First swing at ZipFile support. (Issue #16)

[JosephCatrambone]

These changes are a little messy. I think the IndexedImage file needs a better way to read from an in-memory image. The handling of zipfiles is also not great, and I think when we start monitoring for on-disk changes it's going to get worse. Perhaps the crawler's ingesting methods could be split off into a separate file.

WIP for #16

[JosephCatrambone]

Current change breaks things.

...
Error processing E:\GraphicsAssets\wood_patterned_21-color.jpg: The image format could not be determined
Error processing E:\GraphicsAssets\xoana_splash.png: The image format could not be determined
Error processing E:\GraphicsAssets\youwin.png: The image format could not be determined
...

[JosephCatrambone]

Problem was I was only doing buffer.read() instead of buffer.read_to_end(). The one thing that should probably get cleaned up before this is merged is the view screen, which expects that the file is readable from disk. If it comes in as a zipfile and we can't just open it, we might hit problems. Perhaps the indexedimage needs an extra field for 'image is stored as zip'?