def bc_rel_aspect_fcns_gen(stats, metric):
value_one = stats['data'][sorted(stats['group'])[0]][metric]['lowest']
aspect_fcns = {
'bc_rel': (lambda g, d, m: divide_def0(value_one, float(d[m]['lowest'])
) / float(g))
}
return aspect_fcns
def ft_cost_fcn(g, d, m1, m2, graph):
'''
m1 is the metric we use for comparison; m2 is the opponent.
'''
m1_opt_combo = d[m1]['lowest_combo']
m2_opt_combo = d[m2]['lowest_combo']
apsp = nx.all_pairs_dijkstra_path_length(graph)
apsp_paths = nx.all_pairs_dijkstra_path(graph)
weighted = True
extra_params = None
m1_opt_m1_value = d[m1]['lowest']
m2_opt_m1_value = METRIC_FCNS[m1](graph, m2_opt_combo, apsp, apsp_paths, weighted, extra_params)
print "metric values for %s" % (m1)
print " value for combo %s opt for %s: %s" % (m1_opt_combo, m1, m1_opt_m1_value)
print " value for combo %s opt for %s: %s" % (m2_opt_combo, m2, m2_opt_m1_value)
value = divide_def0(m2_opt_m1_value, m1_opt_m1_value)
print " return value of: %s for k = %s" % (value, g)
return value
def ft_cost_fcn(g, d, m1, m2, graph):
'''
m1 is the metric we use for comparison; m2 is the opponent.
'''
m1_opt_combo = d[m1]['lowest_combo']
m2_opt_combo = d[m2]['lowest_combo']
apsp = nx.all_pairs_dijkstra_path_length(graph)
apsp_paths = nx.all_pairs_dijkstra_path(graph)
weighted = True
extra_params = None
m1_opt_m1_value = d[m1]['lowest']
m2_opt_m1_value = METRIC_FCNS[m1](graph, m2_opt_combo, apsp, apsp_paths,
weighted, extra_params)
print "metric values for %s" % (m1)
print " value for combo %s opt for %s: %s" % (m1_opt_combo, m1,
m1_opt_m1_value)
print " value for combo %s opt for %s: %s" % (m2_opt_combo, m2,
m2_opt_m1_value)
value = divide_def0(m2_opt_m1_value, m1_opt_m1_value)
print " return value of: %s for k = %s" % (value, g)
return value
'miles_cost': 'rx'
},
'aspect_fcns': {
'miles_cost': (lambda g, d, m: d[m]['lowest'] / float(g))
},
'ylabel': (lambda m: metric_fullname(m) + "\nmiles over cost")
},
'ratios': {
'aspect_colors': { #'highest': 'rx',
'mean': 'bo',
'one': 'g+'
},
'aspect_fcns':
{ #'highest': (lambda g, d, m: divide_def0(d[m]['highest'], d[m]['lowest'])),
'mean':
(lambda g, d, m: divide_def0(d[m]['mean'], d[m]['lowest'])),
'one': (lambda g, d, m: 1.0)
},
'ylabel': (lambda m: metric_fullname(m) + "/optimal"),
'max_x': (lambda o: 12.0),
'min_y': (lambda o: 1.0),
'max_y': (lambda o: 2.5)
},
'durations': {
'aspect_colors': {
'duration': 'rx'
},
'aspect_fcns': {
'duration': (lambda g, d, m: d[m]['duration'])
},
'ylabel': (lambda m: metric_fullname(m) + "duration (sec)")
'aspect_fcns': {
'lowest': (lambda g, d, m: d[m]['lowest'])
},
'aspect_colors': {
'lowest': 'g+'
},
'get_data_fcns': ranges_get_data_fcns
},
'ratios_all': {
'aspect_colors': {
'highest': 'rx',
'mean': 'bo',
'one': 'g+'
},
'aspect_fcns': {
'highest': (lambda g, d, m: divide_def0(d[m]['highest'], d[m]['lowest'])),
'mean': (lambda g, d, m: divide_def0(d[m]['mean'], d[m]['lowest'])),
'one': (lambda g, d, m: 1.0)
},
'ylabel': (lambda m: metric_fullname(m) + "/optimal"),
'min_y': (lambda o: 0.8),
'max_y': (lambda o: 6.0),
'get_data_fcns': shared_get_data_fcns
},
'ratios_mean': {
'aspect_colors': {
'mean': 'b*',
'one': 'g+'
},
'aspect_fcns': {
'mean': (lambda g, d, m: divide_def0(d[m]['mean'], d[m]['lowest'])),
def bc_rel_aspect_fcns_gen(stats, metric):
value_one = stats['data'][sorted(stats['group'])[0]][metric]['lowest']
aspect_fcns = {'bc_rel': (lambda g, d, m: divide_def0(value_one, float(d[m]['lowest'])) / float(g))}
return aspect_fcns