79ff3dbb74
BlackRoad Tools - ERP, CRM, manifest profiler, and DevOps utilities: Core Tools: - manifest_profile.py (937 lines) - Kubernetes manifest analysis - pully.py (735 lines) - Pull request automation - build_cluster_manifests.py (515 lines) - K8s cluster builders - erp.py (478 lines) - Enterprise resource planning - crm.py (405 lines) - Customer relationship management - build_metaverse_roster.py (331 lines) - Agent roster management - storage.py (305 lines) - Storage abstractions - agent_test_pipeline.py (304 lines) - Test automation - holo_cli.py (259 lines) - Holographic display CLI DevOps Scripts: - orin_bootstrap.sh - Jetson Orin setup - install_self_heal_pack.sh - Self-healing infrastructure - deploy_openwebui.sh - OpenWebUI deployment - triton_setup.sh - NVIDIA Triton setup Subdirectories: - branch-cleanup/ - Git branch management - kpis/ - KPI dashboards - metrics/ - Metrics collection - miners/ - Data mining tools - pulse/ - Health monitoring - tools-adapter/ - Integration adapters 11,351 lines of Python across 116 code files. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
1.2 KiB
1.2 KiB
Spiral metrics quickstart
This directory collects minimal CLIs around the pitch invariant c = d\ln r / d\theta so experiments can stay reproducible and scriptable.
Claims
spiral_pitch.pyestimates the logarithmic pitch and spiralness of complex traces.
Run
python -m tools.metrics.spiral_pitch samples.csv --real real --imag imag
Good looks like
- Stable pitch estimates across repeated runs (|Δc| < 5e-3).
- Spiralness ≥ 0.85 when the data hugs a clean log spiral.
Spiral pitch CLI (30-second brief)
Measure how log-spiral a 2D trajectory is right from the shell.
- Capture your trace as comma-separated
real,imagsamples. Any delimiter works; skip headers with--skip-rows. - Run the analyser:
python tools/metrics/spiral_pitch.py trace.csv --delimiter ,. - Read the report:
pitch_per_radianis the growth rate (c).pitch_per_turnmultiplies that by (2\pi).spiralness(\in [0,1]) is 1 minus the normalised MSE.
- Need the fitted spiral? Add
--reconstruction fit.csvforreal,imag,magnitude,thetasamples.
Inputs can stream from STDIN (-). Column selection flags (--real-col, --imag-col) unlock multi-column CSVs. Keep zeros out of the trace—the log fit assumes non-zero magnitude.