worst_list = [True]
if options.do_best_worst:
worst_list = [False, True]
start_description = "@@ Starting " + sys.argv[0] + " at " + time.ctime() + " with args " + str(sys.argv[1:])
print start_description
print "base file", base_file
print "algo keys", algo_keys
print "worst list", worst_list
logging.info(start_description)
global base_relation
global base_comments
base_relation, base_comments, base_attrs, base_data = arff.readArff(base_file)
base_data.sort()
print "base_data", len(base_data), len(base_data[0])
print "classifer algorithms used:", algo_keys
def doOneRun():
print "-------------------- START -------------------- " + time.ctime()
print "input file:", base_file
print "test fraction:", options.test_fraction
print "algorithms:", getAlgoDictKeys()
print "class index:", class_index
print "do worst:", getDoWorst()
logging.info("-------------------- START -------------------- " + time.ctime())
logging.info("input file: " + base_file)
logging.info("test fraction: " + str(options.test_fraction))
# Set random seed so that each run gives same results
random.seed(555)
global output_dir
parser = optparse.OptionParser('usage: python ' + sys.argv[0] + ' [options] <input file>')
parser.add_option('-o', '--output', dest='output_dir', default='output', help='output directory')
# parser.add_option('-d', '--dict', action='store_true', dest='dict_only', default=False, help='use dictionaray look-up only')
#parser.add_option('-v', '--verbose', action='store_true', dest='verbose', default=False, help='show details in output')
(options, args) = parser.parse_args()
if len(args) < 1:
print parser.usage
print 'options:', options
print 'args', args
exit()
filename = args[0]
output_dir = options.output_dir
misc.mkDir(output_dir)
relation, comments, attributes, data = arff.readArff(filename)
print 'Algorithms to test:', WC.all_algo_keys
for algo_key in WC.all_algo_keys:
print '======================= findBestAttributes:', filename, algo_key
findBestAttributes(filename, algo_key, data, attributes)
if __name__ == '__main__':
if len(sys.argv) < 3:
print 'Usage: jython get_attribute_subset.py <base-arff-file> <attrs-arff-file>'
sys.exit()
base_filename = sys.argv[1]
attrs_filename = sys.argv[2]
out_filename = os.path.splitext(base_filename)[0] + '.attr_subset' + os.path.splitext(base_filename)[1]
print base_filename
print attrs_filename
print out_filename
relation, comments, attributes, data = arff.readArff(base_filename)
_, _, attributes_subset, _ = arff.readArff(attrs_filename)
attribute_index_map = {}
for i,a in enumerate(attributes):
attribute_index_map[a['name']] = i
names_subset = [a['name'] for a in attributes_subset]
indexes_subset = []
for name in attribute_index_map.keys():
if name in names_subset:
indexes_subset.append(attribute_index_map[name])
out_attributes = [attributes[i] for i in indexes_subset]
out_data = [[d[i] for i in indexes_subset] for d in data]
