saltybet elo tracker
twitch_dump.py needs a directory vid_dir in order to work properly
latest tesseract-ocr training data is in ocr_training/try2/eng2.traineddata this part is fairly rigid, and making it more flexible is on TODO top things TODO to optimize OCR:
- add/makesureisused eng1.user-words to tessdata directory. populate with fighter names automatically somehow
- make the image filter more advanced
possible corrections to matches obtained from challonge: is DAZZLER actually THE DAZZLER? is DRAGONCLAW actually DRAGON CLAW?