Commit ddec66ac authored by Jan Kiene's avatar Jan Kiene
Browse files

Merge branch 'ci/refactor-histogram-creation' into kiene/tmp-branch-for-ltv-split-testing

parents 6379031b a50fb3b8

scripts/create_histograms.py

+7 −0
Original line number Diff line number Diff line
@@ -74,6 +74,8 @@ def create_histograms( 
            ax.set_xlabel(measure)

            if "DIFF" in measure:

                ax.set_xticks(range(len(x)), x, rotation=35)

            else:

                ax.set_xticks(range(len(x)), x)

            ax.set_ylabel("Number of test cases")



            fig.set_figheight(4)
@@ -127,6 +129,11 @@ Use this for visualising diff scores.""", 
    args = parser.parse_args()

    df = pd.read_csv(args.csv_report)



    # filter out missing format/category values

    mask_format_missing = df["format"].isna()

    mask_category_missing = df["category"].isna()

    df = df[~mask_format_missing | ~mask_category_missing]



    bins_for_measures = BINS_FOR_MEASURES

    if args.no_bins:

        bins_for_measures = {}

scripts/parse_xml_report.py

+34 −11
Original line number Diff line number Diff line
@@ -4,12 +4,20 @@ import argparse 
import pandas as pd

from xml.etree import ElementTree

from collections import Counter

from typing import Optional

from enum import Enum





SPLIT_STRING = "_split"

WHOLE_STRING = "_whole"





class Result(str, Enum):

    ERROR = "ERROR"

    FAIL = "FAIL"

    PASS = "PASS"





class TestcaseParser(dict):

    def __init__(self, testcases: list):

        super().__init__()
@@ -32,8 +40,8 @@ class TestcaseParser(dict): 
        # if we already have this testcase, do a sanity check and set result to ERROR

        if fulltestname in self:

            results = [self[fulltestname]["result"], result]

            assert any(r == "ERROR" for r in results)

            self[fulltestname]["result"] = "ERROR"

            assert any(r == Result.ERROR for r in results)

            self[fulltestname]["result"] = Result.ERROR

            return



        ret = {}
@@ -110,24 +118,33 @@ def xml_to_dataframe(xml_report: str) -> pd.DataFrame: 


def get_result_from_testcase(testcase: ElementTree.Element) -> str:

    if testcase.find("failure") is not None:

        testresult = "FAIL"

        testresult = Result.FAIL

    elif testcase.find("error") is not None:

        testresult = "ERROR"

        testresult = Result.ERROR

    else:

        testresult = "PASS"

        testresult = Result.PASS



    return testresult





def main(xml_report, csv_file):

def main(xml_report: str, csv_file: str, split_csv_file: Optional[str]):

    df = xml_to_dataframe(xml_report)

    df.to_csv(csv_file, index=False)



    n_testcases = len(df)

    count = Counter(df["result"])



    if split_csv_file is not None:

        mask_errors = df["result"] == Result.ERROR

        mask_whole = df["split"] == "whole"

        mask_single = mask_errors | mask_whole

        df_split = df[~mask_single]

        df_split.to_csv(split_csv_file, index=False)



        df = df[mask_single]



    df.to_csv(csv_file, index=False)



    print(

        f"Parsed testsuite with {n_testcases} tests: {count['PASS']} passes, {count['FAIL']} failures and {count['ERROR']} errors."

        f"Parsed testsuite with {n_testcases} tests: {count[Result.PASS]} passes, {count[Result.FAIL]} failures and {count[Result.ERROR]} errors."

    )
@@ -140,7 +157,13 @@ if __name__ == "__main__": 
        type=str,

        help="XML junit report input file, e.g. report-junit.xml",

    )

    parser.add_argument("csv_file", type=str, help="Output CSV file, e.g. report.csv")

    parser.add_argument("csv_file", help="Output CSV file, e.g. report.csv")

    parser.add_argument(

        "--split-csv-file",

        type=str,

        default=None,

        help="If given, write the split comparison values to this file separately",

    )



    args = parser.parse_args()

    main(args.xml_report, args.csv_file)
 
    main(args.xml_report, args.csv_file, args.split_csv_file)