# =============================================================================
# SuperModelingFactory - Requirements
# =============================================================================
# Full dependency list of the credit-risk modeling toolkit
# Install: pip install -r requirements.txt

# ---------------------------------------------------------------------------
# Core dependencies (required by every sub-package)
# ---------------------------------------------------------------------------
pandas>=1.5.0,<3
numpy>=1.23.0
scipy>=1.9.0

# ---------------------------------------------------------------------------
# Modeling engine (required by Modeling_Tool)
# ---------------------------------------------------------------------------
scikit-learn>=1.1.0
joblib>=1.2.0
python-dateutil>=2.8.0

# ---------------------------------------------------------------------------
# Gradient boosting models
# ---------------------------------------------------------------------------
lightgbm>=3.3.0
xgboost>=1.7.0
catboost>=1.2.0

# ---------------------------------------------------------------------------
# Visualization
# ---------------------------------------------------------------------------
matplotlib>=3.5.0
seaborn>=0.12.0

# ---------------------------------------------------------------------------
# Excel report engine (required by ExcelMaster / Report)
# ---------------------------------------------------------------------------
xlsxwriter>=3.0.0
openpyxl>=3.0.0
Pillow>=9.0.0

# ---------------------------------------------------------------------------
# Progress bars
# ---------------------------------------------------------------------------
tqdm>=4.64.0

# ---------------------------------------------------------------------------
# Optional dependencies
# ---------------------------------------------------------------------------

# Hyperparameter optimization (used by WOE_Monotone_Binner)
optuna>=4.0.0

# Statistical models (LR stepwise regression, VIF)
statsmodels>=0.13.0

# Imbalanced learning (SMOTE-style sampling, SampleBalancer)
imbalanced-learn>=0.10.0

# Model explainability (Explainability.Model_Explainer; pip install supermodelingfactory[explain])
shap>=0.44.0

# Alibaba Cloud MaxCompute connection (ODPS_Tool)
pyodps>=0.11.0

# H2O machine-learning platform (utils contains an H2O helper function)
# h2o>=3.40.0
