applying clip path to SVG objects

[vkovenko]

Greetings. I need to crop pdf along with its SVG. I came up with an approach consisting of the following steps:

1. Get SVG objects using fitz get_drawings method.
2. Crop objects by applying geometrical functions.
3. Convert objects back to SVG using svgwrite.
4. Convert SVG to PDF by applying cairosvg.
   Everything seemed to work okay, but I found that filling of specific objects with color doesn't work as expected (arrows in black):
   Expected result:

Actual result: Then I found that clip paths can be extracted by setting `extended=True ` in `get_drawings` method. I believe that clip paths are used there in order to fill objects with color w.r.t their shape. The question is how to convert those clip paths back to SVG? Maybe you would have some suggestion. Thank you!

[JorjMcKie]

Sorry for the tardy answer.
Unfortunately, PyMuPDF does not currently support the output of clipped vector graphics or text.
There also are no plans for the foreseeable future.

[more-strive]

May I ask if the author's problem has been resolved

[vkovenko]

Yes. I needed to get clip-path from fitz and assign them to specific or inside SVG. I found the other way around for cropping and rotating SVG objects without using fitz for getting parsed SVG objects in python (I still use fitz for getting svg from PDF in a fast manner, but parse objects myself in a custom way using xml, which gives an opportunity to gather all the tags from SVG along with its attributes, including clip-path).

[more-strive]

Can you share the demo of this section

[vkovenko]

Sorry, but the project I'm working on is under NDA, thus I have no right to share extensive information about the topic.

[more-strive]

Okay， yft-design This repository is my open-source image editor. When loading a PDF, there are still images left that cannot be resolved by the clip. If there is any other help, I would be very grateful