Files
manifests/raw/ekman/cronjobs/arome/cron.yaml
T

152 lines
4.8 KiB
YAML

apiVersion: v1
kind: ConfigMap
metadata:
name: arome-script
namespace: cron
data:
download.py: |
import os
import sys
from netCDF4 import Dataset
import re
from datetime import datetime, timedelta
fname ="https://thredds.met.no/thredds/dodsC/meps25epsarchive/YEAR/MONTH/DAY/meps_det_sfc_YEARMONTHDAYT00Z.ncml"
outdir = "/data/hdd/data/AROME"
def generate_thredds_names(start, stop):
start_date = datetime(int(start.split("-")[0]),
int(start.split("-")[1]),
int(start.split("-")[2]))
end_date = datetime(int(stop.split("-")[0]),
int(stop.split("-")[1]),
int(stop.split("-")[2]))
date_list = []
while start_date <= end_date:
date_list.append(start_date)
start_date += timedelta(days=1)
fileList = []
for date in date_list:
y = str(date.year)
m = (str(date.month)).zfill(2)
d = (str(date.day)).zfill(2)
f = re.sub("YEAR", y, fname)
f = re.sub("MONTH", m, f)
f = re.sub("DAY", d, f)
fileList.append(f)
return fileList
def copy_thredds_file(threddsFile, savename):
dsin = Dataset(threddsFile)
dsout = Dataset(savename, "w")
for dname, the_dim in dsin.dimensions.items():
dsout.createDimension(dname, len(the_dim) if not the_dim.isunlimited() else None)
aromeNames = ["time",
"longitude",
"latitude",
"land_area_fraction",
"air_temperature_2m",
"precipitation_amount_acc",
"water_evaporation_amount",
"relative_humidity_2m",
"integral_of_surface_downwelling_longwave_flux_in_air_wrt_time",
"integral_of_surface_net_downward_shortwave_flux_wrt_time",
"air_pressure_at_sea_level",
"x_wind_10m",
"y_wind_10m"]
for v_name, varin in dsin.variables.items():
if v_name in aromeNames:
fill_value = None
if hasattr(varin, "_FillValue"):
fill_value = varin._FillValue
outVar = dsout.createVariable(v_name, varin.datatype, varin.dimensions, fill_value=fill_value)
outVar.setncatts({k: varin.getncattr(k) for k in varin.ncattrs() if k not in ["_FillValue"]})
outVar[:] = varin[:]
dsout.close()
os.makedirs(outdir, exist_ok=True)
fList = generate_thredds_names("2026-04-24", datetime.today().strftime("%Y-%m-%d"))
failed = False
for fname in fList:
savename = os.path.join(outdir, fname.split("/")[-1].split(".")[0] + ".nc")
if os.path.exists(savename):
print(f"Skipping {savename}, already exists")
continue
print(savename)
try:
try:
copy_thredds_file(fname, savename)
except:
alt_fname = re.sub("sfc", "2_5km", fname)
alt_fname = re.sub("ncml", "nc", alt_fname)
copy_thredds_file(alt_fname, savename)
except Exception as e:
print(f"File not found: {fname} ({e})")
failed = True
if failed:
sys.exit(1)
---
apiVersion: batch/v1
kind: CronJob
metadata:
name: arome
namespace: cron
spec:
schedule: 0 6 * * * # Everyday at 06:00, use https://crontab.guru
concurrencyPolicy: "Forbid" # If only one at at time set to Allow else Forbid
successfulJobsHistoryLimit: 10
failedJobsHistoryLimit: 3
jobTemplate:
spec:
backoffLimit: 3
template:
spec:
restartPolicy: "Never"
containers:
- name: cronpod
image: juselius/busynix:1.1
imagePullPolicy: IfNotPresent
command:
- /bin/sh
- -c
- |
if nix-shell -p 'python3.withPackages(ps: [ps.netcdf4])' --run 'python3 /scripts/download.py'; then
chown -R 5000:5000 /data/hdd/data/AROME
chmod -R g+w /data/hdd/data/AROME
else
echo "Job failed, sleeping 30 minutes before retry..."
sleep 1800
exit 1
fi
resources: {}
volumeMounts:
- name: data
mountPath: /data
- name: script
mountPath: /scripts
securityContext: {}
volumes:
- name: data
persistentVolumeClaim:
claimName: ekman-data
- name: script
configMap:
name: arome-script
defaultMode: 0755