Source code for pycif.plugins.obsparsers.obspack.parse_nc

import xarray as xr
from ....utils.check.errclass import CifError
from logging import info



def parse_nc(obs_file,flags=False):
    # Open the file
    ds = xr.open_dataset(obs_file)

    # Construct the dataframe
    col2use = ["value", "latitude", "longitude",
               "altitude", "obspack_num","value_std_dev","qcflag"]
    df = ds[col2use].to_dataframe().rename(
        columns={
            "value": "obs",
            "longitude": "lon",
            "latitude": "lat",
            "altitude": "alt",
            "value_std_dev":"obserror"
        }
    ).set_index("obspack_num")

    # Deal with dates
    df.loc[:, "date"] = ds["time"].values

    # Deal with time duration
    if "time_interval" in ds:
        df.loc[:, "duration"] = \
            ds["time_interval"].dt.total_seconds().values / 3600

    elif "dataset_data_frequency" in ds.attrs:
        df.loc[:, "duration"] = 1. / ds.attrs["dataset_data_frequency"]

    else:
        raise CifError(
            f"No information about obs duration in the file {obs_file}.\n"
            "The variable 'time_interval' or the attribute 'dataset_data_frequency' "
            "should be defined in your ObsPack file."
        )

    # Replace fixed parameters
    df.loc[:, "station"] = ds.attrs["site_code"]
    df.loc[:, "parameter"] = ds.attrs["dataset_parameter"]
    df.loc[:, "network"] = "ObsPack"
    #df.loc[:, "obserror"] = 0.

    # Rescale depending on unit -> convert to ppb
    unit = ds["value"].attrs["units"]
    info(f'Obs units are {unit}')
    if unit == "mol mol-1":
        info(f'conversion obs from {unit} to ppb')
        df.loc[:, "obs"] *= 1e9
    unit = ds["value_std_dev"].attrs["units"]
    info(f'Obs error units are {unit}')
    if unit == "mol mol-1":
        info(f'conversion obs errors from {unit} to ppb')
        df.loc[:, "obserror"] *= 1e9 
    if flags :
      df["qcflag"] = df["qcflag"].str.decode('utf-8') 
      info(f'Selection of data with flags into {flags}')
      df = df[df["qcflag"].isin(flags)]
      info(f"Data size : {len(df.index)}")
    return df