Skip to content

Uncertainty (quantile)

version: 0.1.0
trait: uncertainty
profile: quantile

View spec source on GitHub

Run with uvx from release on pypi.org:

uvx --with mlwp-data-specs mlwp.validate_dataset_traits --uncertainty quantile <DATASET_PATH_OR_URL>

Warning: mlwp-data-specs is not published on PyPI yet, so this command is included for future release usage.

Run directly from GitHub source:

uvx --from "git+https://github.com/mlwp-tools/mlwp-data-specs" mlwp.validate_dataset_traits --uncertainty quantile <DATASET_PATH_OR_URL>

Python API:

from mlwp_data_specs import validate_dataset

report = validate_dataset(ds, uncertainty="quantile")

1. Introduction

This document defines trait-level requirements for uncertainty representation in MLWP datasets. The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119.

2. Scope

This specification applies to datasets validated with the uncertainty=quantile profile.

3. Structural Requirements

3.1 Accepted Dimension Variants

  • The dataset MUST include the quantile dimension for uncertainty representation.

3.2 Required Coordinates

  • The dataset MUST include the quantile coordinate.

3.3 Optional Coordinates

  • The dataset MAY include optional coordinates for this profile: [].

3.4 Optional Dimensions

  • The dataset MAY include optional dimensions for this profile: [].

4. Coordinate Metadata Requirements

  • quantile MUST have standard_name equal to quantile.
  • quantile MUST have units equal to 1.
  • All quantile coordinate values MUST be within [0, 1].