CLI Reference

Configuration

Configuring the Rule Engine via CLI flags and SDK arguments

Featuresmith v0.1.0 supports programmatic rule configurations in the Python SDK and command-line flags in the CLI. File-based configuration is not yet active.

Roadmap Notice: File-Based Config (.featuresmith.yml)

Layered file-based configuration (via a .featuresmith.yml file at the project root) is a planned enhancement scheduled for Phase 2+. In the current release, configure rules directly in code or use CLI flags.

1. Python SDK Configuration

Configure rules programmatically by passing the enabled_rules list and the rule_config dictionary to the fs.analyze() function:

python

1import featuresmith as fs
2
3result = fs.analyze(
4    "dataset.csv",
5    target_column="churn",
6    enabled_rules=[
7        "quality.missing_value_threshold",
8        "statistical.high_correlation"
9    ],
10    rule_config={
11        "quality.missing_value_threshold": {"threshold": 15.0},
12        "statistical.high_correlation": {"threshold": 0.85}
13    }
14)

2. CLI Options

Control CLI analyze parameters using flags (such as filtering by severity or limiting correlation analysis sizes):

bash

1# Target column and correlation column limits
2featuresmith analyze dataset.csv --target churn --max-correlation-columns 50
3
4# Severity filtering and report generation
5featuresmith analyze dataset.csv --severity warning --output report.txt