This screencast is part 1 of 2. It shows how to scrape imagery and data from the USDA's online collection of pomological watercolours. This sets up an imageset to be used in the generative pattern designer. For more context on the project, you can read my previous post.
The scraping is done using NodeJS with the node-fetch and cheerio libraries. Once all the images and accompanying data are gathered we further process the images by cropping them slightly (10 pixels each way). To this end I demonstrate using ImageMagick's 'chop' command.
You can find the code used in the screencast in the following gist.