main.py

import pandas as pd


def master_data_csv_to_parquet(file_path: str):
    """
    Convert master data export from a CSV to a Parquet file.
    """
    save_path = file_path.replace(".csv", ".parquet")
    categorical_columns = [
        "Monitoring_Location_ID",
        "Activity_Depth_Unit",
        "Sample_Position",
        "Time_Zone",
        "Activity_Type",
        "Waterbody_Class",
        "WBID",
        "Name",
        "Sector",
        "Total_Depth_Unit",
        "Org_Analyte_Name",
    ]

    dtype_dict = {
        "Station_Number": str,
        **{col: "category" for col in categorical_columns},
    }

    return (
        pd.read_csv(file_path, dtype=dtype_dict, low_memory=False)
        .assign(
            Org_Result_Value=lambda df: pd.to_numeric(
                df["Org_Result_Value"].replace("Not Reported", pd.NA), errors="coerce"
            ),
            Activity_Start_Date_Time=lambda df: pd.to_datetime(
                df["Activity_Start_Date_Time"].apply(
                    lambda x: f"{x} 12:00:00" if len(str(x).strip()) <= 10 else x
                ),
                errors="coerce",
            ),
            Org_Result_Unit=lambda df: df["Org_Result_Unit"].str.replace("none", ""),
            Station_Number=lambda x: x["Station_Number"].apply(
                lambda s: f"{float(s):.2f}"
                if pd.notna(s) and str(s).strip() != ""
                else s
            ),
        )
        .to_parquet(save_path)
    )