Streaming GMAO data using fsspec and Xarray

streaming.py

#!/usr/bin/env python

import fsspec
from fsspec.implementations.http import HTTPFile, HTTPFileSystem
import xarray as xr

url = "https://portal.nccs.nasa.gov/datashare/gmao/geos-fp/das/Y2025/M03/D18/GEOS.fp.asm.inst1_2d_lfo_Nx.20250318_0000.V01.nc4"

fs = HTTPFileSystem()
file = fs.open(url)

ds = xr.open_dataset(file, engine="h5netcdf")

# Select a value at a location
ds.isel(lon=500, lat=360)["PS"].values

################################################################################
# Eventually, we would build API structured like this
def open_geospf_data_https(start_date, end_date) -> xr.Dataset:
    # Generate URLs
    if cache:
        return read_from_cache()
    urls = generate_urls_from_dates(start_date, end_date)
    files = [fs.open(url) for url in urls]
    # ds = xr.open_mfdataset(files, engine="h5netcdf")
    ds = fancy_rust_implementation_of_mfdataset(files) #...
    return ds

@fastapi(...)
def get_geospf_timeseries(ds: xr.Dataset, start_date, end_date, lon, lat):
    return ds.sel(lon=lon, lat=lat, time=slice(start_date, end_date)).values

@fastapi(...)
def get_geospf_map(ds: xr.Dataset, bbox, datetime):
    # Generate URLs
    ...

@fastapi(...)
def get_geospf_map(datetime, bbox):
    # Generate URLs
    ...