cotegen 1.1

KAIST CS453 2019 Spring Team 7 Project

Coding Test Input Generator

CoTeGen automatically generates test suite of a typical coding test problem from the bottom. First, an initial test suite can be constructed with minimal cost. Next, search-based input generation with mutation fitness complement the initial test suite. Using mutated reference solution as a simplified form of wrong answers, we aim to kill as many mutants as possible. Hence, mutation coverage is provided as a complementary result.

• PyPI Release: https://pypi.org/project/cotegen/

• Academic Report: Dropbox

Adding New Reference Solution

After installing our cotegen package, You can write a new reference solution as a simple python file to generate corresponding test cases.

Please locate your reference solution file in examples/tasks directory to utilize command line execution.

You should define a new class inherits cotegen.Task and with the same with filename.

input_parameters, solve(reference solution), compare, and convert_input_to_string are essential methods to override.

Please refer the example:

import cotegen

# http://codeforces.com/problemset/problem/1/A

class CF1A(cotegen.Task):
    input_parameters = \
        {
            'n': cotegen.types.Integer(1, 10**9),
            'm': cotegen.types.Integer(1, 10**9),
            'a': cotegen.types.Integer(1, 10**9),
        }

    def solve(n: int, m: int, a: int) -> int:
        # 1 <= n, m, a <= 10**9
        return ((n + (a-1)) // a) * ((m + (a-1)) // a)

    @staticmethod
    def compare(user_answer: int, jury_answer: int) -> bool:
        return user_answer == jury_answer

    @staticmethod
    def convert_input_to_string(n: int, m: int, a: int):
        return "%d %d %d\n" % (n, m, a)


if __name__ == '__main__':
    import os
    CF1A.generate_test_files(os.path.expanduser("~/Downloads/CS453/CF1A"))

Define Input Parameters

Predefine the types of input parameters and the range of each argument as a dictionary format.

Currently, CoTeGen supports following types:

cotegen.types.Integer
cotegen.types.IntegerSequence
cotegen.types.NonIncreasingIntegerSequence
cotegen.types.FixedVariableLengthIntegerSequence
cotegen.types.IntegerPermutation

Define constraint

If needed, define the constraints among input parameters as list.

constraints = [
        cotegen.constraints.Leq('M', 'N')
    ]

You can utilize these predefined constraints:

cotegen.constraints.Leq (Less than or equal)
cotegen.constraints.ListLengthLeqInteger
cotegen.constraints.ListLengthReqInteger

or, use custom constraints giving custom function (which receives parameter dictionary and returns true/false) as an argument

cotegen.constraints.CustomConstraint(lambda test: 'd1' not in test or 'd2' not in test or test['d1'] != test['d2'])

Command Line Usage

python -m cotegen run random --filename <filename>

python -m cotegen run mutation --filename <filename> --show-survived <true or false>

python -m cotegen run kill --filename <filename> --save <directory path> --mutation-fitness <true or false>

For example:

python -m cotegen run random --filename CF158A

python -m cotegen run mutation --filename CF158A

python -m cotegen run kill --filename CF158A --save ~/Downloads --mutation-fitness true