Source code for veriflow.datasinks.cf_compliant_netcdf

"""Read and write netcdf files in a fews compatible format."""

from datetime import datetime, timezone
from pathlib import Path

from veriflow.configuration.default.datasinks import CFCompliantNetCDFConfig
from veriflow.constants import NAME, VERSION
from veriflow.datasinks.base import BaseDatasink
from veriflow.datatree.datatree import VeriflowDataTree

__all__ = [
    "CFCompliantNetCDF",
    "CFCompliantNetCDFConfig",
]


[docs] class CFCompliantNetCDF(BaseDatasink): """For writing veriflow output to NetCDF. This datasink will write one NetCDF file for each verification pair. Input data are included and results are written in NetCDF groups. Each file is named ``"<stem>_<verification_pair_id> <suffix>"``, where ``<stem>``/``<suffix>`` are derived from the configured ``filename``. .. note:: CF-compliancy is not yet fully implemented. """ kind = "cf_compliant_netcdf" config_class = CFCompliantNetCDFConfig def __init__(self, config: CFCompliantNetCDFConfig) -> None: self.config: CFCompliantNetCDFConfig = config @staticmethod def _create_output_directory(directory: Path) -> None: """Create the output directory if it does not exist.""" Path(directory).mkdir(parents=True, exist_ok=True)
[docs] def write_data(self, dt: VeriflowDataTree) -> None: """Write the data in the xarray DataTree to the file as specified in the output config.""" # Create the output directory if it does not exist self._create_output_directory(self.config.directory) filepath = Path(self.config.directory) / self.config.filename # Metadata attrs according to CF-compliancy attrs = { "title": self.config.title, "institution": self.config.institution, "source": f"{NAME}: version: {VERSION}", "history": "", "references": "", "comment": self.config.comment, "time_coverage_start": self.config.verification_period.start.isoformat(), "time_coverage_end": self.config.verification_period.end.isoformat(), "production_time": datetime.now(tz=timezone.utc).isoformat(), "Conventions": "CF-1.11", } filtered_dt = dt.veriflow.filter_nodes( include_input_data=self.config.include_input_data, include_aligned_input_data=self.config.include_aligned_input_data, include_output=self.config.include_output, ) # Persist the `xr.DataTree` attributes to the filtered DataTree filtered_dt.attrs = attrs def validate_write_allowed(filepath: Path, *, force_overwrite: bool) -> None: if not force_overwrite: msg = "File already exists: " + str(filepath) raise FileExistsError(msg) validate_write_allowed(filepath, force_overwrite=self.config.force_overwrite) filtered_dt.to_netcdf(filepath)