Python SubX download_data produces an error before scripts can be created

[kdl0013]

In file, SubX/Python/download_data/generate_full_py_ens_files.ksh the code will produce a list of all files, but it immediately fails at fen='python tmp.py' and produces the following error:

RuntimeError: NetCDF: Access failure oc_open: server error retrieving url: code=6 message="request too large"

The link provided https://iridl.ldeo.columbia.edu/SOURCES/.Models/.SubX/.RSMAS/.CCSM4/.hindcast/.tas/dods does not appear to have any files to open which may be the problem. IRIDL may have updated the site to not allow dods information to populate through.

[raghu330]

I too face the same issue "oc_open: server error retrieving url: code=6 message="request too large"". I added 'decode_times=False', but the error still persists.

[kdl0013]

I too face the same issue "oc_open: server error retrieving url: code=6 message="request too large"". I added 'decode_times=False', but the error still persists.

@raghu330 You can use this python script I created to download SubX data. This script will create a wget list and you can run that script to download the files.

#!/usr/bin/env python3

'''Create wget script for each SubX model to setup download of files in parallel 

Inputs: IRIDL NCAR username and password
Inputs -- you can change the model and variables that you want downloaded

Outputs: Shell script which contains download info. Files will download 20 at a time when run.

'''
import os
import datetime as dt
import numpy as np


username_IRIDL = "usrname"
password_IRIDL = "psswd"

models = ['GMAO']
sources = ['GEOS_V2p1']
#vars = ['huss', 'dswrf','mrso','tas', 'uas', 'vas','tdps','pr','cape']
vars = ['tasmax','tasmin']

# get the dates for all SubX models (this is all the possible dates) I started at year 2000 because other SubX models have all started by then
start_date = dt.date(2000, 1, 5)
end_date = dt.date(2015, 12, 26)

dates = [start_date + dt.timedelta(days=d) for d in range(0, end_date.toordinal() - start_date.toordinal() + 1)]
#GMAO GEOS specifically skips leap days
dates_out = []
for i in dates:
    if i.month == 2 and i.day == 29:
        pass
    else:
        dates_out.append(i)

#Only every 5th date
dates=dates_out[::5]

### GMAO GEOS_V2p1 model


#vars=['tas']


new_dir=(f'/glade/scratch/klesinger/SubX/{models[0]}')
os.system(f'mkdir {new_dir}')

count=0    
output=[]
output.append('#!/bin/bash')
for m_i, model in enumerate(models):
    
    for d_i, date in enumerate(dates):
        
        for v_i, var in enumerate(vars):
        
            date_str = '{}-{}-{}'.format(str(date.year), str(date.month).rjust(2,'0'), str(date.day).rjust(2,'0'))

            command = f"wget -nc --user {username_IRIDL} --password {password_IRIDL} 'http://iridl.ldeo.columbia.edu/SOURCES/.Models/.SubX/.{model}/.{sources[m_i]}/.hindcast/.{var}/S/(1200%20{str(date.day)}%20{date.strftime('%b')}%20{str(date.year)})/VALUES/data.nc' {new_dir}/{var}_{model}_{date_str}.nc &"
            
            if count % 20 == 0:
                output.append('wait')
            
            count+=1
            output.append(command)
            
np.savetxt('wget_GMAO.txt',output, fmt="%s")

os.system("cat wget_GMAO.txt > wget_GMAO.sh")
os.system("rm wget_GMAO.txt")

After running the script, use command line to bash wget_GMAO.sh

[raghu330]

Hello @kdl0013 Thanks very much for the response and for an excellent wget solution. My apologies for the delay in responding. I was able to help the person who came up with his issue to me using a similar but convoluted approach, but yours is better optimized. Since then, I am yet to hear further from him again, so my involvement has reverted back to my official duties. Thank you very much! Cheers Raghu