[go] settling-eval + docs

[Copilot]

This pull request introduces a new offline evaluation tool for LiDAR background grid settling convergence and makes several related updates to documentation and build tooling. The most significant change is the addition of the settling-eval command-line tool, which replays a PCAP file to evaluate convergence metrics and outputs a JSON report. Supporting changes include updates to the Makefile to add a target for this tool, and improvements to documentation and references for clarity and accuracy.

New LiDAR Settling Evaluation Tool:

• Added a new command-line tool settling-eval in cmd/tools/settling-eval/, which replays a PCAP file offline through the background grid manager, evaluates convergence on every frame, and generates a JSON report with metrics and a recommended warmup frame count. [1] [2]

Build and Tooling Updates:

• Updated the Makefile to add a run-settling-eval target, allowing easy invocation of the new evaluation tool with configurable PCAP, port, tuning config, and output file. [1] [2]

Documentation and Reference Improvements:

• Updated references to DESIGN.md in README.md and CONTRIBUTING.md to point to the correct location under docs/ui/. [1] [2]
• Refreshed docs/BACKLOG.md to clarify upcoming features, hardening phases, and roadmap priorities, including new and reordered items for platform simplification and LiDAR foundations.

[Copilot]

Pull request overview

This PR implements Phase 3 of the LiDAR background settling optimization: convergence/evaluation tooling. It provides data-driven metrics to guide WarmupMinFrames tuning by tracking four convergence indicators (coverage rate, spread delta rate, region stability, and mean confidence) through a new API endpoint and CLI tool.

Changes:

• Adds SettlingMetrics computation library with convergence detection
• Exposes /api/lidar/settling_eval endpoint for real-time metric polling
• Provides settling-eval CLI tool that polls metrics and generates recommendations with 20% safety margin
• Updates design documentation and backlog to reflect Phase 3 completion

Reviewed changes

Copilot reviewed 8 out of 9 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
internal/lidar/l3grid/settling_eval.go	Core convergence metrics computation with four indicators and threshold evaluation
internal/lidar/l3grid/settling_eval_test.go	Comprehensive unit tests for metrics computation and convergence detection
internal/lidar/l3grid/background.go	Adds prevSpreads and prevRegionIDs fields to BackgroundGrid for stateful delta tracking
internal/lidar/monitor/webserver.go	New API endpoint returning metrics, thresholds, and convergence status
cmd/tools/settling-eval/main.go	CLI tool that polls API and outputs JSON evaluation report
Makefile	Adds build-settling-eval target
.gitignore	Excludes settling-eval binary
docs/lidar/operations/settling-time-optimization.md	Updates Phase 3 status to complete with implementation details
docs/BACKLOG.md	Moves settling optimization Phase 3 from v0.5 to Complete

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[ddol]

@copilot coverage report says 30 lines missing coverage on this branch. add tests to hit 95.5%+ for changes on this branch

[Copilot]

@copilot coverage report says 30 lines missing coverage on this branch. add tests to hit 95.5%+ for changes on this branch

Added tests in 760bed5:

• IsSettlingComplete() — 4 tests (nil manager, nil grid, not settled, settled) → 0% → 100%
• EvaluateSettling zero-cells branch → 97.8% → 100%
• handleSettlingEval — 3 coverage tests in webserver_coverage_test.go (missing sensor_id, no manager, success with JSON validation)

All new functions now at 100% coverage.