Dynamic loading of NetCDF data

This example illustrates how OMAS can load in memory only the data when it is first requested. What is done here for NC file, also works also for IMAS.

import os
from omas import *

# set OMAS_DEBUG_TOPIC to see when data is loaded dynamically
os.environ['OMAS_DEBUG_TOPIC'] = 'dynamic'

# generate some data and save it as a netcdf file
ods = ODS().sample(ntimes=2)
ods.save('test.nc')

# ODS.open() will keep the file descriptor open so that OMAS
# can load in memory only the data when it is first requested
# NOTE: one can use the `with` statement or open()/close()
ods = ODS()
with ods.open('test.nc'):
    # data gets read from NC file when first requested
    print(ods['equilibrium.time_slice.:.global_quantities.ip'])
    # then it is in memory
    print(ods['equilibrium.time_slice.0.global_quantities.ip'])

# save it as a pickle
ods.save('test.pkl')

# load the data back
ods = ODS()
ods.load('test.pkl')

# the data that was loaded is stored in the pickle
print(ods.flat().keys())

# continue loading more data
with ods.open('test.nc'):
    print(ods['equilibrium.time'])