X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/734335da27f27e2177d3b931b1e5e9e8e83a042f..2a3daf14afb93de4d65108019a7a1d35aa1052ad:/sdk/cwl/arvados_cwl/runner.py diff --git a/sdk/cwl/arvados_cwl/runner.py b/sdk/cwl/arvados_cwl/runner.py index 49d37ebd5a..28de7f368a 100644 --- a/sdk/cwl/arvados_cwl/runner.py +++ b/sdk/cwl/arvados_cwl/runner.py @@ -1,46 +1,69 @@ +# Copyright (C) The Arvados Authors. All rights reserved. +# +# SPDX-License-Identifier: Apache-2.0 + import os import urlparse from functools import partial import logging import json -import re -from cStringIO import StringIO +import subprocess + +from StringIO import StringIO + +from schema_salad.sourceline import SourceLine import cwltool.draft2tool from cwltool.draft2tool import CommandLineTool import cwltool.workflow -from cwltool.process import get_feature, scandeps, UnsupportedRequirement, normalizeFilesDirs +from cwltool.process import get_feature, scandeps, UnsupportedRequirement, normalizeFilesDirs, shortname from cwltool.load_tool import fetch_document -from cwltool.pathmapper import adjustFileObjs, adjustDirObjs +from cwltool.pathmapper import adjustFileObjs, adjustDirObjs, visit_class +from cwltool.utils import aslist +from cwltool.builder import substitute +from cwltool.pack import pack import arvados.collection import ruamel.yaml as yaml from .arvdocker import arv_docker_get_image -from .pathmapper import ArvPathMapper +from .pathmapper import ArvPathMapper, trim_listing from ._version import __version__ +from . import done logger = logging.getLogger('arvados.cwl-runner') -cwltool.draft2tool.ACCEPTLIST_RE = re.compile(r".*") +def trim_anonymous_location(obj): + """Remove 'location' field from File and Directory literals. -def trim_listing(obj): - """Remove 'listing' field from Directory objects that are keep references. + To make internal handling easier, literals are assigned a random id for + 'location'. However, when writing the record back out, this can break + reproducibility. Since it is valid for literals not have a 'location' + field, remove it. - When Directory objects represent Keep references, it redundant and - potentially very expensive to pass fully enumerated Directory objects - between instances of cwl-runner (e.g. a submitting a job, or using the - RunInSingleContainer feature), so delete the 'listing' field when it is - safe to do so. """ - if obj.get("location", "").startswith("keep:") and "listing" in obj: - del obj["listing"] if obj.get("location", "").startswith("_:"): del obj["location"] +def remove_redundant_fields(obj): + for field in ("path", "nameext", "nameroot", "dirname"): + if field in obj: + del obj[field] + +def find_defaults(d, op): + if isinstance(d, list): + for i in d: + find_defaults(i, op) + elif isinstance(d, dict): + if "default" in d: + op(d) + else: + for i in d.itervalues(): + find_defaults(i, op) + def upload_dependencies(arvrunner, name, document_loader, - workflowobj, uri, loadref_run): + workflowobj, uri, loadref_run, include_primary=True): """Upload the dependencies of the workflowobj document to Keep. Returns a pathmapper object mapping local paths to keep references. Also @@ -55,7 +78,7 @@ def upload_dependencies(arvrunner, name, document_loader, loaded = set() def loadref(b, u): - joined = urlparse.urljoin(b, u) + joined = document_loader.fetcher.urljoin(b, u) defrg, _ = urlparse.urldefrag(joined) if defrg not in loaded: loaded.add(defrg) @@ -83,17 +106,39 @@ def upload_dependencies(arvrunner, name, document_loader, sc = scandeps(uri, scanobj, loadref_fields, set(("$include", "$schemas", "location")), - loadref) + loadref, urljoin=document_loader.fetcher.urljoin) normalizeFilesDirs(sc) - if "id" in workflowobj: + if include_primary and "id" in workflowobj: sc.append({"class": "File", "location": workflowobj["id"]}) + if "$schemas" in workflowobj: + for s in workflowobj["$schemas"]: + sc.append({"class": "File", "location": s}) + + def capture_default(obj): + remove = [False] + def add_default(f): + if "location" not in f and "path" in f: + f["location"] = f["path"] + del f["path"] + if "location" in f and not arvrunner.fs_access.exists(f["location"]): + # Remove from sc + sc[:] = [x for x in sc if x["location"] != f["location"]] + # Delete "default" from workflowobj + remove[0] = True + visit_class(obj["default"], ("File", "Directory"), add_default) + if remove[0]: + del obj["default"] + + find_defaults(workflowobj, capture_default) + mapper = ArvPathMapper(arvrunner, sc, "", "keep:%s", "keep:%s/%s", - name=name) + name=name, + single_collection=True) def setloc(p): if "location" in p and (not p["location"].startswith("_:")) and (not p["location"].startswith("keep:")): @@ -101,100 +146,229 @@ def upload_dependencies(arvrunner, name, document_loader, adjustFileObjs(workflowobj, setloc) adjustDirObjs(workflowobj, setloc) + if "$schemas" in workflowobj: + sch = [] + for s in workflowobj["$schemas"]: + sch.append(mapper.mapper(s).resolved) + workflowobj["$schemas"] = sch + return mapper def upload_docker(arvrunner, tool): + """Uploads Docker images used in CommandLineTool objects.""" + if isinstance(tool, CommandLineTool): (docker_req, docker_is_req) = get_feature(tool, "DockerRequirement") if docker_req: + if docker_req.get("dockerOutputDirectory"): + # TODO: can be supported by containers API, but not jobs API. + raise SourceLine(docker_req, "dockerOutputDirectory", UnsupportedRequirement).makeError( + "Option 'dockerOutputDirectory' of DockerRequirement not supported.") arv_docker_get_image(arvrunner.api, docker_req, True, arvrunner.project_uuid) + else: + arv_docker_get_image(arvrunner.api, {"dockerPull": "arvados/jobs"}, True, arvrunner.project_uuid) elif isinstance(tool, cwltool.workflow.Workflow): for s in tool.steps: upload_docker(arvrunner, s.embedded_tool) -def upload_instance(arvrunner, name, tool, job_order): - upload_docker(arvrunner, tool) +def packed_workflow(arvrunner, tool): + """Create a packed workflow. + + A "packed" workflow is one where all the components have been combined into a single document.""" + + return pack(tool.doc_loader, tool.doc_loader.fetch(tool.tool["id"]), + tool.tool["id"], tool.metadata) + +def tag_git_version(packed): + if tool.tool["id"].startswith("file://"): + path = os.path.dirname(tool.tool["id"][7:]) + try: + githash = subprocess.check_output(['git', 'log', '--first-parent', '--max-count=1', '--format=%H'], stderr=subprocess.STDOUT, cwd=path).strip() + except (OSError, subprocess.CalledProcessError): + pass + else: + packed["http://schema.org/version"] = githash + + +def discover_secondary_files(inputs, job_order): + for t in inputs: + def setSecondary(fileobj): + if isinstance(fileobj, dict) and fileobj.get("class") == "File": + if "secondaryFiles" not in fileobj: + fileobj["secondaryFiles"] = [{"location": substitute(fileobj["location"], sf), "class": "File"} for sf in t["secondaryFiles"]] + + if isinstance(fileobj, list): + for e in fileobj: + setSecondary(e) + + if shortname(t["id"]) in job_order and t.get("secondaryFiles"): + setSecondary(job_order[shortname(t["id"])]) + +def upload_job_order(arvrunner, name, tool, job_order): + """Upload local files referenced in the input object and return updated input + object with 'location' updated to the proper keep references. + """ + + discover_secondary_files(tool.tool["inputs"], job_order) - workflowmapper = upload_dependencies(arvrunner, - name, - tool.doc_loader, - tool.tool, - tool.tool["id"], - True) + jobmapper = upload_dependencies(arvrunner, + name, + tool.doc_loader, + job_order, + job_order.get("id", "#"), + False) - jobmapper = upload_dependencies(arvrunner, - os.path.basename(job_order.get("id", "#")), - tool.doc_loader, - job_order, - job_order.get("id", "#"), - False) + if "id" in job_order: + del job_order["id"] - if "id" in job_order: - del job_order["id"] + # Need to filter this out, gets added by cwltool when providing + # parameters on the command line. + if "job_order" in job_order: + del job_order["job_order"] - return workflowmapper + return job_order + +def upload_workflow_deps(arvrunner, tool, override_tools): + # Ensure that Docker images needed by this workflow are available + + upload_docker(arvrunner, tool) + + document_loader = tool.doc_loader + + def upload_tool_deps(deptool): + if "id" in deptool: + upload_dependencies(arvrunner, + "%s dependencies" % (shortname(deptool["id"])), + document_loader, + deptool, + deptool["id"], + False, + include_primary=False) + document_loader.idx[deptool["id"]] = deptool + override_tools[deptool["id"]] = json.dumps(deptool) + + tool.visit(upload_tool_deps) + +def arvados_jobs_image(arvrunner, img): + """Determine if the right arvados/jobs image version is available. If not, try to pull and upload it.""" -def arvados_jobs_image(arvrunner): - img = "arvados/jobs:"+__version__ try: arv_docker_get_image(arvrunner.api, {"dockerPull": img}, True, arvrunner.project_uuid) except Exception as e: raise Exception("Docker image %s is not available\n%s" % (img, e) ) return img +def upload_workflow_collection(arvrunner, name, packed): + collection = arvados.collection.Collection(api_client=arvrunner.api, + keep_client=arvrunner.keep_client, + num_retries=arvrunner.num_retries) + with collection.open("workflow.cwl", "w") as f: + f.write(json.dumps(packed, indent=2, sort_keys=True, separators=(',',': '))) + + filters = [["portable_data_hash", "=", collection.portable_data_hash()], + ["name", "like", name+"%"]] + if arvrunner.project_uuid: + filters.append(["owner_uuid", "=", arvrunner.project_uuid]) + exists = arvrunner.api.collections().list(filters=filters).execute(num_retries=arvrunner.num_retries) + + if exists["items"]: + logger.info("Using collection %s", exists["items"][0]["uuid"]) + else: + collection.save_new(name=name, + owner_uuid=arvrunner.project_uuid, + ensure_unique_name=True, + num_retries=arvrunner.num_retries) + logger.info("Uploaded to %s", collection.manifest_locator()) + + return collection.portable_data_hash() + + class Runner(object): - def __init__(self, runner, tool, job_order, enable_reuse, output_name): + """Base class for runner processes, which submit an instance of + arvados-cwl-runner and wait for the final result.""" + + def __init__(self, runner, tool, job_order, enable_reuse, + output_name, output_tags, submit_runner_ram=0, + name=None, on_error=None, submit_runner_image=None, + intermediate_output_ttl=0): self.arvrunner = runner self.tool = tool self.job_order = job_order self.running = False + if enable_reuse: + # If reuse is permitted by command line arguments but + # disabled by the workflow itself, disable it. + reuse_req, _ = get_feature(self.tool, "http://arvados.org/cwl#ReuseRequirement") + if reuse_req: + enable_reuse = reuse_req["enableReuse"] self.enable_reuse = enable_reuse self.uuid = None self.final_output = None self.output_name = output_name + self.output_tags = output_tags + self.name = name + self.on_error = on_error + self.jobs_image = submit_runner_image or "arvados/jobs:"+__version__ + self.intermediate_output_ttl = intermediate_output_ttl + + if submit_runner_ram: + self.submit_runner_ram = submit_runner_ram + else: + self.submit_runner_ram = 3000 + + if self.submit_runner_ram <= 0: + raise Exception("Value of --submit-runner-ram must be greater than zero") def update_pipeline_component(self, record): pass - def arvados_job_spec(self, *args, **kwargs): - self.name = os.path.basename(self.tool.tool["id"]) - workflowmapper = upload_instance(self.arvrunner, self.name, self.tool, self.job_order) - adjustDirObjs(self.job_order, trim_listing) - return workflowmapper - def done(self, record): - if record["state"] == "Complete": - if record.get("exit_code") is not None: - if record["exit_code"] == 33: - processStatus = "UnsupportedRequirement" - elif record["exit_code"] == 0: - processStatus = "success" + """Base method for handling a completed runner.""" + + try: + if record["state"] == "Complete": + if record.get("exit_code") is not None: + if record["exit_code"] == 33: + processStatus = "UnsupportedRequirement" + elif record["exit_code"] == 0: + processStatus = "success" + else: + processStatus = "permanentFail" else: - processStatus = "permanentFail" + processStatus = "success" else: - processStatus = "success" - else: - processStatus = "permanentFail" + processStatus = "permanentFail" - outputs = None - try: - try: - self.final_output = record["output"] - outc = arvados.collection.CollectionReader(self.final_output, + outputs = {} + + if processStatus == "permanentFail": + logc = arvados.collection.CollectionReader(record["log"], api_client=self.arvrunner.api, keep_client=self.arvrunner.keep_client, num_retries=self.arvrunner.num_retries) + done.logtail(logc, logger, "%s error log:" % self.arvrunner.label(self), maxlen=40) + + self.final_output = record["output"] + outc = arvados.collection.CollectionReader(self.final_output, + api_client=self.arvrunner.api, + keep_client=self.arvrunner.keep_client, + num_retries=self.arvrunner.num_retries) + if "cwl.output.json" in outc: with outc.open("cwl.output.json") as f: - outputs = json.load(f) - def keepify(fileobj): - path = fileobj["location"] - if not path.startswith("keep:"): - fileobj["location"] = "keep:%s/%s" % (record["output"], path) - adjustFileObjs(outputs, keepify) - adjustDirObjs(outputs, keepify) - except Exception as e: - logger.error("While getting final output object: %s", e) + if f.size() > 0: + outputs = json.load(f) + def keepify(fileobj): + path = fileobj["location"] + if not path.startswith("keep:"): + fileobj["location"] = "keep:%s/%s" % (record["output"], path) + adjustFileObjs(outputs, keepify) + adjustDirObjs(outputs, keepify) + except Exception as e: + logger.exception("[%s] While getting final output object: %s", self.name, e) + self.arvrunner.output_callback({}, "permanentFail") + else: self.arvrunner.output_callback(outputs, processStatus) finally: - del self.arvrunner.processes[record["uuid"]] + if record["uuid"] in self.arvrunner.processes: + del self.arvrunner.processes[record["uuid"]]