def testrun(args, parser):
"""TESTRUN
Reserve pages, run pages and optionally upload pages
"""
emop_submit = EmopSubmit(args.config_path)
# Do not run testrun subcommand if not in a valid cluster job environment
# This prevents accidentally running resource intensive program on login nodes
if not emop_submit.scheduler.is_job_environment():
print("Can only use testrun subcommand from within a cluster job environment")
sys.exit(1)
# Reserve pages equal to --num-pages
proc_id = emop_submit.reserve(num_pages=args.testrun_num_pages, r_filter=args.filter)
if not proc_id:
print("Failed to reserve pages")
sys.exit(1)
# Run reserved pages
emop_run = EmopRun(args.config_path, proc_id)
run_status = emop_run.run(force=True)
if not run_status:
sys.exit(1)
# Exit if --no-upload
if args.testrun_no_upload:
sys.exit(0)
# Upload results
emop_upload = EmopUpload(args.config_path)
upload_status = emop_upload.upload_proc_id(proc_id=proc_id)
if not upload_status:
sys.exit(1)
sys.exit(0)
def setUp(self):
self.popen_patcher = mock.patch("emop.lib.utilities.subprocess32.Popen")
self.mock_popen = self.popen_patcher.start()
self.mock_rv = mock.Mock()
self.mock_rv.communicate.return_value = ["", ""]
self.mock_rv.returncode = 0
self.mock_popen.return_value = self.mock_rv
os.environ["SLURM_JOB_ID"] = "2"
self.run = EmopRun(config_path=default_config_path(), proc_id="0001")
def run(args, parser):
"""run command
The run command is intended to be executed on a compute node. This command performs the actual work
of OCRing pages based on the supplied --proc-id value. By default the run command will not run if it detects the PROC_ID
has already run, but this can be modified with --force-run. This is useful when a batch job has been requeued.
"""
emop_run = EmopRun(args.config_path, args.proc_id)
# Do not use run subcommand if not in a valid cluster job environment
# This prevents accidentally running resource intensive program on login nodes
if not emop_run.scheduler.is_job_environment():
print("Can only use run subcommand from within a cluster job environment")
sys.exit(1)
run_status = emop_run.run(force=args.force_run)
if run_status:
sys.exit(0)
else:
sys.exit(1)
class TestEmopRun(TestCase):
def setUp(self):
self.popen_patcher = mock.patch("emop.lib.utilities.subprocess32.Popen")
self.mock_popen = self.popen_patcher.start()
self.mock_rv = mock.Mock()
self.mock_rv.communicate.return_value = ["", ""]
self.mock_rv.returncode = 0
self.mock_popen.return_value = self.mock_rv
os.environ["SLURM_JOB_ID"] = "2"
self.run = EmopRun(config_path=default_config_path(), proc_id="0001")
def tearDown(self):
self.popen_patcher.stop()
@pytest.fixture(autouse=True)
def setup_files(self, tmpdir):
self.tmpdir = tmpdir
os.environ["TMPDIR"] = str(self.tmpdir)
def test_append_result_failed(self):
settings = default_settings()
job = mock_emop_job(settings)
self.run.payload.save_output = mock.MagicMock()
self.run.append_result(job=job, results="Test", failed=True)
payload_save_args, payload_save_kwargs = self.run.payload.save_output.call_args
actual_failed_results = payload_save_kwargs["data"]["job_queues"]["failed"]
actual_completed_results = payload_save_kwargs["data"]["job_queues"]["completed"]
expected_failed = {"id": job.id, "results": "SLURM JOB 2: Test"}
self.assertEqual(1, len(actual_failed_results))
self.assertEqual(0, len(actual_completed_results))
self.assertEqual(expected_failed, actual_failed_results[0])
self.assertTrue(self.run.payload.save_output.called)
def test_append_result_completed(self):
settings = default_settings()
job = mock_emop_job(settings)
self.run.payload.save_output = mock.MagicMock()
self.run.append_result(job=job, results=None)
payload_save_args, payload_save_kwargs = self.run.payload.save_output.call_args
actual_failed_results = payload_save_kwargs["data"]["job_queues"]["failed"]
actual_completed_results = payload_save_kwargs["data"]["job_queues"]["completed"]
expected_completed = [job.id]
self.assertEqual(0, len(actual_failed_results))
self.assertEqual(1, len(actual_completed_results))
self.assertItemsEqual(expected_completed, actual_completed_results)
self.assertTrue(self.run.payload.save_output.called)
def test_get_results(self):
self.run.jobs_completed.append(1)
self.run.jobs_failed.append({"id": 2, "results": "test"})
self.run.page_results.append({"batch_id": 1, "page_id": 2})
self.run.postproc_results.append({"batch_job_id": 1, "page_id": 2})
expected_value = {
"extra_transfers": [],
"font_training_results": [],
"job_queues": {"completed": [1], "failed": [{"id": 2, "results": "test"}]},
"page_results": [{"batch_id": 1, "page_id": 2}],
"postproc_results": [{"batch_job_id": 1, "page_id": 2}],
}
actual_value = self.run.get_results()
self.assertEqual(expected_value, actual_value)
def test_do_process_page_corrector(self):
settings = default_settings()
job = mock_emop_job(settings)
page_corrector = PageCorrector(job=job)
page_corrector.run = mock.MagicMock()
results = mock_results_tuple()
page_corrector.should_run = mock.MagicMock()
page_corrector.should_run.return_value = True
page_corrector.run.return_value = results(stdout=None, stderr=None, exitcode=0)
retval = self.run.do_process(obj=page_corrector, job=job)
self.assertTrue(page_corrector.run.called)
self.assertTrue(retval)
def test_do_process_page_corrector_failed(self):
settings = default_settings()
job = mock_emop_job(settings)
page_corrector = PageCorrector(job=job)
page_corrector.run = mock.MagicMock()
results = mock_results_tuple()
page_corrector.should_run = mock.MagicMock()
page_corrector.should_run.return_value = True
page_corrector.run.return_value = results(stdout=None, stderr="Test", exitcode=1)
self.run.append_result = mock.MagicMock()
retval = self.run.do_process(obj=page_corrector, job=job)
self.run.append_result.assert_called_with(job=job, results="PageCorrector Failed: Test", failed=True)
self.assertFalse(retval)
def test_do_process_page_corrector_skipped(self):
settings = default_settings()
job = mock_emop_job(settings)
page_corrector = PageCorrector(job=job)
page_corrector.run = mock.MagicMock()
results = mock_results_tuple()
page_corrector.run.return_value = results(stdout=None, stderr="Test", exitcode=1)
flexmock(page_corrector).should_receive("should_run").and_return(False)
self.run.append_result = mock.MagicMock()
retval = self.run.do_process(obj=page_corrector, job=job)
self.assertFalse(self.run.append_result.called)
self.assertTrue(retval)
def test_do_process_page_corrector_not_skipped(self):
settings = default_settings()
self.run.settings.controller_skip_existing = False
job = mock_emop_job(settings)
page_corrector = PageCorrector(job=job)
page_corrector.run = mock.MagicMock()
results = mock_results_tuple()
page_corrector.run.return_value = results(stdout=None, stderr=None, exitcode=0)
page_corrector.should_run = mock.MagicMock()
self.run.append_result = mock.MagicMock()
retval = self.run.do_process(obj=page_corrector, job=job)
self.assertFalse(page_corrector.should_run.called)
self.assertTrue(retval)
def test_do_ocr_tesseract(self):
settings = default_settings()
job = mock_emop_job(settings)
results = mock_results_tuple()
tesseract = Tesseract(job=job)
results = mock_results_tuple()
expected_results = results(stdout=None, stderr=None, exitcode=0)
flexmock(os.path).should_receive("isfile").with_args(job.image_path).and_return(True)
mock_mkdirs(job.output_dir)
flexmock(os.path).should_receive("isfile").with_args(job.txt_file).and_return(True)
flexmock(os.path).should_receive("isfile").with_args(job.hocr_file).and_return(True)
flexmock(os.path).should_receive("isfile").with_args(job.xml_file).and_return(True)
flexmock(tesseract).should_receive("run").and_return(expected_results)
retval = self.run.do_ocr(job=job)
self.assertTrue(retval)
def test_do_ocr_tesseract_failed(self):
settings = default_settings()
job = mock_emop_job(settings)
job.page_result.ocr_text_path_exists = False
job.page_result.ocr_xml_path_exists = False
results = mock_results_tuple()
tesseract = Tesseract(job=job)
results = mock_results_tuple()
expected_results = "tesseract OCR Failed: Could not find page image %s" % job.image_path
flexmock(os.path).should_receive("isfile").with_args(job.txt_file).and_return(False)
flexmock(os.path).should_receive("isfile").with_args(job.xml_file).and_return(False)
flexmock(tesseract).should_receive("should_run").and_return(True)
flexmock(os.path).should_receive("isfile").with_args(job.image_path).and_return(False)
flexmock(tesseract).should_receive("run")
self.run.append_result = mock.MagicMock()
retval = self.run.do_ocr(job=job)
self.run.append_result.assert_called_with(job=job, results=expected_results, failed=True)
self.assertFalse(retval)
def test_do_ocr_tesseract_skipped(self):
settings = default_settings()
job = mock_emop_job(settings)
results = mock_results_tuple()
tesseract = Tesseract(job=job)
flexmock(os.path).should_receive("isfile").with_args(job.txt_file).and_return(True)
flexmock(os.path).should_receive("isfile").with_args(job.xml_file).and_return(True)
flexmock(tesseract).should_receive("should_run").and_return(False)
flexmock(tesseract).should_receive("run")
self.run.append_result = mock.MagicMock()
retval = self.run.do_ocr(job=job)
self.assertFalse(self.run.append_result.called)
self.assertTrue(retval)
# This test doesn't correctly validate should_run is not called.
# When self.run.settings.controller_skip_existing is not set to False
# the test still passes
# @skipif(True, reason="Does not work")
def test_do_ocr_tesseract_not_skipped(self):
settings = default_settings()
self.run.settings.controller_skip_existing = False
job = mock_emop_job(settings)
results = mock_results_tuple()
tesseract = Tesseract(job=job)
flexmock(os.path).should_receive("isdir").with_args(tesseract.output_parent_dir).and_return(True)
flexmock(os.path).should_receive("isfile").with_args(job.txt_file).and_return(False)
flexmock(os.path).should_receive("isfile").with_args(job.xml_file).and_return(True)
flexmock(os.path).should_receive("isfile").with_args(job.hocr_file).and_return(True)
flexmock(os.path).should_receive("isfile").with_args(job.image_path).and_return(True)
flexmock(tesseract).should_receive("should_run").never()
flexmock(tesseract).should_receive("run")
retval = self.run.do_ocr(job=job)
self.assertTrue(retval)
def test_do_postprocesses(self):
settings = default_settings()
job = mock_emop_job(settings)
# denoise = Denoise(job=job)
# multi_column_skew = MultiColumnSkew(job=job)
# xml_to_text_proc = XML_To_Text(job=job)
# page_evaluator = PageEvaluator(job=job)
# page_corrector = PageCorrector(job=job)
# juxta_compare = JuxtaCompare(job=job)
# These mocks don't work for some reason
# flexmock(self.run).should_receive("do_process").with_args(obj=denoise, job=job).and_return(True)
# flexmock(self.run).should_receive("do_process").with_args(obj=multi_column_skew, job=job).and_return(True)
# flexmock(self.run).should_receive("do_process").with_args(obj=xml_to_text_proc, job=job).and_return(True)
# flexmock(self.run).should_receive("do_process").with_args(obj=page_evaluator, job=job).and_return(True)
# flexmock(self.run).should_receive("do_process").with_args(obj=page_corrector, job=job).and_return(True)
# flexmock(self.run).should_receive("do_process").with_args(obj=juxta_compare, job=job).and_return(True)
flexmock(self.run).should_receive("do_process").and_return(True)
retval = self.run.do_postprocesses(job=job)
self.assertTrue(retval)
def test_do_postprocesses_failed(self):
settings = default_settings()
job = mock_emop_job(settings)
flexmock(self.run).should_receive("do_process").and_return(False)
retval = self.run.do_postprocesses(job=job)
self.assertFalse(retval)
def test_do_job(self):
settings = default_settings()
job = mock_emop_job(settings)
flexmock(self.run).should_receive("do_ocr").and_return(True)
flexmock(self.run).should_receive("do_postprocesses").and_return(True)
retval = self.run.do_job(job=job)
self.assertTrue(retval)
def test_do_job_failed_ocr(self):
settings = default_settings()
job = mock_emop_job(settings)
flexmock(self.run).should_receive("do_ocr").and_return(False)
flexmock(self.run).should_receive("do_postprocesses").and_return(True)
retval = self.run.do_job(job=job)
self.assertFalse(retval)
def test_do_job_failed_postprocesses(self):
settings = default_settings()
job = mock_emop_job(settings)
flexmock(self.run).should_receive("do_ocr").and_return(True)
flexmock(self.run).should_receive("do_postprocesses").and_return(False)
retval = self.run.do_job(job=job)
self.assertFalse(retval)
