#
# SPDX-License-Identifier: Apache-2.0
+from past.builtins import basestring
+from builtins import object
+from future.utils import viewitems
+
import logging
import re
import copy
from schema_salad.sourceline import SourceLine
+import arvados_cwl.util
import ruamel.yaml as yaml
import arvados.collection
from .perf import Perf
from . import done
from ._version import __version__
+from .util import get_intermediate_collection_info
logger = logging.getLogger('arvados.cwl-runner')
metrics = logging.getLogger('arvados.cwl-runner.metrics')
keep_client=self.arvrunner.keep_client,
num_retries=self.arvrunner.num_retries)
script_parameters["task.vwd"] = {}
- generatemapper = VwdPathMapper([self.generatefiles], "", "",
+ generatemapper = VwdPathMapper(self.generatefiles["listing"], "", "",
separateDirs=False)
with Perf(metrics, "createfiles %s" % self.name):
if vwd:
with Perf(metrics, "generatefiles.save_new %s" % self.name):
- vwd.save_new()
+ info = get_intermediate_collection_info(self.name, None, runtimeContext.intermediate_output_ttl)
+ vwd.save_new(name=info["name"],
+ owner_uuid=self.arvrunner.project_uuid,
+ ensure_unique_name=True,
+ trash_at=info["trash_at"],
+ properties=info["properties"])
for f, p in generatemapper.items():
if p.type == "File":
with Perf(metrics, "arv_docker_get_image %s" % self.name):
(docker_req, docker_is_req) = self.get_requirement("DockerRequirement")
- if docker_req and runtimeContextuse_container is not False:
+ if docker_req and runtimeContext.use_container is not False:
if docker_req.get("dockerOutputDirectory"):
raise SourceLine(docker_req, "dockerOutputDirectory", UnsupportedRequirement).makeError(
"Option 'dockerOutputDirectory' of DockerRequirement not supported.")
- runtime_constraints["docker_image"] = arv_docker_get_image(self.arvrunner.api, docker_req, pull_image, self.arvrunner.project_uuid)
+ runtime_constraints["docker_image"] = arv_docker_get_image(self.arvrunner.api,
+ docker_req,
+ runtimeContext.pull_image,
+ self.arvrunner.project_uuid)
else:
runtime_constraints["docker_image"] = "arvados/jobs"
api_client=self.arvrunner.api,
keep_client=self.arvrunner.keep_client,
num_retries=self.arvrunner.num_retries)
- log = logc.open(logc.keys()[0])
+ log = logc.open(list(logc.keys())[0])
dirs = {
"tmpdir": "/tmpdir",
"outdir": "/outdir",
dirs[g.group(1)] = g.group(2)
if processStatus == "permanentFail":
- done.logtail(logc, logger, "%s error log:" % self.arvrunner.label(self))
+ done.logtail(logc, logger.error, "%s (%s) error log:" % (self.arvrunner.label(self), record["uuid"]), maxlen=40)
with Perf(metrics, "output collection %s" % self.name):
outputs = done.done(self, record, dirs["tmpdir"],
class RunnerJob(Runner):
"""Submit and manage a Crunch job that runs crunch_scripts/cwl-runner."""
- def arvados_job_spec(self, runtimeContext):
+ def arvados_job_spec(self, debug=False):
"""Create an Arvados job specification for this workflow.
The returned dict can be used to create a job (i.e., passed as
a pipeline template or pipeline instance.
"""
- if self.tool.tool["id"].startswith("keep:"):
- self.job_order["cwl:tool"] = self.tool.tool["id"][5:]
+ if self.embedded_tool.tool["id"].startswith("keep:"):
+ self.job_order["cwl:tool"] = self.embedded_tool.tool["id"][5:]
else:
- packed = packed_workflow(self.arvrunner, self.tool, self.merged_map)
+ packed = packed_workflow(self.arvrunner, self.embedded_tool, self.merged_map)
wf_pdh = upload_workflow_collection(self.arvrunner, self.name, packed)
self.job_order["cwl:tool"] = "%s/workflow.cwl#main" % wf_pdh
if self.on_error:
self.job_order["arv:on_error"] = self.on_error
- if runtimeContext.debug:
+ if debug:
self.job_order["arv:debug"] = True
return {
}
def run(self, runtimeContext):
- job_spec = self.arvados_job_spec(runtimeContext)
+ job_spec = self.arvados_job_spec(runtimeContext.debug)
job_spec.setdefault("owner_uuid", self.arvrunner.project_uuid)
find_or_create=self.enable_reuse
).execute(num_retries=self.arvrunner.num_retries)
- for k,v in job_spec["script_parameters"].items():
+ for k,v in viewitems(job_spec["script_parameters"]):
if v is False or v is None or isinstance(v, dict):
job_spec["script_parameters"][k] = {"value": v}
}
def __init__(self, runner, tool, job_order, enable_reuse, uuid,
- submit_runner_ram=0, name=None, merged_map=None):
+ submit_runner_ram=0, name=None, merged_map=None,
+ loadingContext=None):
self.runner = runner
- self.tool = tool
+ self.embedded_tool = tool
self.job = RunnerJob(
runner=runner,
tool=tool,
- job_order=job_order,
enable_reuse=enable_reuse,
output_name=None,
output_tags=None,
submit_runner_ram=submit_runner_ram,
name=name,
- merged_map=merged_map)
+ merged_map=merged_map,
+ loadingContext=loadingContext)
+ self.job.job_order = job_order
self.uuid = uuid
def pipeline_component_spec(self):
job_params = spec['script_parameters']
spec['script_parameters'] = {}
- for param in self.tool.tool['inputs']:
+ for param in self.embedded_tool.tool['inputs']:
param = copy.deepcopy(param)
# Data type and "required" flag...