Source code for pycif.plugins.datastreams.fields.lmdz_outfields_nc.read

import numpy as np
import xarray as xr
import os
import datetime
import tracemalloc
from .....utils.hdf5 import _hdf5_lock




def read(
    self,
    name,
    varnames,
    dates,
    files,
    interpol_flx=False,
    tracer=None,
    model=None,
    **kwargs
):
    """Read a variable from LMDz ``trajq`` files into a pyCIF DataArray.

    Opens each distinct file in ``files`` once, matches each requested
    date against the file's ``time`` coordinate (since the file may cover
    a different month/year range than strictly requested), and stacks the
    selected ``(lev, lat, lon)`` slices into a single array.

    Args:
        self: The field/tracer Plugin.
        name: Name of the component; used as fallback variable name.
        varnames: Name of the variable to read; if empty, ``name`` is used
            instead.
        dates: List of date entries to extract, matching ``files``.
        files: List of files matching ``dates``; each distinct file is
            opened only once.
        interpol_flx: Unused.
        tracer: Unused.
        model: Unused.
        **kwargs: Unused.

    Returns:
        xarray.DataArray: A 4-dimensional ``(time, lev, lat, lon)`` array.
    """
    # Reading required lmdz files
    trcr_lmdz = []
    trcr_dates = []
    data = None
    file_ref = ""
    for dd, file_concs in zip(dates, files):
        if file_concs != file_ref:
            file_ref = file_concs
            # Reading file
            with _hdf5_lock:
                with xr.open_dataset(file_concs) as ds:
                    data = ds[varnames if len(varnames) > 0 else name].load()

                    # Flip upside down
                    # TODO: check once for all whether to flip North-South...
                    # data = data[:, :, ::-1]

                    # Get correct data index
                    # as the month/year may be different
                    # in the file than the one required
                    times = ds["time"].to_pandas().dt.to_pydatetime()

        ind = np.where(times == dd[0])[0][0]
        
        # Appending
        trcr_lmdz.append(data[ind, ...].values)
        trcr_dates.append(dd[0])
    
    # Putting the data into an xarray
    xmod = xr.DataArray(
        trcr_lmdz, coords={"time": trcr_dates}, dims=("time", "lev", "lat", "lon")
    )

    return xmod