Thanks for even looking. This only becomes useful if a lot of people build and poke at it, so anything from a typo fix to a new firmware feature is welcome.
- Embedded / firmware: power management, LoRa uplink, OTA updates, three-axis fusion, fixed-point tuning on the ESP32.
- Seismology / DSP: better characteristic functions, magnitude estimation, telling a real P-wave from a door slam.
- Backend / distributed systems: scaling the gateway, time sync, keeping fake triggers out.
- Hardware: a real PCB, a decent enclosure, better sensor coupling.
- Just building one: seriously, deploy a node and tell me where it breaks. That's the most useful report I can get right now.
git clone https://github.com/tylrcc/tremor-mesh
cd tremor-mesh
pip install -e .
pytest -qpython 5_Algorithms/demo.py and python 7_Simulations/network_sim.py are the quickest way to see things move.
- If you touch the detector, change it in both
tremormesh/stalta.pyand4_Firmware/node/node.ino, and update the tests. The Python is the source of truth. - Tests have to pass before I'll merge. CI runs
pyteston every PR. - Keep PRs small and say why, not just what.
- Don't commit generated junk (plots, build output, binaries). It's git-ignored for a reason.
Use the issue tracker for bugs and ideas. If you find a way to make the network fire a false alarm or stay quiet when it shouldn't, that's a security thing, flag it (see SECURITY.md) so it gets looked at first.
By sending a PR you're agreeing your code goes out under MIT and any hardware/docs under CERN-OHL-P, same as the rest of the repo.