Assignment #3: Fix locator

[1] C. Le Goues, M. Dewey-Vogt, S. Forrest, and W. Weimer, “A systematic study of automated program repair: Fixing 55 out of 105 bugs for $8 each,” in 2012 34th International Conference on Software Engineering (ICSE), Jun. 2012, pp. 3–13. doi: 10.1109/ICSE.2012.6227211.

Your submission must satisfy the following requirements:

• R1. Shall initialize your assignment repository from https://classroom.github.com/a/hN7Qn-Fr
• R2. Write your StatRepair.py in the repository.
• R3. Test your StatRepair.py by using pytest.
• R4. You need to let your TA know your repository URL and your student ID together.
• R5. StatRepair class should be defined in the StatRepair.py
• R6. The above class is tested as:

from StatRepair import StatRepair
from urlparse import urlparse

def test_urlparse():
    debugger = StatRepair()

    with debugger.collect_pass():
        urlparse("http://aaaa.com")

    with debugger.collect_pass():
        urlparse("http://aaaa.com#aaa#bbb")

    with debugger.collect_pass():
        urlparse("https://[2001:db8:85a3:8d3:1319:8a2e:370:7348]:443/")

    with debugger.collect_fail():
        urlparse("http://aaaa.com#aaa#bbb;;;")

    line, dist = debugger.mostsimilarstmt(debugger.rank()[0])

    assert ...

• R7. This assignment assumes that you already have a RankingDebugger, TarantulaDebugger, or OchiaiDebugger class composed from debuggingbook.org.
• R8. You have to add line, dist = mostsimilarstmt(targetloc) to your debugger class, where:
  • targetloc -- A pair of (function name, linu number). A suspicious location in the target program. Usually, this is designated by debugger.rank()[x].
  • line -- A string representing the most similar line with the statement at the targetloc. This should be a stripped string.
  • dist -- An integer value representing the Levenshtein distance between the statement at the targetloc and line.
• R9. The return value of mostsimilarstmt(targetloc) shall avoid statements with dist == 0 as the statements identical with the buggy statement are useless.
• R9.1. The distance should be computed only for the statements executed by at least one test case (passing or failing). If no test case executes a statement, you should ignore it.
• R10. The Levenshtein distance shall be computed after stripping the statements.
• R11. Installing new packages is not allowed. Shall assume that all packages are installed for debuggingbook.
• R12. Use this suspiciousness() function:

def suspiciousness(self, event: Any) -> Optional[float]:
        failed = len(self.collectors_with_event(event, self.FAIL))
        not_in_failed = len(self.collectors_without_event(event, self.FAIL))
        not_in_passed = len(self.collectors_without_event(event, self.PASS))
        passed = len(self.collectors_with_event(event, self.PASS))

        try:
            # return failed / math.sqrt((failed + not_in_failed) * (failed + passed))
            return 2*(failed + (not_in_passed / (passed + not_in_passed)))
        except ZeroDivisionError:
            return None

Note:

• N1. pytest (based on test_urlparse1.py) is just for validating your program. The final grading will be made by other test cases.
• N2. Submissions via GitHub Classroom will only be accepted. Submissions via LMS or any other means are not accepted.
• N3. DO NOT change files in this repository except for StatRepair.py. Adding new files are allowed.
• N4. Extra points (20%) are given if submitted within 4:15pm.
• N5. Late submissions after 4:15pm are allowed, but 20% of points will be discounted for every 30min. Thus, no point is given 150min after 4:15pm.