API Overview

Core modules in this project:

Data and Utilities

• tabnado.data: Data loading and split creation
• tabnado.params: PipelineParams dataclass and PipelineParams.from_yaml() for loading configuration
• tabnado.utils: Logging setup, CLI argument parsing, and training utilities
• tabnado.evaluate: Metrics, scatter plots, and UMAP embeddings

GANDALF Backend

• tabnado.gandalf_sweep: Bayesian hyperparameter sweep via Weights and Biases
• tabnado.gandalf_train: Final GANDALF model training with best hyperparameters
• tabnado.gandalf_shap: SHAP analysis for GANDALF models

XGBoost Backend

• tabnado.xgb_sweep: XGBoost hyperparameter sweep
• tabnado.xgb_train: Final XGBoost model training with best hyperparameters
• tabnado.xgb_shap: SHAP analysis for XGBoost models

Orchestration

• tabnado.api: run_pipeline() orchestrates all stages
• tabnado.cli: CLI entry points for tabnado-run, tabnado-data, tabnado-sweep, tabnado-train, tabnado-evaluate, tabnado-shap, and tabnado-init

Initialization

• tabnado.init: template parameter YAML generation