def test_prev_uploaded(self, upl):
"""Test pathmapper handling previously uploaded files."""
arvrunner = arvados_cwl.ArvCwlRunner(self.api)
arvrunner.add_uploaded(
'tests/hw.py',
MapperEnt(
resolved='keep:99999999999999999999999999999991+99/hw.py',
target='',
type='File'))
upl.side_effect = upload_mock
p = ArvPathMapper(arvrunner, [{
"class": "File",
"location": "tests/hw.py"
}], "", "/test/%s", "/test/%s/%s")
self.assertEqual(
{
'tests/hw.py':
MapperEnt(
resolved='keep:99999999999999999999999999999991+99/hw.py',
target='/test/99999999999999999999999999999991+99/hw.py',
type='File')
}, p._pathmap)
def test_upload(self, statfile, upl):
"""Test pathmapper uploading files."""
arvrunner = arvados_cwl.ArvCwlRunner(self.api)
def statfile_mock(prefix,
fn,
fnPattern="$(file %s/%s)",
dirPattern="$(dir %s/%s/)",
raiseOSError=False):
st = arvados.commands.run.UploadFile("", "tests/hw.py")
return st
upl.side_effect = upload_mock
statfile.side_effect = statfile_mock
p = ArvPathMapper(arvrunner, [{
"class": "File",
"location": "file:tests/hw.py"
}], "", "/test/%s", "/test/%s/%s")
self.assertEqual(
{
'file:tests/hw.py':
MapperEnt(
resolved='keep:99999999999999999999999999999991+99/hw.py',
target='/test/99999999999999999999999999999991+99/hw.py',
type='File',
staged=True)
}, p._pathmap)
def test_statfile(self, statfile, upl):
"""Test pathmapper handling ArvFile references."""
arvrunner = arvados_cwl.ArvCwlRunner(self.api)
# An ArvFile object returned from arvados.commands.run.statfile means the file is located on a
# keep mount, so we can construct a direct reference directly without upload.
def statfile_mock(prefix,
fn,
fnPattern="$(file %s/%s)",
dirPattern="$(dir %s/%s/)"):
st = arvados.commands.run.ArvFile(
"",
fnPattern % ("99999999999999999999999999999991+99", "hw.py"))
return st
upl.side_effect = upload_mock
statfile.side_effect = statfile_mock
p = ArvPathMapper(arvrunner, [{
"class": "File",
"location": "tests/hw.py"
}], "", "/test/%s", "/test/%s/%s")
self.assertEqual(
{
'tests/hw.py':
MapperEnt(
resolved='keep:99999999999999999999999999999991+99/hw.py',
target='/test/99999999999999999999999999999991+99/hw.py',
type='File')
}, p._pathmap)
def test_default_work_api(self):
arvados_cwl.add_arv_hints()
api = mock.MagicMock()
api._rootDesc = copy.deepcopy(get_rootDesc())
del api._rootDesc.get('resources')['jobs']['methods']['create']
runner = arvados_cwl.ArvCwlRunner(api)
self.assertEqual(runner.work_api, 'containers')
def test_make_output_collection(self, reader, col):
keep_client = mock.MagicMock()
runner = arvados_cwl.ArvCwlRunner(self.api, keep_client=keep_client)
runner.project_uuid = 'zzzzz-j7d0g-zzzzzzzzzzzzzzz'
final = mock.MagicMock()
col.return_value = final
readermock = mock.MagicMock()
reader.return_value = readermock
final_uuid = final.manifest_locator()
num_retries = runner.num_retries
cwlout = StringIO.StringIO()
openmock = mock.MagicMock()
final.open.return_value = openmock
openmock.__enter__.return_value = cwlout
_, runner.final_output_collection = runner.make_output_collection("Test output", ["foo"], "tag0,tag1,tag2", {
"foo": {
"class": "File",
"location": "keep:99999999999999999999999999999991+99/foo.txt",
"size": 3,
"basename": "foo.txt"
},
"bar": {
"class": "File",
"location": "keep:99999999999999999999999999999992+99/bar.txt",
"basename": "baz.txt",
"size": 4
}
})
final.copy.assert_has_calls([mock.call('bar.txt', 'baz.txt', overwrite=False, source_collection=readermock)])
final.copy.assert_has_calls([mock.call('foo.txt', 'foo.txt', overwrite=False, source_collection=readermock)])
final.save_new.assert_has_calls([mock.call(ensure_unique_name=True, name='Test output', owner_uuid='zzzzz-j7d0g-zzzzzzzzzzzzzzz', storage_classes=['foo'])])
self.assertEqual("""{
"bar": {
"basename": "baz.txt",
"class": "File",
"location": "baz.txt",
"size": 4
},
"foo": {
"basename": "foo.txt",
"class": "File",
"location": "foo.txt",
"size": 3
}
}""", cwlout.getvalue())
self.assertIs(final, runner.final_output_collection)
self.assertIs(final_uuid, runner.final_output_collection.manifest_locator())
self.api.links().create.assert_has_calls([mock.call(body={"head_uuid": final_uuid, "link_class": "tag", "name": "tag0"}), mock.call().execute(num_retries=num_retries)])
self.api.links().create.assert_has_calls([mock.call(body={"head_uuid": final_uuid, "link_class": "tag", "name": "tag1"}), mock.call().execute(num_retries=num_retries)])
self.api.links().create.assert_has_calls([mock.call(body={"head_uuid": final_uuid, "link_class": "tag", "name": "tag2"}), mock.call().execute(num_retries=num_retries)])
def test_missing_file(self, stat):
"""Test pathmapper handling missing references."""
arvrunner = arvados_cwl.ArvCwlRunner(self.api)
stat.side_effect = OSError(2, "No such file or directory")
with self.assertRaises(OSError):
p = ArvPathMapper(arvrunner, [{
"class": "File",
"location": "file:tests/hw.py"
}], "", "/test/%s", "/test/%s/%s")
def test_keepref(self):
"""Test direct keep references."""
arvrunner = arvados_cwl.ArvCwlRunner(self.api)
p = ArvPathMapper(arvrunner, [{
"class": "File",
"location": "keep:99999999999999999999999999999991+99/hw.py"
}], "", "/test/%s", "/test/%s/%s")
self.assertEqual({'keep:99999999999999999999999999999991+99/hw.py': MapperEnt(resolved='keep:99999999999999999999999999999991+99/hw.py', target='/test/99999999999999999999999999999991+99/hw.py', type='File', staged=True)},
p._pathmap)
def test_upload(self, upl):
"""Test pathmapper uploading files."""
arvrunner = arvados_cwl.ArvCwlRunner(mock.MagicMock())
upl.side_effect = upload_mock
p = ArvPathMapper(arvrunner, [{
"class": "File",
"location": "tests/hw.py"
}], "", "/test/%s", "/test/%s/%s")
self.assertEqual(
{
'tests/hw.py':
MapperEnt(
resolved='keep:99999999999999999999999999999991+99/hw.py',
target='/test/99999999999999999999999999999991+99/hw.py',
type='File')
}, p._pathmap)
def test_run(self, list_images_in_arv, mockcollection,
mockcollectionreader):
arvados_cwl.add_arv_hints()
api = mock.MagicMock()
api._rootDesc = get_rootDesc()
runner = arvados_cwl.ArvCwlRunner(api)
self.assertEqual(runner.work_api, 'jobs')
list_images_in_arv.return_value = [["zzzzz-4zz18-zzzzzzzzzzzzzzz"]]
runner.api.collections().get().execute.return_vaulue = {
"portable_data_hash": "99999999999999999999999999999993+99"
}
runner.api.collections().list().execute.return_vaulue = {
"items": [{
"portable_data_hash":
"99999999999999999999999999999993+99"
}]
}
runner.project_uuid = "zzzzz-8i9sb-zzzzzzzzzzzzzzz"
runner.ignore_docker_for_reuse = False
runner.num_retries = 0
document_loader, avsc_names, schema_metadata, metaschema_loader = cwltool.process.get_schema(
"v1.0")
make_fs_access = functools.partial(
arvados_cwl.CollectionFsAccess,
collection_cache=arvados_cwl.CollectionCache(runner.api, None, 0))
document_loader.fetcher_constructor = functools.partial(
arvados_cwl.CollectionFetcher,
api_client=api,
fs_access=make_fs_access(""))
document_loader.fetcher = document_loader.fetcher_constructor(
document_loader.cache, document_loader.session)
document_loader.fetch_text = document_loader.fetcher.fetch_text
document_loader.check_exists = document_loader.fetcher.check_exists
tool, metadata = document_loader.resolve_ref("tests/wf/scatter2.cwl")
metadata["cwlVersion"] = tool["cwlVersion"]
mockcollection(
).portable_data_hash.return_value = "99999999999999999999999999999999+118"
arvtool = arvados_cwl.ArvadosWorkflow(runner,
tool,
work_api="jobs",
avsc_names=avsc_names,
basedir="",
make_fs_access=make_fs_access,
loader=document_loader,
makeTool=runner.arv_make_tool,
metadata=metadata)
arvtool.formatgraph = None
it = arvtool.job({},
mock.MagicMock(),
basedir="",
make_fs_access=make_fs_access)
it.next().run()
it.next().run()
with open("tests/wf/scatter2_subwf.cwl") as f:
subwf = StripYAMLComments(f.read())
runner.api.jobs().create.assert_called_with(
body=JsonDiffMatcher(
{
'minimum_script_version':
'a3f2cb186e437bfce0031b024b2157b73ed2717d',
'repository': 'arvados',
'script_version': 'master',
'script': 'crunchrunner',
'script_parameters': {
'tasks': [{
'task.env': {
'HOME': '$(task.outdir)',
'TMPDIR': '$(task.tmpdir)'
},
'task.vwd': {
'workflow.cwl':
'$(task.keep)/99999999999999999999999999999999+118/workflow.cwl',
'cwl.input.yml':
'$(task.keep)/99999999999999999999999999999999+118/cwl.input.yml'
},
'command': [
u'cwltool', u'--no-container',
u'--move-outputs',
u'--preserve-entire-environment',
u'workflow.cwl#main', u'cwl.input.yml'
],
'task.stdout':
'cwl.output.json'
}]
},
'runtime_constraints': {
'min_scratch_mb_per_node': 2048,
'min_cores_per_node': 1,
'docker_image': 'arvados/jobs',
'min_ram_mb_per_node': 1024
},
'owner_uuid': 'zzzzz-8i9sb-zzzzzzzzzzzzzzz'
}),
filters=[['repository', '=', 'arvados'],
['script', '=', 'crunchrunner'],
[
'script_version', 'in git',
'a3f2cb186e437bfce0031b024b2157b73ed2717d'
], ['docker_image_locator', 'in docker', 'arvados/jobs']],
find_or_create=True)
mockcollection().open().__enter__().write.assert_has_calls(
[mock.call(subwf)])
mockcollection().open().__enter__().write.assert_has_calls([
mock.call('''{
"fileblub": {
"basename": "token.txt",
"class": "File",
"location": "/keep/99999999999999999999999999999999+118/token.txt"
},
"sleeptime": 5
}''')
])
def test_run(self, mockcollection):
try:
arvados_cwl.add_arv_hints()
runner = arvados_cwl.ArvCwlRunner(mock.MagicMock())
runner.project_uuid = "zzzzz-8i9sb-zzzzzzzzzzzzzzz"
runner.ignore_docker_for_reuse = False
runner.num_retries = 0
document_loader, avsc_names, schema_metadata, metaschema_loader = cwltool.process.get_schema(
"v1.0")
tool, metadata = document_loader.resolve_ref(
"tests/wf/scatter2.cwl")
metadata["cwlVersion"] = tool["cwlVersion"]
mockcollection(
).portable_data_hash.return_value = "99999999999999999999999999999999+118"
make_fs_access = functools.partial(arvados_cwl.CollectionFsAccess,
api_client=runner.api)
arvtool = arvados_cwl.ArvadosWorkflow(
runner,
tool,
work_api="jobs",
avsc_names=avsc_names,
basedir="",
make_fs_access=make_fs_access,
loader=document_loader,
makeTool=runner.arv_make_tool,
metadata=metadata)
arvtool.formatgraph = None
it = arvtool.job({},
mock.MagicMock(),
basedir="",
make_fs_access=make_fs_access)
it.next().run()
it.next().run()
with open("tests/wf/scatter2_subwf.cwl") as f:
subwf = f.read()
mockcollection().open().__enter__().write.assert_has_calls(
[mock.call(subwf)])
mockcollection().open().__enter__().write.assert_has_calls(
[mock.call('{sleeptime: 5}')])
runner.api.jobs(
).create.assert_called_with(body={
'minimum_script_version':
'9e5b98e8f5f4727856b53447191f9c06e3da2ba6',
'repository':
'arvados',
'script_version':
'master',
'script':
'crunchrunner',
'script_parameters': {
'tasks': [{
'task.env': {
'HOME': '$(task.outdir)',
'TMPDIR': '$(task.tmpdir)'
},
'task.vwd': {
'workflow.cwl':
'$(task.keep)/99999999999999999999999999999999+118/workflow.cwl',
'cwl.input.yml':
'$(task.keep)/99999999999999999999999999999999+118/cwl.input.yml'
},
'command': [
u'cwltool', u'--no-container', u'--move-outputs',
u'--preserve-entire-environment',
u'workflow.cwl#main', u'cwl.input.yml'
],
'task.stdout':
'cwl.output.json'
}]
},
'runtime_constraints': {
'min_scratch_mb_per_node': 2048,
'min_cores_per_node': 1,
'docker_image': 'arvados/jobs',
'min_ram_mb_per_node': 1024
},
'owner_uuid':
'zzzzz-8i9sb-zzzzzzzzzzzzzzz'
},
filters=
[['repository', '=', 'arvados'],
['script', '=', 'crunchrunner'],
[
'script_version', 'in git',
'9e5b98e8f5f4727856b53447191f9c06e3da2ba6'
],
[
'docker_image_locator',
'in docker', 'arvados/jobs'
]],
find_or_create=True)
except:
logging.exception("")
def run():
# Timestamps are added by crunch-job, so don't print redundant timestamps.
arvados.log_handler.setFormatter(logging.Formatter('%(name)s %(levelname)s: %(message)s'))
# Print package versions
logger.info(arvados_cwl.versionstring())
api = arvados.api("v1")
arvados_cwl.add_arv_hints()
runner = None
try:
job_order_object = arvados.current_job()['script_parameters']
toolpath = "file://%s/%s" % (os.environ['TASK_KEEPMOUNT'], job_order_object.pop("cwl:tool"))
pdh_path = re.compile(r'^[0-9a-f]{32}\+\d+(/.+)?$')
def keeppath(v):
if pdh_path.match(v):
return "keep:%s" % v
else:
return v
def keeppathObj(v):
if "location" in v:
v["location"] = keeppath(v["location"])
for k,v in job_order_object.items():
if isinstance(v, basestring) and arvados.util.keep_locator_pattern.match(v):
job_order_object[k] = {
"class": "File",
"location": "keep:%s" % v
}
adjustFileObjs(job_order_object, keeppathObj)
adjustDirObjs(job_order_object, keeppathObj)
normalizeFilesDirs(job_order_object)
output_name = None
output_tags = None
enable_reuse = True
on_error = "continue"
debug = False
if "arv:output_name" in job_order_object:
output_name = job_order_object["arv:output_name"]
del job_order_object["arv:output_name"]
if "arv:output_tags" in job_order_object:
output_tags = job_order_object["arv:output_tags"]
del job_order_object["arv:output_tags"]
if "arv:enable_reuse" in job_order_object:
enable_reuse = job_order_object["arv:enable_reuse"]
del job_order_object["arv:enable_reuse"]
if "arv:on_error" in job_order_object:
on_error = job_order_object["arv:on_error"]
del job_order_object["arv:on_error"]
if "arv:debug" in job_order_object:
debug = job_order_object["arv:debug"]
del job_order_object["arv:debug"]
runner = arvados_cwl.ArvCwlRunner(api_client=arvados.safeapi.ThreadSafeApiCache(
api_params={"model": OrderedJsonModel()}, keep_params={"num_retries": 4}),
output_name=output_name, output_tags=output_tags)
make_fs_access = functools.partial(CollectionFsAccess,
collection_cache=runner.collection_cache)
t = load_tool(toolpath, runner.arv_make_tool,
fetcher_constructor=functools.partial(CollectionFetcher,
api_client=runner.api,
fs_access=make_fs_access(""),
num_retries=runner.num_retries))
if debug:
logger.setLevel(logging.DEBUG)
logging.getLogger('arvados').setLevel(logging.DEBUG)
logging.getLogger("cwltool").setLevel(logging.DEBUG)
args = argparse.Namespace()
args.project_uuid = arvados.current_job()["owner_uuid"]
args.enable_reuse = enable_reuse
args.on_error = on_error
args.submit = False
args.debug = debug
args.quiet = False
args.ignore_docker_for_reuse = False
args.basedir = os.getcwd()
args.name = None
args.cwl_runner_job={"uuid": arvados.current_job()["uuid"], "state": arvados.current_job()["state"]}
args.make_fs_access = make_fs_access
args.trash_intermediate = False
args.intermediate_output_ttl = 0
args.priority = arvados_cwl.DEFAULT_PRIORITY
args.do_validate = True
args.disable_js_validation = False
runner.arv_executor(t, job_order_object, **vars(args))
except Exception as e:
if isinstance(e, WorkflowException):
logging.info("Workflow error %s", e)
else:
logging.exception("Unhandled exception")
if runner and runner.final_output_collection:
outputCollection = runner.final_output_collection.portable_data_hash()
else:
outputCollection = None
api.job_tasks().update(uuid=arvados.current_task()['uuid'],
body={
'output': outputCollection,
'success': False,
'progress':1.0
}).execute()
def test_overall_resource_singlecontainer(self, list_images_in_arv, mockcollection, mockcollectionreader):
arvados_cwl.add_arv_hints()
api = mock.MagicMock()
api._rootDesc = get_rootDesc()
runner = arvados_cwl.ArvCwlRunner(api)
self.assertEqual(runner.work_api, 'jobs')
list_images_in_arv.return_value = [["zzzzz-4zz18-zzzzzzzzzzzzzzz"]]
runner.api.collections().get().execute.return_vaulue = {"portable_data_hash": "99999999999999999999999999999993+99"}
runner.api.collections().list().execute.return_vaulue = {"items": [{"portable_data_hash": "99999999999999999999999999999993+99"}]}
runner.project_uuid = "zzzzz-8i9sb-zzzzzzzzzzzzzzz"
runner.ignore_docker_for_reuse = False
runner.num_retries = 0
loadingContext, runtimeContext = self.helper(runner)
tool, metadata = loadingContext.loader.resolve_ref("tests/wf/echo-wf.cwl")
metadata["cwlVersion"] = tool["cwlVersion"]
mockcollection.side_effect = lambda *args, **kwargs: CollectionMock(mock.MagicMock(), *args, **kwargs)
arvtool = arvados_cwl.ArvadosWorkflow(runner, tool, loadingContext)
arvtool.formatgraph = None
it = arvtool.job({}, mock.MagicMock(), runtimeContext)
it.next().run(runtimeContext)
it.next().run(runtimeContext)
with open("tests/wf/echo-subwf.cwl") as f:
subwf = StripYAMLComments(f.read())
runner.api.jobs().create.assert_called_with(
body=JsonDiffMatcher({
'minimum_script_version': 'a3f2cb186e437bfce0031b024b2157b73ed2717d',
'repository': 'arvados',
'script_version': 'master',
'script': 'crunchrunner',
'script_parameters': {
'tasks': [{'task.env': {
'HOME': '$(task.outdir)',
'TMPDIR': '$(task.tmpdir)'},
'task.vwd': {
'workflow.cwl': '$(task.keep)/99999999999999999999999999999996+99/workflow.cwl',
'cwl.input.yml': '$(task.keep)/99999999999999999999999999999996+99/cwl.input.yml'
},
'command': [u'cwltool', u'--no-container', u'--move-outputs', u'--preserve-entire-environment', u'workflow.cwl#main', u'cwl.input.yml'],
'task.stdout': 'cwl.output.json'}]},
'runtime_constraints': {
'min_scratch_mb_per_node': 4096,
'min_cores_per_node': 3,
'docker_image': 'arvados/jobs',
'min_ram_mb_per_node': 1024
},
'owner_uuid': 'zzzzz-8i9sb-zzzzzzzzzzzzzzz'}),
filters=[['repository', '=', 'arvados'],
['script', '=', 'crunchrunner'],
['script_version', 'in git', 'a3f2cb186e437bfce0031b024b2157b73ed2717d'],
['docker_image_locator', 'in docker', 'arvados/jobs']],
find_or_create=True)
def run():
# Print package versions
logger.info(arvados_cwl.versionstring())
api = arvados.api("v1")
arvados_cwl.add_arv_hints()
runner = None
try:
job_order_object = arvados.current_job()['script_parameters']
pdh_path = re.compile(r'^[0-9a-f]{32}\+\d+(/.+)?$')
def keeppath(v):
if pdh_path.match(v):
return "keep:%s" % v
else:
return v
def keeppathObj(v):
v["location"] = keeppath(v["location"])
job_order_object["cwl:tool"] = "file://%s/%s" % (
os.environ['TASK_KEEPMOUNT'], job_order_object["cwl:tool"])
for k, v in job_order_object.items():
if isinstance(
v,
basestring) and arvados.util.keep_locator_pattern.match(v):
job_order_object[k] = {
"class": "File",
"location": "keep:%s" % v
}
adjustFileObjs(job_order_object, keeppathObj)
adjustDirObjs(job_order_object, keeppathObj)
normalizeFilesDirs(job_order_object)
adjustDirObjs(
job_order_object,
functools.partial(
getListing,
arvados_cwl.fsaccess.CollectionFsAccess("", api_client=api)))
output_name = None
output_tags = None
enable_reuse = True
if "arv:output_name" in job_order_object:
output_name = job_order_object["arv:output_name"]
del job_order_object["arv:output_name"]
if "arv:output_tags" in job_order_object:
output_tags = job_order_object["arv:output_tags"]
del job_order_object["arv:output_tags"]
if "arv:enable_reuse" in job_order_object:
enable_reuse = job_order_object["arv:enable_reuse"]
del job_order_object["arv:enable_reuse"]
runner = arvados_cwl.ArvCwlRunner(api_client=arvados.api(
'v1', model=OrderedJsonModel()),
output_name=output_name,
output_tags=output_tags)
t = load_tool(job_order_object, runner.arv_make_tool)
args = argparse.Namespace()
args.project_uuid = arvados.current_job()["owner_uuid"]
args.enable_reuse = enable_reuse
args.submit = False
args.debug = True
args.quiet = False
args.ignore_docker_for_reuse = False
args.basedir = os.getcwd()
args.cwl_runner_job = {
"uuid": arvados.current_job()["uuid"],
"state": arvados.current_job()["state"]
}
outputObj = runner.arv_executor(t, job_order_object, **vars(args))
except Exception as e:
if isinstance(e, WorkflowException):
logging.info("Workflow error %s", e)
else:
logging.exception("Unhandled exception")
if runner and runner.final_output_collection:
outputCollection = runner.final_output_collection.portable_data_hash(
)
else:
outputCollection = None
api.job_tasks().update(uuid=arvados.current_task()['uuid'],
body={
'output': outputCollection,
'success': False,
'progress': 1.0
}).execute()
