by Liuser on 3/4/17, 2:04 PM with 2 comments
by Liuser on 3/4/17, 2:07 PM
The core is written in Python and nested in the Cocoa app. It uses scikit-image and PIL libraries. I actually originally wrote this project as a web application that allowed users to convert images online. It had an easy to use mobile web interface for users to upload images and download conversion. The upsell was for users to pay for high resolution images. I ultimately decided to remove the web app because I felt the quality of the images it was producing wasn't that great due to certain technical constraints.
It takes real estate to draw the nested shapes, thus drawing work better on larger images. The conversion is inherently CPU intensive. The server I was renting had one processor and 512 RAM, which works well for images less than 1000 pixels. Anything larger requires some gymnastics such as cutting the picture into 100 smaller pictures, processing, then stitching it back together. I also didn't want to upgrade the servers too much because that becomes a re-occuring high monthly cost for a small project with a low feature set.
Next steps: - Working on some additional effects. - Vector output. PIL doesn't support vector output by default, but was easy to hit the ground running with. pycairo would be a good candidate.
by owly on 3/5/17, 1:44 PM