X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/e838828374ceed5ef6da260939251e86f72b6f27..09cbdc3074b3f1e69c9c537875146f6da0a6ed8f:/sdk/cwl/arvados_cwl/arvworkflow.py diff --git a/sdk/cwl/arvados_cwl/arvworkflow.py b/sdk/cwl/arvados_cwl/arvworkflow.py index c6bcd04776..56226388d7 100644 --- a/sdk/cwl/arvados_cwl/arvworkflow.py +++ b/sdk/cwl/arvados_cwl/arvworkflow.py @@ -11,21 +11,25 @@ import copy import logging from schema_salad.sourceline import SourceLine, cmap +import schema_salad.ref_resolver + +import arvados.collection from cwltool.pack import pack -from cwltool.load_tool import fetch_document +from cwltool.load_tool import fetch_document, resolve_and_validate_document from cwltool.process import shortname from cwltool.workflow import Workflow, WorkflowException, WorkflowStep -from cwltool.pathmapper import adjustFileObjs, adjustDirObjs, visit_class +from cwltool.utils import adjustFileObjs, adjustDirObjs, visit_class, normalizeFilesDirs from cwltool.context import LoadingContext import ruamel.yaml as yaml from .runner import (upload_dependencies, packed_workflow, upload_workflow_collection, trim_anonymous_location, remove_redundant_fields, discover_secondary_files, - make_builder) + make_builder, arvados_jobs_image) from .pathmapper import ArvPathMapper, trim_listing from .arvtool import ArvadosCommandTool, set_cluster_target +from ._version import __version__ from .perf import Perf @@ -35,10 +39,91 @@ metrics = logging.getLogger('arvados.cwl-runner.metrics') max_res_pars = ("coresMin", "coresMax", "ramMin", "ramMax", "tmpdirMin", "tmpdirMax") sum_res_pars = ("outdirMin", "outdirMax") -def upload_workflow(arvRunner, tool, job_order, project_uuid, uuid=None, - submit_runner_ram=0, name=None, merged_map=None): +def make_wrapper_workflow(arvRunner, main, packed, project_uuid, name, git_info, tool): + col = arvados.collection.Collection(api_client=arvRunner.api, + keep_client=arvRunner.keep_client) + + with col.open("workflow.json", "wt") as f: + json.dump(packed, f, sort_keys=True, indent=4, separators=(',',': ')) + + pdh = col.portable_data_hash() + + toolname = tool.tool.get("label") or tool.metadata.get("label") or os.path.basename(tool.tool["id"]) + if git_info and git_info.get("http://arvados.org/cwl#gitDescribe"): + toolname = "%s (%s)" % (toolname, git_info.get("http://arvados.org/cwl#gitDescribe")) + + existing = arvRunner.api.collections().list(filters=[["portable_data_hash", "=", pdh], ["owner_uuid", "=", project_uuid]]).execute(num_retries=arvRunner.num_retries) + if len(existing["items"]) == 0: + col.save_new(name=toolname, owner_uuid=project_uuid, ensure_unique_name=True) + + # now construct the wrapper + + step = { + "id": "#main/" + toolname, + "in": [], + "out": [], + "run": "keep:%s/workflow.json#main" % pdh, + "label": name + } + + newinputs = [] + for i in main["inputs"]: + inp = {} + # Make sure to only copy known fields that are meaningful at + # the workflow level. In practice this ensures that if we're + # wrapping a CommandLineTool we don't grab inputBinding. + # Right now also excludes extension fields, which is fine, + # Arvados doesn't currently look for any extension fields on + # input parameters. + for f in ("type", "label", "secondaryFiles", "streamable", + "doc", "id", "format", "loadContents", + "loadListing", "default"): + if f in i: + inp[f] = i[f] + newinputs.append(inp) + + wrapper = { + "class": "Workflow", + "id": "#main", + "inputs": newinputs, + "outputs": [], + "steps": [step] + } + + for i in main["inputs"]: + step["in"].append({ + "id": "#main/step/%s" % shortname(i["id"]), + "source": i["id"] + }) + + for i in main["outputs"]: + step["out"].append({"id": "#main/step/%s" % shortname(i["id"])}) + wrapper["outputs"].append({"outputSource": "#main/step/%s" % shortname(i["id"]), + "type": i["type"], + "id": i["id"]}) + + wrapper["requirements"] = [{"class": "SubworkflowFeatureRequirement"}] + + if main.get("requirements"): + wrapper["requirements"].extend(main["requirements"]) + if main.get("hints"): + wrapper["hints"] = main["hints"] + + doc = {"cwlVersion": "v1.2", "$graph": [wrapper]} + + if git_info: + for g in git_info: + doc[g] = git_info[g] - packed = packed_workflow(arvRunner, tool, merged_map) + return json.dumps(doc, sort_keys=True, indent=4, separators=(',',': ')) + +def upload_workflow(arvRunner, tool, job_order, project_uuid, + runtimeContext, uuid=None, + submit_runner_ram=0, name=None, merged_map=None, + submit_runner_image=None, + git_info=None): + + packed = packed_workflow(arvRunner, tool, merged_map, runtimeContext, git_info) adjustDirObjs(job_order, trim_listing) adjustFileObjs(job_order, trim_anonymous_location) @@ -54,26 +139,38 @@ def upload_workflow(arvRunner, tool, job_order, project_uuid, uuid=None, name = tool.tool.get("label", os.path.basename(tool.tool["id"])) upload_dependencies(arvRunner, name, tool.doc_loader, - packed, tool.tool["id"], False) + packed, tool.tool["id"], False, + runtimeContext) + + wf_runner_resources = None + + hints = main.get("hints", []) + found = False + for h in hints: + if h["class"] == "http://arvados.org/cwl#WorkflowRunnerResources": + wf_runner_resources = h + found = True + break + if not found: + wf_runner_resources = {"class": "http://arvados.org/cwl#WorkflowRunnerResources"} + hints.append(wf_runner_resources) + + wf_runner_resources["acrContainerImage"] = arvados_jobs_image(arvRunner, + submit_runner_image or "arvados/jobs:"+__version__, + runtimeContext) if submit_runner_ram: - hints = main.get("hints", []) - found = False - for h in hints: - if h["class"] == "http://arvados.org/cwl#WorkflowRunnerResources": - h["ramMin"] = submit_runner_ram - found = True - break - if not found: - hints.append({"class": "http://arvados.org/cwl#WorkflowRunnerResources", - "ramMin": submit_runner_ram}) - main["hints"] = hints + wf_runner_resources["ramMin"] = submit_runner_ram + + main["hints"] = hints + + wrapper = make_wrapper_workflow(arvRunner, main, packed, project_uuid, name, git_info, tool) body = { "workflow": { "name": name, "description": tool.tool.get("doc", ""), - "definition":json.dumps(packed, sort_keys=True, indent=4, separators=(',',': ')) + "definition": wrapper }} if project_uuid: body["workflow"]["owner_uuid"] = project_uuid @@ -132,15 +229,21 @@ class ArvadosWorkflowStep(WorkflowStep): **argv ): # type: (...) -> None - super(ArvadosWorkflowStep, self).__init__(toolpath_object, pos, loadingContext, *argc, **argv) - self.tool["class"] = "WorkflowStep" + if arvrunner.fast_submit: + self.tool = toolpath_object + self.tool["inputs"] = [] + self.tool["outputs"] = [] + else: + super(ArvadosWorkflowStep, self).__init__(toolpath_object, pos, loadingContext, *argc, **argv) + self.tool["class"] = "WorkflowStep" self.arvrunner = arvrunner def job(self, joborder, output_callback, runtimeContext): runtimeContext = runtimeContext.copy() runtimeContext.toplevel = True # Preserve behavior for #13365 - builder = make_builder({shortname(k): v for k,v in viewitems(joborder)}, self.hints, self.requirements, runtimeContext) + builder = make_builder({shortname(k): v for k,v in viewitems(joborder)}, self.hints, self.requirements, + runtimeContext, self.metadata) runtimeContext = set_cluster_target(self.tool, self.arvrunner, builder, runtimeContext) return super(ArvadosWorkflowStep, self).job(joborder, output_callback, runtimeContext) @@ -160,7 +263,7 @@ class ArvadosWorkflow(Workflow): def job(self, joborder, output_callback, runtimeContext): - builder = make_builder(joborder, self.hints, self.requirements, runtimeContext) + builder = make_builder(joborder, self.hints, self.requirements, runtimeContext, self.metadata) runtimeContext = set_cluster_target(self.tool, self.arvrunner, builder, runtimeContext) req, _ = self.get_requirement("http://arvados.org/cwl#RunInSingleContainer") @@ -172,26 +275,31 @@ class ArvadosWorkflow(Workflow): with SourceLine(self.tool, None, WorkflowException, logger.isEnabledFor(logging.DEBUG)): if "id" not in self.tool: raise WorkflowException("%s object must have 'id'" % (self.tool["class"])) - document_loader, workflowobj, uri = (self.doc_loader, self.doc_loader.fetch(self.tool["id"]), self.tool["id"]) discover_secondary_files(self.arvrunner.fs_access, builder, self.tool["inputs"], joborder) + normalizeFilesDirs(joborder) with Perf(metrics, "subworkflow upload_deps"): upload_dependencies(self.arvrunner, os.path.basename(joborder.get("id", "#")), - document_loader, + self.doc_loader, joborder, joborder.get("id", "#"), - False) + False, + runtimeContext) if self.wf_pdh is None: - workflowobj["requirements"] = dedup_reqs(self.requirements) - workflowobj["hints"] = dedup_reqs(self.hints) + packed = pack(self.loadingContext, self.tool["id"], loader=self.doc_loader) - packed = pack(document_loader, workflowobj, uri, self.metadata) + for p in packed["$graph"]: + if p["id"] == "#main": + p["requirements"] = dedup_reqs(self.requirements) + p["hints"] = dedup_reqs(self.hints) def visit(item): + if "requirements" in item: + item["requirements"] = [i for i in item["requirements"] if i["class"] != "DockerRequirement"] for t in ("hints", "requirements"): if t not in item: continue @@ -211,9 +319,6 @@ class ArvadosWorkflow(Workflow): raise WorkflowException("Non-top-level ResourceRequirement in single container cannot have expressions") if not dyn: self.static_resource_req.append(req) - if req["class"] == "DockerRequirement": - if "http://arvados.org/cwl#dockerCollectionPDH" in req: - del req["http://arvados.org/cwl#dockerCollectionPDH"] visit_class(packed["$graph"], ("Workflow", "CommandLineTool"), visit) @@ -222,10 +327,11 @@ class ArvadosWorkflow(Workflow): upload_dependencies(self.arvrunner, runtimeContext.name, - document_loader, + self.doc_loader, packed, - uri, - False) + self.tool["id"], + False, + runtimeContext) # Discover files/directories referenced by the # workflow (mainly "default" values) @@ -289,12 +395,22 @@ class ArvadosWorkflow(Workflow): if self.wf_pdh is None: adjustFileObjs(packed, keepmount) adjustDirObjs(packed, keepmount) - self.wf_pdh = upload_workflow_collection(self.arvrunner, shortname(self.tool["id"]), packed) + self.wf_pdh = upload_workflow_collection(self.arvrunner, shortname(self.tool["id"]), packed, runtimeContext) self.loadingContext = self.loadingContext.copy() self.loadingContext.metadata = self.loadingContext.metadata.copy() self.loadingContext.metadata["http://commonwl.org/cwltool#original_cwlVersion"] = "v1.0" + if len(job_res_reqs) == 1: + # RAM request needs to be at least 128 MiB or the workflow + # runner itself won't run reliably. + if job_res_reqs[0].get("ramMin", 1024) < 128: + job_res_reqs[0]["ramMin"] = 128 + + arguments = ["--no-container", "--move-outputs", "--preserve-entire-environment", "workflow.cwl", "cwl.input.yml"] + if runtimeContext.debug: + arguments.insert(0, '--debug') + wf_runner = cmap({ "class": "CommandLineTool", "baseCommand": "cwltool", @@ -314,7 +430,7 @@ class ArvadosWorkflow(Workflow): }] }], "hints": self.hints, - "arguments": ["--no-container", "--move-outputs", "--preserve-entire-environment", "workflow.cwl#main", "cwl.input.yml"], + "arguments": arguments, "id": "#" }) return ArvadosCommandTool(self.arvrunner, wf_runner, self.loadingContext).job(joborder_resolved, output_callback, runtimeContext)