Get number of Ceph teuthology test cases over time

teuthology-stats.py

#/usr/bin/python3

"""
Get statistics out of teuthology
This needs a ceph repo git checkout.

Author: Thomas Bechtold <tbechtold@suse.com>
License: Apache-2.0
"""

import argparse
import subprocess
import os
import contextlib
import json
import csv


@contextlib.contextmanager
def remember_cwd():
    curdir = os.getcwd()
    try:
        yield
    finally:
        os.chdir(curdir)


def _git_ref_before_date(year, month):
    cmd = [
        '/usr/bin/git',
        'rev-list', '-n', '1',
        '--first-parent',
        '--before="{}-{:02d}-01 00:00"'.format(year, month),
        'master'
    ]
    return subprocess.check_output(cmd, encoding='utf-8').strip()


def _git_checkout(rev):
    cmd = [
        '/usr/bin/git',
        'checkout', '--force',
        '{}'.format(rev)
    ]
    return subprocess.check_output(cmd, encoding='utf-8').strip()


def _teuthology_get_tests_from_suite(suite_dir):
    cmd = [
        'teuthology-describe-tests',
        suite_dir,
        '-c',
        '--format', 'json',
    ]

    print('Calling now: ' + ' '.join(cmd))
    data = json.loads(subprocess.check_output(cmd, encoding='utf-8'))
    return data['data']


def parse_args():
    """parse command line arguments"""
    parser = argparse.ArgumentParser()
    parser.add_argument('--ceph-repo', default='/home/tom/devel/ceph/ceph/',
                        help='Path to a ceph git repo (default: %(default)s)')
    parser.add_argument(
        '--suites', default=['big', 'buildpackages', 'cephmetrics', 'experimental',
                             'fs', 'hadoop', 'kcephfs', 'krbd', 'marginal', 'mixed-clients',
                             'multimds', 'perf-basic', 'powercycle', 'rados', 'rbd', 'rgw',
                             'samba', 'smoke', 'stress', 'teuthology', 'tgt', 'upgrade'],
        nargs='*',
        help='Path to a ceph git repo (default: %(default)s)')
    parser.add_argument('--years', default=[2018, 2019], nargs='*',
                        help='Years to get data for (default: %(default)s)')
    return parser.parse_args()


def main():
    args = parse_args()
    print(args)

    if not os.path.isdir(args.ceph_repo):
        raise Exception('ceph repo dir "{}" does not exist'.format(args.ceph_repo))

    with remember_cwd():
        os.chdir(args.ceph_repo)

        data = {}

        for year in args.years:
            for month in range(0, 6):
                month = month * 2 + 1
                ym = '{}-{}'.format(year, month)
                data[ym] = {}
                ref = _git_ref_before_date(year, month)
                _git_checkout(ref)
                print('git checkout of ref {} done ({})'.format(ref, ym))
                for suite in args.suites:
                    suite_dir = os.path.join(args.ceph_repo, 'qa', 'suites', suite)
                    if not os.path.exists(suite_dir):
                        print('suite "{}" not found ({})'.format(suite, ym))
                        data[ym][suite] = 0
                    else:
                        print('get suite data for {} ({})'.format(suite, ym))
                        suite_data = _teuthology_get_tests_from_suite(suite_dir)
                        data[ym][suite] = len(suite_data)

        print(data)
        # print the whole data
        with open('data.csv', 'w', newline='') as csvfile:
            fieldnames = ['date', 'suite_name', 'suite_tests']
            writer = csv.DictWriter(csvfile, fieldnames=fieldnames)
            writer.writeheader()

            for date, suite_data in data.items():
                for suite_name, suite_tests in suite_data.items():
                    writer.writerow({'date': date, 'suite_name': suite_name, 'suite_tests': suite_tests})



if __name__ == "__main__":
    main()