skym

A Python package exposing fuzzy text matching based on the Rust creates skim and fuzzy-matcher.

Benchmarking

This project includes a comprehensive benchmarking suite built with Criterion.rs.

Running Benchmarks

We use just as a task runner to simplify benchmark operations. Here are the common benchmarking commands:

Basic Benchmarking

# Run all benchmarks
just bench

# Open the benchmark report in your browser
just bench-report

Comparing Implementations

When making optimizations, you can compare performance between different implementations:

# Run benchmarks and save as a baseline named "original"
just bench-save original

# Make your code changes, then run and compare with the baseline
just bench-compare original

# Generate a comparison table between two sets of results
just bench-table original new

Managing Benchmark Results

# Clean all benchmark results
just bench-clean

Understanding Benchmark Results

After running benchmarks, Criterion generates comprehensive reports in target/criterion/. The most useful files are:

1. HTML Report: Open target/criterion/report/index.html in your browser for a visual representation of benchmark results including:

   • Line graphs showing performance over time
   • Violin plots showing distribution of execution times
   • Detailed statistical analysis

2. JSON Data: Raw benchmark data is stored in JSON format under target/criterion/<benchmark_name>/ for each benchmark:

   • estimates.json: Contains summary statistics (mean, median, etc.)
   • sample.json: Contains raw sample timing data
   • baseline_estimates.json: Available when comparing with a baseline