Detect regions with different backgrounds

[maiiabocharova]

Hello, suppose there is a PDF page containing a column with some background color different from the main white background.
I am wondering is there a way to extract the coordinates of the regions based only on background colors using PyMuPDF? I know how to do this using PIL, but maybe there is some better and easier way without conversion, counting colors and determining the regions using some additional logic.

[JorjMcKie]

Hello, there is method page.get_drawings() which returns lines, rectangles, curves together with any fill colors.
So you can look for a rectangle in that list of dictionaries. In your case it should indeed have a fill color and be large enough, etc.
Then take that rect as the clip parameter in page.get_text(option, clip=rect, ...more args).

This are only slightly different if the background was not created by PDF drawings but by embedded images (which is rare).

[maiiabocharova]

Thank you a lot! It solved the issue, indeed much easier compared to doing in with PIL

[JorjMcKie]

Wow, that was quick!