def arv_docker_get_image(api_client, dockerRequirement, pull_image, project_uuid,
- force_pull, tmp_outdir_prefix, match_local_docker):
+ force_pull, tmp_outdir_prefix, match_local_docker, copy_deps):
"""Check if a Docker image is available in Keep, if not, upload it using arv-keepdocker."""
if "http://arvados.org/cwl#dockerCollectionPDH" in dockerRequirement:
image_tag=image_tag,
project_uuid=None)
- images = arvados.commands.keepdocker.list_images_in_arv(api_client, 3,
- image_name=image_name,
- image_tag=image_tag,
- project_uuid=project_uuid)
+ if copy_deps:
+ # Only images that are available in the destination project
+ images = arvados.commands.keepdocker.list_images_in_arv(api_client, 3,
+ image_name=image_name,
+ image_tag=image_tag,
+ project_uuid=project_uuid)
+ else:
+ images = out_of_project_images
if match_local_docker:
local_image_id = determine_image_id(dockerRequirement["dockerImageId"])
if not runtimeContext.name:
runtimeContext.name = self.name = updated_tool.tool.get("label") or updated_tool.metadata.get("label") or os.path.basename(updated_tool.tool["id"])
+ if runtimeContext.copy_deps is None and (runtimeContext.create_workflow or runtimeContext.update_workflow):
+ runtimeContext.copy_deps = True
+
+ if runtimeContext.update_workflow and self.project_uuid is None:
+ # If we are updating a workflow, make sure anything that
+ # gets uploaded goes into the same parent project, unless
+ # an alternate --project-uuid was provided.
+ existing_wf = self.api.workflows().get(uuid=runtimeContext.update_workflow).execute()
+ self.project_uuid = existing_wf["owner_uuid"]
+
# Upload local file references in the job order.
job_order = upload_job_order(self, "%s input" % runtimeContext.name,
updated_tool, job_order)
else:
tool = updated_tool
- if runtimeContext.update_workflow and self.project_uuid is None:
- existing_wf = self.api.workflows().get(uuid=runtimeContext.update_workflow).execute()
- self.project_uuid = existing_wf["owner_uuid"]
-
# Upload direct dependencies of workflow steps, get back mapping of files to keep references.
# Also uploads docker images.
merged_map = upload_workflow_deps(self, tool)
import schema_salad.validate as validate
import arvados.collection
+import arvados.util
from .util import collectionUUID
from ruamel.yaml import YAML
from ruamel.yaml.comments import CommentedMap, CommentedSeq
single_collection=True,
optional_deps=optional_deps)
+ keeprefs = set()
+ def addkeepref(k):
+ keeprefs.add(collection_pdh_pattern.match(k).group(1))
+
def setloc(p):
loc = p.get("location")
if loc and (not loc.startswith("_:")) and (not loc.startswith("keep:")):
p["location"] = mapper.mapper(p["location"]).resolved
+ addkeepref(p["location"])
return
if not loc:
gp = collection_uuid_pattern.match(loc)
if not gp:
+ # Not a uuid pattern (must be a pdh pattern)
+ addkeepref(p["location"])
return
+
uuid = gp.groups()[0]
if uuid not in uuid_map:
raise SourceLine(p, "location", validate.ValidationException).makeError(
for d in discovered:
discovered_secondaryfiles[mapper.mapper(d).resolved] = discovered[d]
+ if arvrunner.runtimeContext.copy_deps:
+ # Find referenced collections and copy them into the
+ # destination project, for easy sharing.
+ already_present = list(arvados.util.keyset_list_all(arvrunner.api.collections().list,
+ filters=[["portable_data_hash", "in", list(keeprefs)],
+ ["owner_uuid", "=", arvrunner.project_uuid]],
+ select=["uuid", "portable_data_hash", "created_at"]))
+
+ keeprefs = keeprefs - set(a["portable_data_hash"] for a in already_present)
+ for kr in keeprefs:
+ col = arvrunner.api.collections().list(filters=[["portable_data_hash", "=", kr]],
+ order="created_at desc",
+ select=["name", "description", "properties", "portable_data_hash", "manifest_text", "storage_classes_desired", "trash_at"],
+ limit=1).execute()
+ if len(col["items"]) == 0:
+ logger.warning("Cannot find collection with portable data hash %s", kr)
+ continue
+ col = col["items"][0]
+ try:
+ arvrunner.api.collections().create(body={"collection": {
+ "owner_uuid": arvrunner.project_uuid,
+ "name": col["name"],
+ "description": col["description"],
+ "properties": col["properties"],
+ "portable_data_hash": col["portable_data_hash"],
+ "manifest_text": col["manifest_text"],
+ "storage_classes_desired": col["storage_classes_desired"],
+ "trash_at": col["trash_at"]
+ }}, ensure_unique_name=True).execute()
+ except Exception as e:
+ logger.warning("Unable copy collection to destination: %s", e)
+
if "$schemas" in workflowobj:
sch = CommentedSeq()
for s in workflowobj["$schemas"]:
arvados_cwl.arvdocker.arv_docker_get_image(arvrunner.api, docker_req, True, arvrunner.project_uuid,
arvrunner.runtimeContext.force_docker_pull,
arvrunner.runtimeContext.tmp_outdir_prefix,
- arvrunner.runtimeContext.match_local_docker)
+ arvrunner.runtimeContext.match_local_docker,
+ arvrunner.runtimeContext.copy_deps)
else:
arvados_cwl.arvdocker.arv_docker_get_image(arvrunner.api, {"dockerPull": "arvados/jobs:"+__version__},
True, arvrunner.project_uuid,
arvrunner.runtimeContext.force_docker_pull,
arvrunner.runtimeContext.tmp_outdir_prefix,
- arvrunner.runtimeContext.match_local_docker)
+ arvrunner.runtimeContext.match_local_docker,
+ arvrunner.runtimeContext.copy_deps)
elif isinstance(tool, cwltool.workflow.Workflow):
for s in tool.steps:
upload_docker(arvrunner, s.embedded_tool)
arvrunner.project_uuid,
arvrunner.runtimeContext.force_docker_pull,
arvrunner.runtimeContext.tmp_outdir_prefix,
- arvrunner.runtimeContext.match_local_docker)
+ arvrunner.runtimeContext.match_local_docker,
+ arvrunner.runtimeContext.copy_deps)
for l in v:
visit(v[l], cur_id)
if isinstance(v, list):
return arvados_cwl.arvdocker.arv_docker_get_image(arvrunner.api, {"dockerPull": img}, True, arvrunner.project_uuid,
arvrunner.runtimeContext.force_docker_pull,
arvrunner.runtimeContext.tmp_outdir_prefix,
- arvrunner.runtimeContext.match_local_docker)
+ arvrunner.runtimeContext.match_local_docker,
+ arvrunner.runtimeContext.copy_deps)
except Exception as e:
raise Exception("Docker image %s is not available\n%s" % (img, e) )