def get_options(): global options parser = lp.prediction_options() parser = lp.ga_options(parser) parser = lp.data_options(parser) parser.add_option("--model", dest="model", help="The model class that the genomes instantiate", default=None) parser.add_option( "--test-set", dest="test_set", action="store_true", help= "Test the genomes on the test set, rather than on the training set", default=False) parser.add_option("--plot", dest="plot", action="store_true", help="Make a plot (in combination with --test-set)", default=False) (options, args) = parser.parse_args() lp.options = options if options.model is None: print >> sys.stderr, "Model argument is required." sys.exit(1)
def parse_args(args=None): parser = argparse.ArgumentParser(description="Estimate vector models with different levels of regularization.") parser.add_argument( "--save-path", required=True, help="path to file where results should be saved. " 'Will be appended with ".losses.npy" and ".forecasts.npy".', ) parser.add_argument( "--bc-data", action="store_const", const="bc-transmission", dest="dataset", help="Use transmission data from BC Columbia.", ) parser.add_argument( "--total-load", action="store_const", const="total-load", dest="dataset", help="Use distribution data from Norway.", ) parser.add_argument( "--gef-data", action="store_const", const="gefcom-2012", dest="dataset", help="Use GEFCom 2012 data." ) parser.add_argument( "--vanilla", action="store_true", dest="vanilla", help="Use the vectorized vanilla model instead of ARX." ) # Trying to load mpi4py causes segfault on some systems if the # program is not actually launched under MPI (i.e. via mpirun or # similar). parser.add_argument("--no-mpi", action="store_true", dest="no_mpi", help="Do not load any MPI libraries.") # parser.add_argument('--total-load', action='store_true', help='Use distribution data.') # parser.add_argument('--gef-data', action='store_true', help='Use GEFCom 2012 data.') parsed_args = parser.parse_args(args) lp_cmdline = ["--data-seed=0", "--standardize"] if not parsed_args.vanilla: lp_cmdline += ["--subtract-weekly-pattern"] if parsed_args.dataset is None: raise RuntimeError("A dataset argument must be supplied.") if parsed_args.dataset == "bc-transmission": lp_cmdline += ["--bc-data", "--remove-holidays"] elif parsed_args.dataset == "total-load": lp_cmdline += ["--total-load"] elif parsed_args.dataset == "gefcom-2012": lp_cmdline += ["--gef-data"] raise RuntimeError( "GEFCom data is not supported yet, remains to assign a " "set of weights for the different temperature zones." ) else: raise RuntimeError("Unrecognized dataset: {}".format(parsed_args.dataset)) parser = lp.prediction_options() parser = lp.data_options(parser) parser = ga.ga_options(parser) lp_options, _ = parser.parse_args(lp_cmdline) lp.options = lp_options return parsed_args, lp_options
def parse_args(args=None): parser = argparse.ArgumentParser(description= 'Estimate vector models with different levels of regularization.') parser.add_argument('--save-path', required=True, help='path to file where results should be saved. '\ 'Will be appended with ".losses.npy" and ".forecasts.npy".') parser.add_argument( '--bc-data', action='store_const', const='bc-transmission', dest='dataset', help='Use transmission data from BC Columbia.') parser.add_argument( '--total-load', action='store_const', const='total-load', dest='dataset', help='Use distribution data from Norway.') parser.add_argument( '--gef-data', action='store_const', const='gefcom-2012', dest='dataset', help='Use GEFCom 2012 data.') parser.add_argument( '--vanilla', action='store_true', dest='vanilla', help='Use the vectorized vanilla model instead of ARX.') # Trying to load mpi4py causes segfault on some systems if the # program is not actually launched under MPI (i.e. via mpirun or # similar). parser.add_argument( '--no-mpi', action='store_true', dest='no_mpi', help='Do not load any MPI libraries.') # parser.add_argument('--total-load', action='store_true', help='Use distribution data.') # parser.add_argument('--gef-data', action='store_true', help='Use GEFCom 2012 data.') parsed_args = parser.parse_args(args) lp_cmdline = ['--data-seed=0', '--standardize'] if not parsed_args.vanilla: lp_cmdline += ['--subtract-weekly-pattern'] if parsed_args.dataset is None: raise RuntimeError('A dataset argument must be supplied.') if parsed_args.dataset == 'bc-transmission': lp_cmdline += ['--bc-data', '--remove-holidays'] elif parsed_args.dataset == 'total-load': lp_cmdline += ['--total-load'] elif parsed_args.dataset == 'gefcom-2012': lp_cmdline += ['--gef-data'] raise RuntimeError( 'GEFCom data is not supported yet, remains to assign a ' 'set of weights for the different temperature zones.') else: raise RuntimeError('Unrecognized dataset: {}'.format(parsed_args.dataset)) parser = lp.prediction_options() parser = lp.data_options(parser) parser = ga.ga_options(parser) lp_options, _ = parser.parse_args(lp_cmdline) lp.options = lp_options return parsed_args, lp_options
def get_options(): global options parser = lp.prediction_options() parser = lp.ga_options(parser) parser = lp.data_options(parser) parser.add_option("--model", dest="model", help="The model class that the genomes instantiate", default=None) parser.add_option("--test-set", dest="test_set", action="store_true", help="Test the genomes on the test set, rather than on the training set", default=False) parser.add_option("--plot", dest="plot", action="store_true", help="Make a plot (in combination with --test-set)", default=False) (options, args) = parser.parse_args() lp.options = options if options.model is None: print >>sys.stderr, "Model argument is required." sys.exit(1)
'--print-pop'] cmdline = cmdline_base + \ ['--remove-holidays', '--data-seed=0', '--bc-data', '--standardize', '--difference=1', '--subtract-weekly-pattern'] cmdline = cmdline_base + \ ['--data-seed=15', '--gef-data', '--subtract-weekly-pattern'] parser = lp.prediction_options() parser = lp.ga_options(parser) parser = lp.data_options(parser) options, _ = parser.parse_args(cmdline) lp.options = options def reseed(seed=None): if seed is None: seed = options.seed random.seed(seed) np.random.seed(seed) reseed() def new_model():