What is this?

An attempt at shape recognition for the reMarkable tablet. It's pretty rough around the edges, so it won't win any user experience awards any time soon.
Currently, the shape recognizer from Xournal++ is used - it operates on stroke data directly instead of trying to use bitmap recognition. You can see a description of how it works here.

Eventually I might write my own shape recognizer...

Building

rmkit is included as a submodule - make sure to clone using --recursive, or git submodule update --init after cloning.

make docker will use the rmkit Docker container and output into artifacts/symbiote - this is recommended.
Otherwise, just running make will use the locally installed toolchain, and output into build/symbiote.

Using

Currently there is literally no UI - run it via SSH, draw something, then tap anywhere to stop and draw out the shapes.

Acknowledgements

• rmkit - a light library that handles most of the things I want to deal with
• iago/lamp from rmkit - reference on how to fake input events, inspiration