Discussion: New parser to prevent lossy conversion

[Sodium-Hydrogen]

svg2shenzhen currently uses kicads bitmap2component tool for some layers and a svg parser for the other unsupported layers. I don't think this is quite ideal as converting back and forth from bitmaps is a trade off of processing time and quality.

Would there be interest in overhauling backend to use a tool like svg2mod so it is vector to vector conversion which I see as less lossy and it has the advantage of parsing all layers uniformly?

Svg2mod has received a lot of recent upgrades and is a lot more reliable now and has the added benefit of easier installation for use with other python scripts like inkscape plugins.

[badgeek]

Hi, initialy i used svg2mod as converter, see dev log here http://wiki.8bitmixtape.cc/#/4_7.1-PCB-Art-with-Kicad-and-Inkscape?id=development-log but there was problem with overlapping shape, and i just aware that you maintain svg2mod now, one idea came to my mind, what if we can choose each layer converter? use bitmap2component or svg2mod

[Sodium-Hydrogen]

I would be interested in helping that happen. Yesterday I officially listed svg2mod as a pypi package that can be installed with pip so setup for using it would not be to complicated. There are still some bugs with multi hole parsing (see svg2mod/svg2mod#8) but I am trying to work on fixing these issues and migrating features that svg2shenzhen has to svg2mod so they are comparable tools.

Funny enough I was looking at svg2shenzhen to see how it worked and to see if it would be something I was interested in using instead of svg2mod. I like the ui, but I've already put a lot of time to updating svg2mod and I didn't want to abandon that after I had just said I would take over maintenance. And I don't know why but vector to bitmap to vector has always bothered me 🤷‍♂️