Homework Marking

A python code to judge and mark python homework automatically.

Requirements

• Python >= 3.10
• VSCode

Please use uv sync or install the python requirements (in pyproject.toml) yourself :]

Usage

Create 2 directories submissions and testcase, and put the zips of the students' codes and the testcases in them respectively, for example

homework1/
    submissions.zip/  # or `submissions` dir
        123zhangsan_useless_suffix.zip
            question1.py
            question2.py
        456lisi_456_LISI.zip
            question1.py
            question2.py
        789wangwu_wangwu-789-homework1.zip
            question1.py
            question2.py
    testcase/
        question1/
            1.in
            1.out
            2.in
            2.out
        question2/
            1.py  # inner data: assert(Func(1) == 2)
            2.py  # inner data: assert(Func(2) == 3)

Extract

Use

uv run extract_all.py
# python extract_all.py

to extract the zips/rars in homework1/submissions.zip (or submissions/) to homework1/extract/.

As shown above, the name of the .zip/.rar file must be <id><name>_<any other suffix>, or you can modify the name split rules in lib/pattern.py.

Some students' zip files may contain invalid contents. To inspect and resolve the potential errors, see homework1/error.txt for more information.

Mark

After resolving the errors, use

uv run mark_all.py
# python mark_all.py

to mark scores. Just follow the prompt, and the result will be saved to an excel file named homework1/result.xlsx.

Moreover, In mark_all.py, you can stop at any time you want (Ctrl + C), and start from any index you want. The result data of students will only be saved at the end of each marking process (prompt is OK / Retry (o -> OK / r -> Retry):).

Differentiate (Duplication Check)

Use

uv run diff_all.py
# python diff_all.py

to check the codes' duplication rate. If you have some other codes as benchmark to check the student's codes, also put them in the more_diff/ directory as shown below.

more_diff/
    any_name_you_want/
        homework1/
            question1.py
            question2.py
        homework2/
            question1.py
            question2.py
    benchmark_2/
        homework2/
            question2.py

Testcase

.in and .out are like

1
2
3

6

which will just be the input and output of students' .py files. Judging code will detect whether the output is the same (or other detect way, see JudgeProject.__judge_text() code in lib/judge.py).

.py is like

print(getPrimeNum(3))
assert(getPrimeNum(3) == 5)

which will append to students' .py files and running. Judging code only marks it as false when exception occurred (which means assert() failed here), and it will output both the exception (in its stderr) and the normal output (in its stdout) to your screen.