clankers-cli 0.1.3

clankers-cli-0.1.3 is not a library.

Visit the last successful build: clankers-cli-0.1.6

Scaffold projects, replay MCAP logs, validate ONNX models, and run the camera-perception demo — the clankers binary for day-to-day robotics workflows.

Install

cargo install clankers-cli

Templates for clankers new are bundled in the crate — no repo clone required.

clankers --help

Quick start

# New project from a template (works from any directory)
clankers new my_robot --template perception-node
cd my_robot
clankers run

# Inspect and replay your own MCAP log
clankers inspect path/to/camera_log.mcap
clankers replay path/to/camera_log.mcap

# Validate ONNX against stored PyTorch references
clankers validate-model \
  --onnx models/detector.onnx \
  --samples path/to/inputs/

# Golden-path demo (requires a repo clone for sample_data/)
git clone https://github.com/PvRao-29/clankeRS.git && cd clankeRS
clankers demo camera-perception

Commands

Command	Purpose
`new`	Scaffold from `basic-node`, `perception-node`, `ml-inference-node`, or `replay-test-node`
`run`	Run the current project's node
`test`	Run replay tests and `cargo test`
`inspect`	Summarize an MCAP file
`replay`	Replay a log (optionally through a node)
`latency`	Latency stats from a replay
`compare`	Diff two MCAP files
`validate-model`	ONNX vs PyTorch reference outputs
`bench`	Benchmark inference latency and copy/allocation stats
`import-pytorch`	Export a checkpoint to ONNX
`add-model`	Register a model in `clankeRS.toml`
`visualize`	MCAP summary for Foxglove / Rerun
`demo`	Run bundled demos (`camera-perception`)
`record`	MCAP recording (stub — not complete yet)

Learn more

License

MIT — see LICENSE.