coordinate quirk when exporting to SVG

[jesuslop]

Hi, I have an e-ink reader with stylus to sketch notes and diagrams that I get as pdf. Now am writing something to convert the pds to SVG that I can paste in a note-taking app. It's very pleasant that pymupdf does the heavy lifting of conversion. Thanks! I attach here what I have. All this prologue is to ask about using coordinate systems in pymupdf. I wrote a first version of the converter that worked charms and without problems. It exported a whole A4 sized svg version of my pdf. But that has too many white-space around the figure for practical pasting. Next I wanted to implement automatic fit-to-contents. A thing that works is chaning the media-box of the pdf before exporting the svg. The root of my question is that in the rectrangle that I'm feeding to Page.set_mediabox() the y-coordinate seems to be upside-down!

At some point I have this code

	rect = get_mediabox_rect(page)
	
	fixed_rect = pymupdf.Rect(rect)
	fixed_rect.y0=page.rect.height-rect.y1
	fixed_rect.y1=page.rect.height-rect.y0

        page.set_mediabox(fixed_rect)

that actually works OK. But I don't understand why I cannot just feed rect to set_mediabox (instead of fixed_rect (a fix I found only empirically).

Maybe it is a bug, or maybe I don't understand something about pymupdf.

My work-around solves my case, I'm reporting this for others' benefit.

py_pdf2svg_converter.py.txt
Blank Portrait_20241106.pdf

[jamie-lemon]

@jesuslop Regarding the coordinate system for PyMuPDF maybe this helps? https://pymupdf.readthedocs.io/en/latest/app3.html#coordinates

[jesuslop]

Hi @jamie-lemon, good point! Things should be more easy for the user had there been only one coordinate system in the interface.