Source code for pycif.plugins.datastreams.fields.oldlmdz_photochem.fetch

import os
import datetime
import pandas as pd
import xarray as xr
from .....utils import path
from .....utils.hdf5 import _hdf5_lock




def fetch(ref_dir, ref_file, date_interval, target_dir, tracer=None, **kwargs):
    """Fetch legacy LMDZ4 photolysis-rate files and their sub-intervals.

    For each date in ``date_interval`` (stepped at ``tracer.file_freq``),
    links the corresponding file into ``target_dir`` if it exists, and
    reads its ``time_counter`` dimension to build one 24h sub-interval
    per time step found in the file.

    Args:
        ref_dir: directory where the original files are found.
        ref_file: (template) name of the original files.
        date_interval: list of two dates, the beginning and end of the
            simulation.
        target_dir: directory where links to the original files are
            created.
        tracer: the fields Plugin, giving access to ``file_freq``.

    Returns:
        list_files: for each date, the file path repeated once per time
            step found in that file.
        list_dates: for each date, the list of 24h ``[start, end]``
            sub-intervals covered by that file's time steps.
    """
    # Reshape input interval to include full months
    datei, datef = date_interval

    list_period_dates = \
        pd.date_range(datei, datef, freq=tracer.file_freq).to_pydatetime()
    list_dates = {}
    list_files = {}
    for dd in list_period_dates:
        dout = datetime.datetime(dd.year + 1, 1, 1)
        file = dd.strftime(f"{ref_dir}/{ref_file}")
        if os.path.isfile(file):
            # Fetching
            target_file = f"{target_dir}/{os.path.basename(file)}"
            path.link(file, target_file)

            # Time stamps
            with _hdf5_lock:
                ds = xr.open_dataset(target_file)
                file_hours = list(
                    ds["time_counter"].to_pandas().index.to_pydatetime()
                )
            list_dates[dd] = [
                [hh0, hh0 + datetime.timedelta(hours=24)]
                for hh0 in file_hours]

            list_files[dd] = (len(list_dates[dd]) * [target_file])

    return list_files, list_dates