entropice/cds.py

"""Download ERA5 data from the Copernicus Data Store.

Web platform: https://cds.climate.copernicus.eu
"""

import os
import re
from pathlib import Path

import cdsapi
import cyclopts
from rich import pretty, print, traceback

traceback.install()
pretty.install()

DATA_DIR = Path(os.environ.get("DATA_DIR", "data")) / "entropyc-rts"


def hourly(years: str):
    """Download ERA5 data from the Copernicus Data Store.

    Args:
        years (str): Years to download, seperated by a '-'.

    """
    assert re.compile(r"^\d{4}-\d{4}$").match(years), "Years must be in the format 'YYYY-YYYY'"
    start_year, end_year = map(int, years.split("-"))
    assert 1950 <= start_year <= end_year <= 2024, "Years must be between 1950 and 2024"

    dataset = "reanalysis-era5-single-levels"
    client = cdsapi.Client(wait_until_complete=False)

    outdir = (DATA_DIR / "era5/cds").resolve()
    outdir.mkdir(parents=True, exist_ok=True)

    print(f"Downloading ERA5 data from {start_year} to {end_year}...")
    for y in range(start_year, end_year + 1):
        for month in [f"{i:02d}" for i in range(1, 13)]:
            request = {
                "product_type": ["reanalysis"],
                "variable": [
                    "2m_temperature",
                    "total_precipitation",
                    "snow_depth",
                    "snow_density",
                    "snowfall",
                    "lake_ice_temperature",
                    "surface_sensible_heat_flux",
                ],
                "year": [str(y)],
                "month": [month],
                "day": [f"{i:02d}" for i in range(1, 32)],
                "time": [f"{i:02d}:00" for i in range(0, 24)],
                "data_format": "netcdf",
                "download_format": "unarchived",
                "area": [85, -180, 50, 180],
            }

            outpath = outdir / f"era5_{y}_{month}.zip"
            client.retrieve(dataset, request).download(str(outpath))
            print(f"Downloaded {dataset} for {y}-{month}")


if __name__ == "__main__":
    cyclopts.run(hourly)