v0.1.0

First release of causal-perception-implementation — machine learning research code for causal perception.

• Linear Additive Noise Model with a configurable causal DAG (Chiappa, 2019), do-operator and counterfactual (abduction → intervention → prediction) inference.
• Three 1-D distribution distances (Wasserstein-2, KL via KDE, Total Variation) with bootstrap confidence intervals.
• Fair-decisions experiment (demographic-parity / equal-opportunity gaps, ROC/PR, decision disagreement) on the German Credit dataset.
• German Credit data fetched on demand from OpenML (not redistributed).
• Apache-2.0 licensed. Copyright held by the author, José M. Álvarez; open-source release distributed by Santander AI Lab.

Test suite (pytest, 86% coverage) and CI/CD (lint, type-check, license & SPDX checks, dependency & pattern scans).