Takes an SVG image consisting of a number of filled paths, breaks it up into shapes, and then re-arranges the shapes to make a 'deconstructed' image.
I took a few drawing classes online, and these classes caused me to notice more of the beauty already around me. I started to view trees more as visual objects, pure creations of simple concepts like, color, and shape. Trees are incredible works of art, when viewed only in terms of how they use these basic forms.
For example, consider line. The only things that characterize a line are its angle and width. When I looked at how a tree 'uses' lines, I noticed that there is a wide distribution of line widths. The trunk is a very wide line, branches get increasingly smaller, until we get to twigs, and even the veins in the leaf of a tree. The bigger lines tend to have very restricted angles - the trunk usually goes straight up. As lines get smaller and smaller, they start to go every which way.
I was really hoping to see what this distribution of line angles by width would actually look like. As a first pass, I thought i'd try breaking up an SVG image of a tree. I naively thought that the svg image would be a rough approximation of the tree's geometry, but this turns out to not be the case.
I still haven't gotten to see what I actually watned to see, but now I have found one approach that didn't work - although it did make some neat pictures.
This uses python 3, and you'll need to install the libraries requirements.txt using pip. Run the program
python parse_tree.py --file_name any-svg-file.svg > output-here.svg
The program may through an error if it encounters svg path elements different from the ones in the test images i used.
Enjoy!