lawkit

日本語

Statistical law analysis toolkit. Analyze data for Benford's law, Pareto principle, Zipf's law, and more. Detect anomalies and assess data quality.

Why lawkit?

Traditional tools analyze one pattern at a time. lawkit analyzes multiple statistical laws simultaneously:

$ lawkit benf financial_data.csv
Benford Law Analysis Results

Dataset: financial_data.csv
Numbers analyzed: 1000
[LOW] Dataset conforms to Benford's Law

First Digit Distribution:
1: ██████████████████████████████  30.1% (expected: 30.1%)
2: ██████████████████              17.8% (expected: 17.6%)
3: ████████████                    12.5% (expected: 12.5%)
...

$ lawkit pareto sales.csv
80/20 Rule: Top 20% owns 79.2% of total

Installation

# As CLI tool
cargo install lawkit

# As library (Cargo.toml)
[dependencies]
lawkit-core = "2.5"

Usage

# Analysis commands
lawkit benf data.csv          # Benford's law (fraud detection)
lawkit pareto data.csv        # Pareto principle (80/20 rule)
lawkit zipf data.csv          # Zipf's law (frequency distribution)
lawkit normal data.csv        # Normal distribution (quality control)
lawkit poisson data.csv       # Poisson distribution (rare events)

# Integration commands
lawkit analyze data.csv       # Multi-law analysis
lawkit validate data.csv      # Data validation
lawkit diagnose data.csv      # Detailed diagnostics

# Generate test data
lawkit generate benf -s 1000
lawkit generate pareto -s 500

# Utility commands
lawkit list                   # List available laws
lawkit selftest               # Run self-test

Supported Input

File: lawkit benf data.csv
URL: lawkit benf https://example.com/data.json
stdin: cat data.csv | lawkit benf -

Formats: CSV, JSON, YAML, plain text (one number per line)

Main Options

-f, --format <FORMAT>      # Output: text, csv, json, yaml, toml, xml
-q, --quiet                # Minimal output
-v, --verbose              # Detailed output
--filter <RANGE>           # Filter numbers (e.g., >=100, <1000, 50-500)
-c, --min-count <N>        # Minimum data count (default: 10)
--no-color                 # Disable colors
-t, --threshold <LEVEL>    # Detection threshold: low, medium, high, critical
--confidence <LEVEL>       # Confidence level (0.01-0.99, default: 0.95)
--sample-size <N>          # Max sample size for large datasets
--min-value <VALUE>        # Minimum value to include

Risk Levels

Level	Meaning	Exit Code
LOW	Data conforms to expected distribution	0
MEDIUM	Minor deviation, likely normal	0
HIGH	Significant deviation (p ≤ 0.05)	10
CRITICAL	Severe anomaly (p ≤ 0.01)	11

Other exit codes: 1 (general error), 2 (argument error)

CI/CD Usage

# Detect anomalies in financial data
if ! lawkit benf transactions.csv --quiet; then
  echo "Anomaly detected"
  lawkit benf transactions.csv --format json > report.json
fi

# Validate data quality with cross-validation
lawkit validate data.csv --cross-validation

Documentation

License

MIT

lawkit 2.5.16