args.py

# -*- coding: utf-8 -*-

import argparse

def process_args():
    def percent_value(string):
        val = float(string)
        if 0.0 <= val <=1.0:
            return val
        else:
            raise argparse.ArgumentTypeError(u"{0} is not from the range of 0..1".format(string))

    def algorithm_name(string):
        if string.lower() not in ('apriori', 'dic'):
            raise argparse.ArgumentTypeError(u"No algorithm named {0} is known.".format(string))
        else:
            return string.lower()

    def dic_granularity(string):
        val = int(string)
        #if val < 0:
            #raise argparse.ArgumentTypeError(u"DIC's granularity must be positive or 0 for automatic M calculation.")
        #else:
            #return val
        return val
            
    def boolean(string):
        if string == "False":
            val = False
        elif string == "True":
            val = True
        else:
            int_val = int(string)
            if int_val in (0, 1):
                val = bool(int_val)
            else:
                raise argparse.ArgumentTypeError(u"{0} should be from \"True\" \\ \"False\" or \"0\" \\ \"1\" ".format(string))
        return bool(val)


    parser = argparse.ArgumentParser(
            description=u"Induction of associative rules from datasets.",
            epilog=u"For current version check 'https://github.com/PowerLlamas/associative_rules'."
        )
    parser.add_argument('infile',
            help=u"input file")
    parser.add_argument('outfile',
            nargs='?',
            help=u"output file"
        )
    parser.add_argument('-a, --algorithm',
            dest='algorithm',
            type=algorithm_name,
            default='dic',
            help=u"name of preferred algorithm - apriori or dic"
        )
    parser.add_argument('-r, --randomize',
            dest='randomize',
            type=boolean,
            default='False',
            help=u"Randomize transactions order or not"
        )
    parser.add_argument('-c, --confidence',
            dest='minconf',
            type=percent_value,
            default=0.7,
            help=u"minimum confidence (0..1)"
        )
    parser.add_argument('-s, --support',
            dest="minsup",
            type=percent_value,
            default=0.4,
            help=u"minimum support (0..1)"
        )
    parser.add_argument('-m',
            dest="m",
            type=dic_granularity,
            default=100,
            help=u"DIC's granularity or 0 for automatic (M = transactions number * min support"
        )
    parser.add_argument('-p, --partial',
            dest="partial",
            type=boolean,
            default=False,
            help=u"Switches DIC partial aproach: support is calculated using transactions counted till moment of support calculation."
        )
     
    return parser.parse_args()