JPlag not producing 100% similarity on identical submissions

[AjayvirS]

Describe the bug

When submitting two identical programming exercises, plagiarism detection will never produce a 100% similarity because the template code is considered part of the total tokens by which the tool presumably divides the similarity value.

Screenshots

No response

Which version of Artemis are you seeing the problem on?

9.8.3

What browsers do you see the problem on?

Firefox

Additional context

No response

Relevant log output

No response

To Reproduce

1. Create a programming exercise Sort exercise template, e.g., Java
2. log in as User 1 and submit the following piece of code in BubbleSort.java:

    public static boolean isPalindrome(String input) {
        if (input == null) {
            return false;
        }
        String normalized = input.replaceAll("[^a-zA-Z0-9]", "").toLowerCase();
        return normalized.equals(new StringBuilder(normalized).reverse().toString());
    }

3. Login as User 2 and submit the same piece of code in BubbleSort.java
4. log in as a tutor and navigate to Plagiarism in the exercise dashboard
5. Run plagiarism detection
6. Observe a <100% similarity despite the identical repositories for both submissions

As we increase the number of identical lines, we see that the similarity tends to be 100% but never reaches 100%, which makes me assume that the template code is somehow factored in. I confirmed this with a Python exercise from which I removed the template code and only submitted identical code with both users, producing a 100% similarity.

Submissions with multiple identical lines:

similarity.mp4

Expected behavior

Similarity value should reflect the actual content similarity of the submissions.
Solution approach suggestion: Add a checkbox that does not consider template code as part of the input to JPlag

Screenshots

No response

Which version of Artemis are you seeing the problem on?

9.8.3

What browsers are you seeing the problem on?

Firefox

Additional context

No response

Relevant log output

No response

[krusche]

seems to be a duplicate of #7174
I suggest to continue the discussion in #7174