Summary

Ensure aprender APIs support the CODEX pipeline in verificar for Python-to-Rust training data generation.

Required APIs

• RandomForestClassifier with predict_proba
• GradientBoostingClassifier with predict_proba
• KMeans clustering with centroids() access
• PCA for dimensionality reduction
• Expose silhouette_score in metrics (if not already)
• TF-IDF vectorizer in text module (for code embeddings)

Integration Points

• verificar VER-050: Quality gate uses RandomForest
• verificar VER-051: Bug predictor uses GradientBoosting
• verificar VER-052: Active learning uses KMeans + PCA
• verificar VER-054: Data quality scoring uses silhouette

Acceptance Criteria

• All required APIs available and documented
• Example usage in verificar integration tests
• No breaking changes to existing APIs

Ref

verificar: docs/specifications/codex-multi-tech-python-to-rust-spec.md