command
Version:
v0.7.3
Opens a new window with list of versions in this module.
Published: May 7, 2025
License: GPL-3.0
Opens a new window with license information.
Imports: 19
Opens a new window with list of imports.
Imported by: 0
Opens a new window with list of known importers.
README
¶
cmd/bench
cmd/bench hosts the benchmarking CLI for Launchrail.
Responsibility
- Provide a
launchrail bench command to run performance benchmarks.
- Load CSV flight and benchmark datasets via
LoadFlightInfo and related helpers.
- Compare actual vs expected metrics with configurable tolerances.
- Exit with status code 0 on success, 1 on any failure (for CI integration).
Contract
- Exposes
main.go and benchmark.go defining Benchmark and BenchmarkSuite.
- Implements
BenchmarkSuite.Run() to orchestrate data loads and result comparison.
Test Suite Overview
main_test.go (using testify) covers:
- CSV data loading (
datastore_test.go).
- Example hipr-euroc24 benchmark logic (
hipr_euroc24_test.go).
- Exit-code behavior for pass/fail.
Decisions & Gotchas
- Zero expected values switch tolerance logic to absolute difference.
- Tolerance scaling chosen to balance noise vs strictness; may need tuning per dataset.
- CSV schema changes must be mirrored in
datastore.go loaders.
Documentation
¶
There is no documentation for this package.
Source Files
¶
Click to show internal directories.
Click to hide internal directories.