use-ml-feature

Feature metadata primitives for RustUse machine-learning workflows.

Experimental

use-ml-feature is experimental while use-ml remains below 0.3.0.

Example

use use_ml_feature::{MlFeatureKind, MlFeatureName, MlFeatureRole};

let name = MlFeatureName::new("sepal_width")?;
let kind: MlFeatureKind = "numeric".parse()?;

assert_eq!(name.as_str(), "sepal_width");
assert_eq!(kind, MlFeatureKind::Numeric);
assert_eq!(MlFeatureRole::Input.as_str(), "input");
# Ok::<(), use_ml_feature::MlFeatureError>(())

Scope

• Feature names, identifiers, kinds, roles, and sources.
• Transform, encoding, scaling, missing-value, and drift-status labels.
• Lightweight validation for ASCII-safe feature names.

Non-goals

• Performing feature engineering, transformations, or pipeline orchestration.
• Prompt variables, message fields, context-window fields, or tool-call arguments.

License

Licensed under either Apache-2.0 or MIT.