def tabular_munge_data()

in src/autotrain/project.py [0:0]

• 49 lines of code
• 17 McCabe index (conditional complexity)

def tabular_munge_data(params, local):
    if isinstance(params.target_columns, str):
        col_map_label = [params.target_columns]
    else:
        col_map_label = params.target_columns
    task = params.task
    if task == "classification" and len(col_map_label) > 1:
        task = "tabular_multi_label_classification"
    elif task == "classification" and len(col_map_label) == 1:
        task = "tabular_multi_class_classification"
    elif task == "regression" and len(col_map_label) > 1:
        task = "tabular_multi_column_regression"
    elif task == "regression" and len(col_map_label) == 1:
        task = "tabular_single_column_regression"
    else:
        raise Exception("Please select a valid task.")

    exts = ["csv", "jsonl"]
    ext_to_use = None
    for ext in exts:
        path = f"{params.data_path}/{params.train_split}.{ext}"
        if os.path.exists(path):
            ext_to_use = ext
            break

    train_data_path = f"{params.data_path}/{params.train_split}.{ext_to_use}"
    if params.valid_split is not None:
        valid_data_path = f"{params.data_path}/{params.valid_split}.{ext_to_use}"
    else:
        valid_data_path = None
    if os.path.exists(train_data_path):
        dset = AutoTrainDataset(
            train_data=[train_data_path],
            task=task,
            token=params.token,
            project_name=params.project_name,
            username=params.username,
            column_mapping={"id": params.id_column, "label": col_map_label},
            valid_data=[valid_data_path] if valid_data_path is not None else None,
            percent_valid=None,  # TODO: add to UI
            local=local,
            ext=ext_to_use,
        )
        params.data_path = dset.prepare()
        params.valid_split = "validation"
        params.id_column = "autotrain_id"
        if len(col_map_label) == 1:
            params.target_columns = ["autotrain_label"]
        else:
            params.target_columns = [f"autotrain_label_{i}" for i in range(len(col_map_label))]
    return params