def main()

in build_pipeline/pipelines/preprocess.py [0:0]

• 16 lines of code
• 4 McCabe index (conditional complexity)

def main(base_dir):
    # Input data files
    input_dir = os.path.join(base_dir, "input/data")
    input_file_list = glob.glob(f"{input_dir}/*.csv")
    logger.info(f"Input file list: {input_file_list}")
    if len(input_file_list) == 0:
        raise Exception(f"No input files found in {input_dir}")

    # Input zones file
    zones_dir = os.path.join(base_dir, "input/zones")
    zones_file = os.path.join(zones_dir, "taxi_zones.zip")
    if not os.path.exists(zones_file):
        raise Exception(f"Zones file {zones_file} does not exist")

    # Extract and load taxi zones geopandas dataframe
    extract_zones(zones_file, zones_dir)
    zone_df = load_zones(zones_dir)

    # Load input files
    data_df = load_data(input_file_list)
    data_df = enrich_data(data_df, zone_df)
    data_df = clean_data(data_df)
    return save_files(base_dir, data_df)