Collect judge data interactively instead of hand-editing judges.json?

Seems we could eliminate the hand-editing-JSON step in the [contribution guidelines](https://github.com/freelawproject/judge-pics/blob/master/README.md#contributing) by adding some interactive prompts in the `convert_images.py` script, or a wrapper around it. That might make it a little easier for novices to contribute. On the other hand, the effort might better spent making a web form instead.
