A codebase for running Tensorflow inference and Tesseract OCR on thumbnails of Twitch streams to create relationships between Twitch accounts and OSRS character names. The automation of managing the OSRS Streamers plugin
3 sections of the site, built with React.
- Verified Manually create, update, remove character names from streamers
- Review Double check the inference and OCR has correctly identified a new streamer character name, and quickly edit, add or dismiss
- JSON Retrieve the resulting JSON to update the streamers.json file for the OSRS Streamers plugin
- The application has been split into two dockerised containers, one responsible for crawling Twitch, running inference and OCR, the other for providing an interface for manual verification.
- The goal is to mount the same directory from the host machine on both containers to share data (streamer-getter provides, verify-streamers consumes)
- Run the verify-streamers build command from the 'verify-streamers' folder, and the streamer-getter build command from the repository root.
Building commands
docker build -t royporter7/verify-streamers . && docker run -d --name verify-streamers -it -v <ABSOLUTE PATH DIRECTORY TO THIS REPO>\streamers:/streamers --restart always -p 127.0.0.1:3002:8080/tcp -p 127.0.0.1:3003:3000/tcp royporter7/verify-streamers:latest
docker build -t royporter7/streamer-getter -f streamer-getter/Dockerfile . && docker run -d --name streamer-getter -it -v <ABSOLUTE PATH DIRECTORY TO THIS REPO>\streamers:/streamers -v <ABSOLUTE PATH DIRECTORY TO THIS REPO>\streamer-getter:/streamer-getter-dev --restart always royporter7/streamer-getter:latest
You can run the streamer-getter logic periodically from the host container using docker exec commands (bat, shell, etc), be sure to pass a Twitch user auth token.
docker exec streamer-getter python /streamer-getter.py <TWITCH USER AUTH TOKEN>