Processing command-line arguments: module ''argparse''

A lot of useful tools for computer users have a form of commands/scripts which are run from command line (ls, dir, grep, …). These tools usually accept command-line arguments which follow the command name (e.g., ls -la) and which affect the behavior of the command. Python contains standard module argparse (tutorial) which allows us easilly configure possible parameters of a script and process them.

The script you have to create should behave according to specifications. This can be achieved by configuring the argparse module in the following way:

import argparse
 
def setup_arg_parser():
    parser = argparse.ArgumentParser(description='Learn and classify image data.')
    parser.add_argument('train_path', type=str, help='path to the training data directory')
    parser.add_argument('test_path', type=str, help='path to the testing data directory')
    mutex_group = parser.add_mutually_exclusive_group(required=True)
    mutex_group.add_argument('-k', type=int, 
                             help='run k-NN classifier (if k is 0 the code may decide about proper K by itself')
    mutex_group.add_argument("-b", 
                             help="run Naive Bayes classifier", action="store_true")
    parser.add_argument("-o", metavar='filepath', 
                        default='classification.dsv',
                        help="path (including the filename) of the output .dsv file with the results")
    return parser

When you use this function in a script as follows:

    parser = setup_arg_parser()
    args = parser.parse_args()
    print(args.train_path)

variable args will contain all the information passed to the script via command-line arguments, such that you can work with them easily.

The above code can be found in file classifier.py which can be used as the skeleton of your solution. If you run the downloaded module as

python3 classifier.py -k 3 ./train_data ./test_data 

Training data directory: ./train_data
Testing data directory: ./test_data
Output file: classification.dsv
Running k-NN classifier with k=3

• paths to directories with training and testing data passed on the command line,
• the name of the output file which was not specified on the command line and argparse thus filled in the default value, and
• the chosen classifier type (k-NN) with a numeric argument. (But you have to implement the classifier yourselves. )