summarize_json.py

import json
from collections import OrderedDict
import argparse
import sys
import json
from modules import validate_args
import os.path
from operator import itemgetter
import copy


def outputResultsJSON(
    output_filename: str,
    data: list[dict],
    baseline: OrderedDict,
    failure_manifest_filename="detection_failure_manifest.json",
    output_folder: str = "",
    summarization_reproduce_failure_config: dict = {},
) -> tuple[bool, int, int, int, int]:
    success = True

    try:
        test_count = len(data)
        # Passed
        pass_count = len([x for x in data if x["success"] == True])

        # A failure or an error
        fail_count = len([x for x in data if x["success"] == False])

        # An error (every error is also a failure)
        fail_and_error_count = len([x for x in data if x["error"] == True])

        # A failure without an error
        fail_without_error_count = len(
            [x for x in data if x["success"] == False and x["error"] == False]
        )

        # This number should always be zero...
        error_and_success_count = len(
            [x for x in data if x["success"] == True and x["error"] == True]
        )
        if error_and_success_count > 0:
            print(
                "Error - a test was successful, but also included an error. This should be impossible.",
                file=sys.stderr,
            )
            success = False

        if test_count != (pass_count + fail_count):
            print(
                "Error - the total tests [%d] does not equal the pass[%d]/fails[%d]"
                % (test_count, pass_count, fail_count)
            )
            success = False

        if fail_count > 0:
            result = "FAIL for %d detections" % (fail_count)
            success = False
        else:
            result = "PASS for all %d detections" % (pass_count)

        summary = {
            "TOTAL_TESTS": test_count,
            "TESTS_PASSED": pass_count,
            "TOTAL_FAILURES": fail_count,
            "FAIL_ONLY": fail_without_error_count,
            "PASS_RATE": calculate_pass_rate(pass_count, test_count),
            "FAIL_AND_ERROR": fail_and_error_count,
        }

        data_sorted = sorted(
            data, key=lambda k: (-k["error"], k["success"], k["detection_file"])
        )
        with open(os.path.join(output_folder, output_filename), "w") as jsonFile:
            json.dump(
                {"summary": summary, "baseline": baseline, "results": data_sorted},
                jsonFile,
                indent="    ",
            )

        # Generate a failure that the user can download to reproduce and test ONLY the failures locally.
        # This makes it easy to test and debug ONLY those that failed.  No need to test the ones
        # that succeeded!

        fail_list = [
            os.path.join("security_content/detections", x["detection_file"])
            for x in data_sorted
            if x["success"] == False
        ]

        if len(fail_list) > 0:
            print("FAILURES:")
            for failed_test in fail_list:
                print(f"\t{failed_test}")
            failures_test_override = copy.deepcopy(
                summarization_reproduce_failure_config
            )
            # Force all tests to be interactive, even if they don't fail (because they failed on this test)
            failures_test_override.update(
                {
                    "detections_list": fail_list,
                    "no_interactive_failure": False,
                    "interactive": True,
                    "num_containers": 1,
                    "branch": baseline["branch"],
                    "commit_hash": baseline["commit_hash"],
                    "mode": "selected",
                    "show_splunk_app_password": True,
                }
            )
            with open(
                os.path.join(output_folder, failure_manifest_filename), "w"
            ) as failures:
                validate_args.validate_and_write(failures_test_override, failures)
    except Exception as e:
        print(
            "There was an error generating [%s]: [%s]" % (output_filename, str(e)),
            file=sys.stderr,
        )
        print(data)
        raise (e)
        # success = False
        # return success, False

    # note that total failures is fail_count, fail_and_error count is JUST errors (and every error is also a failure)
    return success, test_count, pass_count, fail_count, fail_and_error_count


def calculate_pass_rate(pass_count: int, test_count: int) -> float:
    if test_count == 0:
        # Assume this means 100% pass rate to avoid divide by zero
        pass_rate = 1
    else:
        pass_rate = pass_count / test_count
    return pass_rate


def print_summary(
    test_count: int, pass_count: int, fail_count: int, error_count: int
) -> None:
    print(
        "Summary:"
        f"\n\tTotal Tests: {test_count}"
        f"\n\tTotal Pass : {pass_count}"
        f"\n\tTotal Fail : {fail_count} ({error_count} of these were ERRORS))"
        f"\n\tPass  Rate : {calculate_pass_rate(pass_count, test_count):.3f}"
    )


def exit_with_status(
    test_pass: bool, test_count: int, pass_count: int, fail_count: int, error_count: int
) -> None:
    if not test_pass:
        print("Result: FAIL")
        # print("DURING TESTING, THIS WILL STILL EXIT WITH AN EXIT CODE OF 0 (SUCCESS) TO ALLOW THE WORKFLOW "
        #      "TO PASS AND CI/CD TO CONTINUE.  THIS WILL BE CHANGED IN A FUTURE VERSION.")
        # sys.exit(0)
        sys.exit(1)
    else:
        print("Result: PASS!")
        sys.exit(0)


def finish(
    test_pass: bool, test_count: int, pass_count: int, fail_count: int, error_count: int
) -> None:
    print_summary(test_count, pass_count, fail_count, error_count)
    exit_with_status(test_pass, test_count, pass_count, fail_count, error_count)


def main():
    parser = argparse.ArgumentParser(description="Results Merger")
    parser.add_argument(
        "-f",
        "--files",
        type=argparse.FileType("r"),
        required=True,
        nargs="+",
        help="The json files you would like to combine into a single file",
    )
    parser.add_argument(
        "-o",
        "--output_filename",
        type=str,
        required=True,
        help="The name of the output file",
    )
    parser.add_argument("--smoketest", action=argparse.BooleanOptionalAction)
    args = parser.parse_args()

    all_data = OrderedDict()
    try:
        print("We will summarize the files: %s" % (str([f.name for f in args.files])))
        for f in args.files:
            if not f.name.endswith(".json"):
                print(
                    "Error: passed in file must end in .json - you passed in [%s].\n\tQuitting..."
                    % (f.name)
                )
                sys.exit(1)
            data = json.loads(f.read())
            if "baseline" in all_data:
                # everything has the same baseline, only need to do it once
                pass
            else:
                all_data["baseline"] = data["baseline"]
            if "results" in all_data:
                # this is a list of dictionaries, so add to it
                all_data["results"].extend(data["results"])
            else:
                all_data["results"] = data["results"]

        IGNORE_MESSAGES = [
            "Model does not exist",  # model not generated by baseline
            "Data model 'Identity_Management' was not found",  # missing datamodel included with es
            "get_asset",  # missing macro included with es
            "Failed to load model",  # when running a model that has not been downloaded separately
            "UEBA",  # Another missing ES asset
        ]
        if args.smoketest:
            new_results = []
            for result in all_data["results"]:
                if result.get("detection_error", None):
                    message = result.get("detection_error", None)
                    ignore = False
                    for ignore_message in IGNORE_MESSAGES:
                        if ignore_message in message:
                            ignore = True
                            break
                    if not ignore:
                        new_results.append(result)
                        pass

            all_data["results"] = new_results

        test_pass, test_count, pass_count, fail_count, error_count = outputResultsJSON(
            args.output_filename, all_data["results"], all_data["baseline"]
        )
        finish(test_pass, test_count, pass_count, fail_count, error_count)

    except Exception as e:
        print("Error generating the summary file: [%s].\n\tQuitting..." % (str(e)))
        sys.exit(1)


if __name__ == "__main__":
    main()