Examples

The repository includes two notebooks that walk through the main ifcfill workflows with small pandas DataFrames:

• Basic usage without categorical encoding
• Label encoding for synthetic data

The notebooks cover:

• Using ifcfill around tabular synthetic-data workflows without a target variable
• Creating sample tabular data with numeric, categorical, datetime, missing, and constant columns
• Running IFCTransformer.fit_transform
• Keeping categoricals unencoded with the default cat_encoding="none"
• Label encoding filled categorical variables with inverse-compatible mappings using cat_encoding="label"
• Inspecting mapping dictionaries with get_category_mappings() and get_category_mapping(...)
• Inspecting inferred column types, fill values, dropped constants, and missing_report_
• Saving and loading fitted transformation state for later inverse transformation
• Restoring the original table structure with inverse_transform

To run the notebooks:

pip install "ifcfill[examples]"
jupyter notebook

Run these commands in your preferred Python environment.

The base install stays lean:

pip install ifcfill

Use the examples extra only when you want the notebook dependencies.