def evaluate()

in scripts/regressor_finder.py [0:0]

• 74 lines of code
• 27 McCabe index (conditional complexity)

def evaluate(bug_introducing_commits):
    logger.info("Downloading commits database...")
    assert db.download(repository.COMMITS_DB)

    logger.info("Downloading bugs database...")
    assert db.download(bugzilla.BUGS_DB)

    logger.info("Building bug -> commits map...")
    bug_to_commits_map = defaultdict(list)
    for commit in tqdm(repository.get_commits()):
        bug_to_commits_map[commit["bug_id"]].append(commit["node"])

    logger.info("Loading known regressors using regressed-by information...")
    known_regressors = {}
    for bug in tqdm(bugzilla.get_bugs()):
        if bug["regressed_by"]:
            known_regressors[bug["id"]] = bug["regressed_by"]
    logger.info("Loaded %d known regressors", len(known_regressors))

    fix_to_regressors_map = defaultdict(list)
    for bug_introducing_commit in bug_introducing_commits:
        if not bug_introducing_commit["bug_introducing_rev"]:
            continue

        fix_to_regressors_map[bug_introducing_commit["bug_fixing_rev"]].append(
            bug_introducing_commit["bug_introducing_rev"]
        )

    logger.info("%d fixes linked to regressors", len(fix_to_regressors_map))
    logger.info(
        "%d regressors linked to fixes",
        sum(len(regressors) for regressors in fix_to_regressors_map.values()),
    )

    logger.info("Measuring how many known regressors SZZ was able to find correctly...")
    all_regressors = 0
    perfect_regressors = 0
    found_regressors = 0
    misassigned_regressors = 0
    for bug_id, regressor_bugs in tqdm(known_regressors.items()):
        # Get all commits which fixed the bug.
        fix_commits = bug_to_commits_map[bug_id] if bug_id in bug_to_commits_map else []
        if len(fix_commits) == 0:
            continue

        # Skip bug/regressor when we didn't analyze the commits to fix the bug (as
        # certainly we can't have found the regressor in this case).
        if not any(fix_commit in fix_to_regressors_map for fix_commit in fix_commits):
            continue

        # Get all commits linked to the regressor bug.
        regressor_commits = []
        for regressor_bug in regressor_bugs:
            if regressor_bug not in bug_to_commits_map:
                continue

            regressor_commits += (
                commit for commit in bug_to_commits_map[regressor_bug]
            )

        if len(regressor_commits) == 0:
            continue

        found_good = False
        found_bad = False
        for fix_commit in fix_commits:
            # Check if we found at least a correct regressor.
            if any(
                regressor_commit in regressor_commits
                for regressor_commit in fix_to_regressors_map[fix_commit]
            ):
                found_good = True

            # Check if we found at least a wrong regressor.
            if any(
                regressor_commit not in regressor_commits
                for regressor_commit in fix_to_regressors_map[fix_commit]
            ):
                found_bad = True

        all_regressors += 1

        if found_good and not found_bad:
            perfect_regressors += 1
        if found_good:
            found_regressors += 1
        if found_bad:
            misassigned_regressors += 1

    logger.info(
        "Perfectly found %d regressors out of %d", perfect_regressors, all_regressors
    )
    logger.info("Found %d regressors out of %d", found_regressors, all_regressors)
    logger.info(
        "Misassigned %d regressors out of %d", misassigned_regressors, all_regressors
    )