main.py

# Markov decision process
from iofile.inputs import load_network, load_activity
from iofile.outputs import export_multi_run_data
from loading.assign import find_fixed_point
# from viewer.plot import draw_zone_population
from allocating.generators import gen_solo_activity_util, gen_path_set, find_shortest_path
from allocating.creators import set_corr
from stats.timer import print_current_time

# TODO: replace the dictionary with numpy.array
# TODO: use integers to index multidimension arrays


def run_multi_scenarios(case_name, corr_list):
    # try distinct corrs
    for corr in corr_list:
        print "       data set: %s" % case_name
        print "    correlation: %.2f" % corr

        # correlations between household members
        set_corr(1, 2, corr)

        # run the iterative procedure
        find_fixed_point(2, case_name, corr)

        # output the raw results
        # export_data(case_name+'_r'+str(corr))

        # generate visual results
        # draw_zone_population(4)


def main():
    # initialize the timer
    print_current_time()

    # the data set name
    case_name = '6node'
    # load activity data
    load_activity(case_name)
    # load network data
    load_network(case_name)

    # generate path sets
    gen_path_set()
    find_shortest_path()
    # generate utils
    gen_solo_activity_util()

    print '\n DATA LOADED'
    print_current_time()

    # run multiple scenarios
    corr_list = [corr / 10.0 for corr in range(10, -1, -5)]
    run_multi_scenarios(case_name, corr_list)

    # export multi-run data
    export_multi_run_data(case_name)


if __name__ == '__main__':
    # import cProfile
    # cProfile.run('main()', 'main_profile.log')
    main()