[ARVADOS] created: 53deed9cb00b94a69a3be1f959cbfda342646718
Git user
git at public.curoverse.com
Wed May 18 16:40:57 EDT 2016
at 53deed9cb00b94a69a3be1f959cbfda342646718 (commit)
commit 53deed9cb00b94a69a3be1f959cbfda342646718
Author: Peter Amstutz <peter.amstutz at curoverse.com>
Date: Wed May 18 16:40:48 2016 -0400
8653: Update cwl-runner to match changes in sdk/arvados-cwl-runner
diff --git a/crunch_scripts/cwl-runner b/crunch_scripts/cwl-runner
index 7c73493..e011177 100755
--- a/crunch_scripts/cwl-runner
+++ b/crunch_scripts/cwl-runner
@@ -32,6 +32,8 @@ try:
def keeppath(v):
if arvados.util.keep_locator_pattern.match(v):
return "file://%s/%s" % (os.environ['TASK_KEEPMOUNT'], v)
+ else:
+ return v
job_order_object["cwl:tool"] = keeppath(job_order_object["cwl:tool"])
@@ -55,7 +57,8 @@ try:
args.debug = True
args.quiet = False
args.ignore_docker_for_reuse = False
- outputObj = runner.arvExecutor(t, job_order_object, "", args, cwl_runner_job={"uuid": arvados.current_job()["uuid"], "state": arvados.current_job()["state"]})
+ args.cwl_runner_job={"uuid": arvados.current_job()["uuid"], "state": arvados.current_job()["state"]}
+ outputObj = runner.arvExecutor(t, job_order_object, **vars(args))
files = {}
def capture(path):
commit c8d9a898cde654b53200bda0b0ef8b406dd71739
Author: Peter Amstutz <peter.amstutz at curoverse.com>
Date: Wed May 18 16:33:57 2016 -0400
8653: cwl-runner crunch script rewrites keep file paths into CWL File objects.
Clean up argument handling in arvados-cwl-runner so that --create-template
doesn't require a job object, and that --help doesn't present options that are
irrelevant or don't work.
diff --git a/crunch_scripts/cwl-runner b/crunch_scripts/cwl-runner
index 8a4de24..7c73493 100755
--- a/crunch_scripts/cwl-runner
+++ b/crunch_scripts/cwl-runner
@@ -35,6 +35,13 @@ try:
job_order_object["cwl:tool"] = keeppath(job_order_object["cwl:tool"])
+ for k,v in job_order_object.items():
+ if arvados.util.keep_locator_pattern.match(v):
+ job_order_object[k] = {
+ "class": "File",
+ "path": keeppath(v)
+ }
+
adjustFiles(job_order_object, keeppath)
runner = arvados_cwl.ArvCwlRunner(api_client=arvados.api('v1', model=OrderedJsonModel()))
diff --git a/sdk/cwl/arvados_cwl/__init__.py b/sdk/cwl/arvados_cwl/__init__.py
index 8341624..61fa07e 100644
--- a/sdk/cwl/arvados_cwl/__init__.py
+++ b/sdk/cwl/arvados_cwl/__init__.py
@@ -11,13 +11,12 @@ import arvados.events
import arvados.util
import copy
import cwltool.docker
-import cwltool.draft2tool
+from cwltool.draft2tool import revmap_file, remove_hostfs, CommandLineTool
from cwltool.errors import WorkflowException
import cwltool.main
-from cwltool.process import shortname
import cwltool.workflow
import fnmatch
-import functools
+from functools import partial
import json
import logging
import os
@@ -25,8 +24,10 @@ import pkg_resources # part of setuptools
import re
import sys
import threading
+from schema_salad.ref_resolver import Loader
+from cwltool.builder import Builder
-from cwltool.process import get_feature, adjustFiles, scandeps
+from cwltool.process import shortname, get_feature, adjustFiles, adjustFileObjs, scandeps
from arvados.api import OrderedJsonModel
logger = logging.getLogger('arvados.cwl-runner')
@@ -315,7 +316,7 @@ class RunnerJob(object):
pass
def upload_docker(self, tool):
- if isinstance(tool, cwltool.draft2tool.CommandLineTool):
+ if isinstance(tool, CommandLineTool):
(docker_req, docker_is_req) = get_feature(tool, "DockerRequirement")
if docker_req:
arv_docker_get_image(self.arvrunner.api, docker_req, True, self.arvrunner.project_uuid)
@@ -342,7 +343,7 @@ class RunnerJob(object):
files.add(path)
return path
- document_loader, _, _ = cwltool.process.get_schema()
+ document_loader = Loader({"cwl": "https://w3id.org/cwl/cwl#", "id": "@id"})
def loadref(b, u):
return document_loader.resolve_ref(u, base_url=b)[0]
@@ -350,8 +351,8 @@ class RunnerJob(object):
set(("$import", "run")),
set(("$include", "$schemas", "path")),
loadref)
- adjustFiles(sc, functools.partial(visitFiles, workflowfiles))
- adjustFiles(self.job_order, functools.partial(visitFiles, jobfiles))
+ adjustFiles(sc, partial(visitFiles, workflowfiles))
+ adjustFiles(self.job_order, partial(visitFiles, jobfiles))
workflowmapper = ArvPathMapper(self.arvrunner, workflowfiles, "",
"%s",
@@ -410,6 +411,10 @@ class RunnerJob(object):
outc = arvados.collection.Collection(record["output"])
with outc.open("cwl.output.json") as f:
outputs = json.load(f)
+ def keepify(path):
+ if not path.startswith("keep:"):
+ return "keep:%s/%s" % (record["output"], path)
+ adjustFiles(outputs, keepify)
except Exception as e:
logger.error("While getting final output object: %s", e)
self.arvrunner.output_callback(outputs, processStatus)
@@ -504,7 +509,7 @@ class RunnerTemplate(object):
},
"name": self.job.name,
"owner_uuid": self.runner.project_uuid,
- }).execute(num_retries=self.runner.num_retries)
+ }, ensure_unique_name=True).execute(num_retries=self.runner.num_retries)
self.uuid = response["uuid"]
logger.info("Created template %s", self.uuid)
@@ -512,7 +517,7 @@ class RunnerTemplate(object):
class ArvPathMapper(cwltool.pathmapper.PathMapper):
"""Convert container-local paths to and from Keep collection ids."""
- def __init__(self, arvrunner, referenced_files, basedir,
+ def __init__(self, arvrunner, referenced_files, input_basedir,
collection_pattern, file_pattern, name=None, **kwargs):
self._pathmap = arvrunner.get_uploaded()
uploadfiles = set()
@@ -525,7 +530,7 @@ class ArvPathMapper(cwltool.pathmapper.PathMapper):
if "#" in src:
src = src[:src.index("#")]
if src not in self._pathmap:
- ab = cwltool.pathmapper.abspath(src, basedir)
+ ab = cwltool.pathmapper.abspath(src, input_basedir)
st = arvados.commands.run.statfile("", ab, fnPattern=file_pattern)
if kwargs.get("conformance_test"):
self._pathmap[src] = (src, ab)
@@ -560,7 +565,7 @@ class ArvPathMapper(cwltool.pathmapper.PathMapper):
return super(ArvPathMapper, self).reversemap(target)
-class ArvadosCommandTool(cwltool.draft2tool.CommandLineTool):
+class ArvadosCommandTool(CommandLineTool):
"""Wrap cwltool CommandLineTool to override selected methods."""
def __init__(self, arvrunner, toolpath_object, **kwargs):
@@ -637,45 +642,45 @@ class ArvCwlRunner(object):
def add_uploaded(self, src, pair):
self.uploaded[src] = pair
- def arvExecutor(self, tool, job_order, input_basedir, args, **kwargs):
- self.debug = args.debug
+ def arvExecutor(self, tool, job_order, **kwargs):
+ self.debug = kwargs.get("debug")
- if args.quiet:
+ if kwargs.get("quiet"):
logger.setLevel(logging.WARN)
logging.getLogger('arvados.arv-run').setLevel(logging.WARN)
useruuid = self.api.users().current().execute()["uuid"]
- self.project_uuid = args.project_uuid if args.project_uuid else useruuid
+ self.project_uuid = kwargs.get("project_uuid") if kwargs.get("project_uuid") else useruuid
self.pipeline = None
- if args.create_template:
- tmpl = RunnerTemplate(self, tool, job_order, args.enable_reuse)
+ if kwargs.get("create_template"):
+ tmpl = RunnerTemplate(self, tool, job_order, kwargs.get("enable_reuse"))
tmpl.save()
# cwltool.main will write our return value to stdout.
return tmpl.uuid
- if args.submit:
- runnerjob = RunnerJob(self, tool, job_order, args.enable_reuse)
- if not args.wait:
+ if kwargs.get("submit"):
+ runnerjob = RunnerJob(self, tool, job_order, kwargs.get("enable_reuse"))
+ if not kwargs.get("wait"):
runnerjob.run()
return runnerjob.uuid
events = arvados.events.subscribe(arvados.api('v1'), [["object_uuid", "is_a", "arvados#job"]], self.on_message)
- self.debug = args.debug
- self.ignore_docker_for_reuse = args.ignore_docker_for_reuse
- self.fs_access = CollectionFsAccess(input_basedir)
+ self.debug = kwargs.get("debug")
+ self.ignore_docker_for_reuse = kwargs.get("ignore_docker_for_reuse")
+ self.fs_access = CollectionFsAccess(kwargs["basedir"])
kwargs["fs_access"] = self.fs_access
- kwargs["enable_reuse"] = args.enable_reuse
+ kwargs["enable_reuse"] = kwargs.get("enable_reuse")
kwargs["outdir"] = "$(task.outdir)"
kwargs["tmpdir"] = "$(task.tmpdir)"
if kwargs.get("conformance_test"):
- return cwltool.main.single_job_executor(tool, job_order, input_basedir, args, **kwargs)
+ return cwltool.main.single_job_executor(tool, job_order, **kwargs)
else:
- if args.submit:
+ if kwargs.get("submit"):
jobiter = iter((runnerjob,))
else:
components = {}
@@ -692,7 +697,6 @@ class ArvCwlRunner(object):
logger.info("Pipeline instance %s", self.pipeline["uuid"])
jobiter = tool.job(job_order,
- input_basedir,
self.output_callback,
docker_outdir="$(task.outdir)",
**kwargs)
@@ -717,11 +721,6 @@ class ArvCwlRunner(object):
self.cond.wait(1)
events.close()
-
- if self.final_output is None:
- raise cwltool.workflow.WorkflowException("Workflow did not return a result.")
-
- # create final output collection
except:
if sys.exc_info()[0] is KeyboardInterrupt:
logger.error("Interrupted, marking pipeline as failed")
@@ -733,6 +732,9 @@ class ArvCwlRunner(object):
finally:
self.cond.release()
+ if self.final_output is None:
+ raise cwltool.workflow.WorkflowException("Workflow did not return a result.")
+
return self.final_output
def versionstring():
@@ -746,9 +748,25 @@ def versionstring():
"arvados-python-client", arvpkg[0].version,
"cwltool", cwlpkg[0].version)
-def main(args, stdout, stderr, api_client=None):
- args.insert(0, "--leave-outputs")
- parser = cwltool.main.arg_parser()
+def arg_parser(): # type: () -> argparse.ArgumentParser
+ parser = argparse.ArgumentParser(description='Arvados executor for Common Workflow Language')
+
+ parser.add_argument("--conformance-test", action="store_true")
+ parser.add_argument("--basedir", type=str)
+ parser.add_argument("--outdir", type=str, default=os.path.abspath('.'),
+ help="Output directory, default current directory")
+
+ parser.add_argument("--eval-timeout",
+ help="Time to wait for a Javascript expression to evaluate before giving an error.",
+ type=float)
+ parser.add_argument("--version", action="store_true", help="Print version and exit")
+
+ exgroup = parser.add_mutually_exclusive_group()
+ exgroup.add_argument("--verbose", action="store_true", help="Default logging")
+ exgroup.add_argument("--quiet", action="store_true", help="Only print warnings and errors.")
+ exgroup.add_argument("--debug", action="store_true", help="Print even more logging")
+
+ parser.add_argument("--tool-help", action="store_true", help="Print command line help for tool")
exgroup = parser.add_mutually_exclusive_group()
exgroup.add_argument("--enable-reuse", action="store_true",
@@ -776,6 +794,19 @@ def main(args, stdout, stderr, api_client=None):
exgroup.add_argument("--no-wait", action="store_false", help="Submit workflow runner job and exit.",
default=True, dest="wait")
+ parser.add_argument("workflow", type=str, nargs="?", default=None)
+ parser.add_argument("job_order", nargs=argparse.REMAINDER)
+
+ return parser
+
+def main(args, stdout, stderr, api_client=None):
+ parser = arg_parser()
+
+ job_order_object = None
+ arvargs = parser.parse_args(args)
+ if arvargs.create_template:
+ job_order_object = ({}, "")
+
try:
if api_client is None:
api_client=arvados.api('v1', model=OrderedJsonModel())
@@ -784,10 +815,10 @@ def main(args, stdout, stderr, api_client=None):
logger.error(e)
return 1
- return cwltool.main.main(args,
+ return cwltool.main.main(args=arvargs,
stdout=stdout,
stderr=stderr,
executor=runner.arvExecutor,
makeTool=runner.arvMakeTool,
- parser=parser,
- versionfunc=versionstring)
+ versionfunc=versionstring,
+ job_order_object=job_order_object)
diff --git a/sdk/cwl/setup.py b/sdk/cwl/setup.py
index c665a00..c069843 100644
--- a/sdk/cwl/setup.py
+++ b/sdk/cwl/setup.py
@@ -30,7 +30,7 @@ setup(name='arvados-cwl-runner',
'bin/arvados-cwl-runner'
],
install_requires=[
- 'cwltool==1.0.20160427142240',
+ 'cwltool==1.0.20160518201549',
'arvados-python-client>=0.1.20160322001610'
],
test_suite='tests',
-----------------------------------------------------------------------
hooks/post-receive
--
More information about the arvados-commits
mailing list