Automation pipeline for large-scale Cooja simulations. Proof of concept for training ML on a dataset of simulations.

This directory contains a cleaned, publishable subset of the automation used to run headless Cooja simulations and parse results. It is intended to live alongside the original repository without overwriting legacy scripts.

Highlights:

• Modular CSC generation with a pluggable PlatformSpec (z1 default; sky/wismote preview).
• Per-run, headless execution pattern (JSON → CSC → log → CSV).
• Parsing utilities including an XY/XYZ augmenter for coordinates extracted from .csc.

Examples

1. Generate a CSC from a topology JSON:

python3 core/topology/topology_generator.py grid -n 5 -s 10 -o sample_topologies/grid5.json
python3 core/csc/csc_generator.py sample_topologies/grid5.json artifacts/example/sim.csc --build-root artifacts/example/build

2. Select a different platform (preview):

python3 core/csc/csc_generator.py sample_topologies/grid5.json artifacts/example_sky/sim.csc --platform sky

3. Augment a parsed CSV with coordinates (optional Z support):

python3 core/parse/rpl_log_add_xyz_to_csv.py --grid artifacts/example/sim.csc --input artifacts/example/results.csv --output artifacts/example/results_xy.csv
# With Z column
python3 core/parse/rpl_log_add_xyz_to_csv.py --grid artifacts/example/sim.csc --input artifacts/example/results.csv --output artifacts/example/results_xyz.csv --enable-z --z-col-name z

Notes

• Batch sweep currently sweeps the Cartesian product of N × spacing × seed. Other parameters (duration, tx, ix) are fixed per sweep; extend as needed.
• Parser and DB loader are tailored to the case study; they can be adapted for other firmware/log schemas.
• Platform specs beyond z1 are previews; validate before large sweeps / simulation runs.

• Quick: make test (runs pytest -q against tests/)
• Direct: pytest -q cooja-automation-repo/tests

Notes:

• Tests are self-contained and do not require Cooja; they validate topology JSON/CSC generation, parser behavior on a tiny synthetic log, and XY/XYZ augmentation.

To actually run headless Cooja simulations via the run scripts, you need a working Contiki-NG+Cooja setup and Java/Gradle:

• Install Contiki‑NG and Cooja per the official docs: Contiki‑NG documentation
• Ensure --cooja-root points to your Cooja folder (typically <contiki-ng>/tools/cooja).
• A JDK 17+ is required; you can pass --java-home to the run scripts if needed.

If you only want to use the topology/CSC tools and parsing utilities, you do not need Contiki‑NG/Cooja.