JSONLogic Benchmark

This benchmark suite allows you to compare the performance of different versions of the JSONLogic Go implementation.

Methodology

The benchmark compares two versions of the JSONLogic Go implementation:

1. Published Version: The specified release version from GitHub
2. Current Version: Your local code changes

The benchmark:

• Creates temporary Go modules for both versions
• Runs the Go benchmark tests with the same set of test cases for both versions
• Compares the results across multiple metrics:
  • Execution time (ns/op)
  • Memory usage (B/op)
  • Memory allocations (allocs/op)
• Calculates improvement/degradation percentages for each metric
• Provides an overall summary showing average improvements

Test Cases

The benchmark uses a variety of JSONLogic operations to thoroughly test performance:

• Simple equality checks
• Complex conditions with logical operators
• Variable access with nested paths
• Array operations
• Conditional logic

Requirements

• Go (1.21+)
• Python 3.x
• Python packages: colorama (installed automatically with uv)
• uv (Python package manager)

Usage

Run the benchmark by executing the bench script with a published version to compare against:

./bench v3.7.5

Where v3.7.5 should be replaced with the specific version tag you want to compare against.

Understanding the Results

The benchmark results display:

• A side-by-side comparison of each test case for both versions
• Percentage differences for time, memory, and allocations
• Positive percentages (green) indicate improvements in your current code
• Negative percentages (yellow) indicate performance regressions
• Average improvements across all test cases