evaluate.py

from argparse import ArgumentParser

def evaluate_results(test_file_path, correct_test_file_path):
    total_words = 0
    correct_marked_tags = 0
    is_mismatch = False
    with open(test_file_path, "r") as tf, open(correct_test_file_path, "r") as cf:
        for tsent, csent in zip(tf, cf):
            tsent = tsent.strip(" \n")
            csent = csent.strip(" \n")
            for tword, cword in zip(tsent.split(" "), csent.split(" ")):
                ttoken, ttag = tword.split("/")
                ctoken, ctag = cword.split("/")
                if ttoken == ctoken :
                    if ttag == ctag:
                        correct_marked_tags+=1
                    total_words+=1
                else:
                    is_mismatch = True
                    break
            if is_mismatch:
                print("Refactor Mismatch in test and reference file!")
                break
    print("{0} words tag correctly Out of {1}".format(correct_marked_tags, total_words))    
    print("Accuracy: {}".format(round(correct_marked_tags/total_words, 2)))

def get_args():
    parser = ArgumentParser()
    parser.add_argument('-t', '--test_output', required=True, help='Test Output file contains tagged sentences generated by Model')
    parser.add_argument('-c', '--correct_output', required=True,
            help='Test Output file contains correct tagged sentence') 
    return parser.parse_args()

def main():
    args = get_args()
    evaluate_results(args.test_output, args.correct_output)

if __name__ == '__main__':
    main()