Simple example of how to use the ORM

Import necessary stuff



In [1]:

    
import modelmeta
from sqlalchemy import create_engine
from sqlalchemy.orm import sessionmaker
from optparse import OptionParser

Create the database session



In [2]:

    
engine = create_engine('postgresql://httpd_meta@atlas.pcic/pcic_meta?sslmode=require')
Session = sessionmaker(bind=engine)
session = Session()

Query all of some type of object



In [3]:

    
q = session.query(modelmeta.Ensemble)
q.count(), [x.name for x in q.all()]









    Out[3]:





(9,
 ['canada_map',
  'test_ensemble_small',
  'downscaled_canada',
  'bccaq_downscale_canada',
  'bcsd_downscale_canada',
  'test_ensemble',
  'bc_prism',
  'vic_gen1',
  'bccaq_extremes'])

Look up objects based on query parameters



In [4]:

    
ensemble_name = 'canada_map'
mydatafilevars = session.query(modelmeta.DataFileVariable).\
join(modelmeta.EnsembleDataFileVariables).\
join(modelmeta.Ensemble).\
filter(modelmeta.Ensemble.name == ensemble_name).all()
print('Ensemble: ' + ensemble_name + ' with ' + str(len(mydatafilevars)) + ' data_file_vars')









    



Ensemble: canada_map with 330 data_file_vars

Or better yet, use the built in relational mapping



In [5]:

    
myensemble = session.query(modelmeta.Ensemble).filter(modelmeta.Ensemble.name == 'canada_map').first()
print('Ensemble: ' + myensemble.name + ' with ' + str(len(myensemble.data_file_variables)) + ' data_file_vars')









    



Ensemble: canada_map with 330 data_file_vars

You can map all the way down the rabbit hole



In [6]:

    
mymodel = myensemble.data_file_variables[0].file.run.model
print(mymodel.short_name)









    



BCCI+ANUSPLIN300+CanESM2

And back up again



In [7]:

    
mymodel.runs[0].files[0].data_file_variables[0].ensembles[0].name









    Out[7]:





'canada_map'

And around in circles



In [8]:

    
myensemble.data_file_variables[0].\
file.run.emission.runs[0].files[0].data_file_variables[0].\
grid.data_file_variables[0].variable_alias.data_files[0].\
data_file_variables[0].ensembles[0].name









    Out[8]:





'canada_map'



In [ ]: