test_forensics.py

"""Self-check for the static forensics layer. Run: python3 test_forensics.py"""

from forensics import forensic_signals, static_features

TYPED_BY_HAND = """\
n = int(input())
a = list(map(int, input().split()))
best = 0
for i in range(n):
    s = 0
    for j in range(i, n):
        s += a[j]
        if s > best: best = s
print(best)
"""

PASTED_FROM_CHAT = """\
def max_subarray_sum(arr: list) -> int:
    \"\"\"Return the maximum subarray sum using Kadane’s algorithm.\"\"\"
    # Step 1: initialize the running and best totals
    current_sum = 0
    best_sum = arr[0]
    # Step 2: iterate through the array
    for value in arr:
        # Step 3: extend the window or start a new one
        current_sum = max(value, current_sum + value)
        # Step 4: update the best total seen so far
        best_sum = max(best_sum, current_sum)
    return best_sum
"""


def run():
    typed = static_features(TYPED_BY_HAND)
    pasted = static_features(PASTED_FROM_CHAT)

    assert typed["smart_punct"] == 0
    assert pasted["smart_punct"] == 1, "curly apostrophe should be caught"
    assert pasted["ai_style_comments"] >= 3, pasted["ai_style_comments"]
    assert typed["ai_style_comments"] == 0
    assert typed["single_char_ratio"] > pasted["single_char_ratio"]

    typed_score, typed_signals = forensic_signals(typed)
    pasted_score, pasted_signals = forensic_signals(pasted)
    assert typed_score == 0, typed_signals
    assert pasted_score >= 60, (pasted_score, pasted_signals)
    assert {s["code"] for s in pasted_signals} >= {
        "unicode_punctuation",
        "ai_narration_comments",
    }

    # Indentation forensics: a snippet that mixes tabs and spaces was assembled.
    mixed = static_features("def f():\n\tx = 1\n    y = 2\n    return x\n")
    assert mixed["mixed_indent"] is True
    assert "mixed_indentation" in {s["code"] for s in forensic_signals(mixed)[1]}

    # Empty input must not divide by zero.
    assert static_features("")["lines"] == 0

    print("forensics self-check OK")


if __name__ == "__main__":
    run()