Source code for openghg.store._eulerian_model

from __future__ import annotations
import numpy as np
import logging
from typing import Any

from openghg.store import DataSchema
from openghg.store.base import BaseStore

logger = logging.getLogger("openghg.store")
logger.setLevel(logging.DEBUG)  # Have to set level for logger as well as handler

__all__ = ["EulerianModel"]


# TODO: Currently built around these keys but will probably need more unique distiguishers for different setups
# model name
# species
# date (start_date)
# ...
# setup (included as option for now)




class EulerianModel(BaseStore):
    """This class is used to process Eulerian model data"""

    _data_type = "eulerian_model"
    _root = "EulerianModel"
    _uuid = "63ff2365-3ba2-452a-a53d-110140805d06"
    _metakey = f"{_root}/uuid/{_uuid}/metastore"



    def format_inputs(self, **kwargs: Any) -> dict:
        """
        Apply appropriate formatting for expected inputs for EulerianModel. Expected
        inputs will typically be defined within the openghg.standardise.standardise_eulerian()
        function.

        Args:
            kwargs: Set of keyword arguments. Selected keywords will be
                appropriately formatted.
        Returns:
            dict: Formatted parameters for this data type.
        """
        from openghg.util import clean_string, synonyms

        params = kwargs.copy()

        # Apply clean string formatting
        params["species"] = clean_string(params.get("species"))
        params["model"] = clean_string(params.get("model"))
        params["start_date"] = clean_string(params.get("start_date"))
        params["end_date"] = clean_string(params.get("end_date"))
        params["setup"] = clean_string(params.get("setup"))

        # Apply individual formatting as appropriate
        # - apply synonyms substitution for species
        species = params.get("species")
        if species is not None:
            params["species"] = synonyms(species)

        return params




    @staticmethod
    def schema() -> DataSchema:  # type:ignore[override]
        """
        Define schema for Eulerian model Dataset.

        At present, this doesn't check the variables but does check that
        "lat", "lon", "time" are included as appropriate types.

        Returns:
            DataSchema : Contains dummy schema for EulerianModel.

        TODO: Decide on data_vars checks as we build up the use of this data_type
        """
        data_vars: dict[str, tuple[str, ...]] = {}
        dtypes = {"lat": np.floating, "lon": np.floating, "time": np.datetime64}

        data_format = DataSchema(data_vars=data_vars, dtypes=dtypes)

        return data_format