def main():
    parser = ArgumentParser()
    parser.add_argument('output', help='ZipNum Cluster Output directory')
    parser.add_argument('inputs', nargs='+', help='CDX Input glob eg: /cdx/*.cdx.gz')
    parser.add_argument('-s', '--shards', default=10, type=int,
                        help='Number of ZipNum Cluster shards to create')

    parser.add_argument('-l', '--numlines', default=3000, type=int,
                        help='Number of lines per gzip block (default 3000)')

    parser.add_argument('-p', '--parallel', action='store_true',
                        help='Run in parllel (multiple maps/reducer processes)')

    r = parser.parse_args()

    MRJobLauncher.set_up_logging(quiet=False,
                                 verbose=False,
                                 stream=sys.stderr)

    log.setLevel(logging.INFO)
    compat_log = logging.getLogger('mrjob.compat')
    compat_log.setLevel(logging.ERROR)

    run_job(r.inputs, r.output, r.shards, r.parallel, r.numlines)
    build_summary_and_loc(r.output)

 def test_hadoop_runner(self):
     # you can't instantiate a HadoopJobRunner without Hadoop installed
     launcher = MRJobLauncher(args=["--no-conf", "-r", "hadoop", "", "--hadoop-streaming-jar", "HUNNY"])
     with no_handlers_for_logger("mrjob.runner"):
         with patch.dict(os.environ, {"HADOOP_HOME": "100-Acre Wood"}):
             with launcher.make_runner() as runner:
                 self.assertIsInstance(runner, HadoopJobRunner)

 def test_hadoop_runner(self):
     # you can't instantiate a HadoopJobRunner without Hadoop installed
     launcher = MRJobLauncher(args=['--no-conf', '-r', 'hadoop', '',
                                    '--hadoop-streaming-jar', 'HUNNY'])
     with no_handlers_for_logger('mrjob.runner'):
         with patch.dict(os.environ, {'HADOOP_HOME': '100-Acre Wood'}):
             with launcher.make_runner() as runner:
                 self.assertIsInstance(runner, HadoopJobRunner)

 def test_no_output(self):
     launcher = MRJobLauncher(args=['--no-conf', '--no-output', ''])
     launcher.sandbox()
     with patch.object(launcher, 'make_runner') as m_make_runner:
         runner = RunnerStub()
         _mock_context_mgr(m_make_runner, runner)
         runner.stream_output.return_value = ['a line']
         launcher.run_job()
         self.assertEqual(launcher.stdout.getvalue(), '')
         self.assertEqual(launcher.stderr.getvalue(), '')

 def test_no_output(self):
     launcher = MRJobLauncher(args=["--no-conf", "--no-output", ""])
     launcher.sandbox()
     with patch.object(launcher, "make_runner") as m_make_runner:
         runner = Mock()
         _mock_context_mgr(m_make_runner, runner)
         runner.stream_output.return_value = ["a line"]
         launcher.run_job()
         self.assertEqual(launcher.stdout.getvalue(), "")
         self.assertEqual(launcher.stderr.getvalue(), "")

    def test_no_file_args_required(self):
        words1 = self.makefile('words1', b'kit and caboodle\n')
        words2 = self.makefile('words2', b'baubles\nbangles and beads\n')

        job = MRJobLauncher(
            args=['-r', 'local', tests.sr_wc.__file__, words1, words2])
        job.sandbox()

        with job.make_runner() as runner:
            runner.run()

            lines = list(to_lines(runner.cat_output()))
            self.assertEqual(len(lines), 1)
            self.assertEqual(int(lines[0]), 7)

    def _make_launcher(self, *args):
        """Make a launcher, add a mock runner (``launcher.mock_runner``), and
        set it up so that ``launcher.make_runner().__enter__()`` returns
        ``launcher.mock_runner()``.
        """
        launcher = MRJobLauncher(args=['--no-conf', ''] + list(args))
        launcher.sandbox()

        launcher.mock_runner = Mock()
        launcher.mock_runner.stream_output.return_value = [b'a line\n']

        launcher.make_runner = MagicMock()  # include __enter__
        launcher.make_runner.return_value.__enter__.return_value = (
            launcher.mock_runner)

        return launcher

 def test_emr_runner(self):
     launcher = MRJobLauncher(args=['--no-conf', '-r', 'emr', ''])
     with no_handlers_for_logger('mrjob.runner'):
         with launcher.make_runner() as runner:
             self.assertIsInstance(runner, EMRJobRunner)

 def test_emr_runner(self):
     launcher = MRJobLauncher(args=["--no-conf", "-r", "emr", ""])
     with no_handlers_for_logger("mrjob"):
         with patch_fs_s3():
             with launcher.make_runner() as runner:
                 self.assertIsInstance(runner, EMRJobRunner)

 def test_local_runner(self):
     launcher = MRJobLauncher(args=["--no-conf", "-r", "local", ""])
     with no_handlers_for_logger("mrjob.runner"):
         with launcher.make_runner() as runner:
             self.assertIsInstance(runner, LocalMRJobRunner)