Dataset/library of parsed ffmpeg codec information

To know ahead of time information about codecs (ie. whether a given codec can be muxed into a given container), we need to have a database of codec information. While this wikipedia page has some of this info, it's not complete and doesn't match with ffprobe's output.

That's where this comes in. This repo generates a JSON dataset of file extensions to the container corresponding to this extension's supported ffmpeg codecs. It does this using a best effort upper bound approach by parsing FFMPEG's source code for codec ids in a file containing the muxer associated with a given extension. This means that it's possible for some false positives where a certain codec is marked as supported when it's not. For these cases we have a blacklist that explicitly removes these false positives.