+from future import standard_library
+standard_library.install_aliases()
+from builtins import str
+from past.builtins import basestring
# Copyright (C) The Arvados Authors. All rights reserved.
#
# SPDX-License-Identifier: Apache-2.0
import logging
import uuid
import os
-import urllib
+import urllib.request, urllib.parse, urllib.error
-from arvados_cwl.util import get_current_container, get_intermediate_collection_info
+import arvados_cwl.util
import arvados.commands.run
import arvados.collection
src = src[:src.index("#")]
if isinstance(src, basestring) and ArvPathMapper.pdh_dirpath.match(src):
- self._pathmap[src] = MapperEnt(src, self.collection_pattern % urllib.unquote(src[5:]), srcobj["class"], True)
+ self._pathmap[src] = MapperEnt(src, self.collection_pattern % urllib.parse.unquote(src[5:]), srcobj["class"], True)
debug = logger.isEnabledFor(logging.DEBUG)
if isinstance(st, arvados.commands.run.UploadFile):
uploadfiles.add((src, ab, st))
elif isinstance(st, arvados.commands.run.ArvFile):
- self._pathmap[src] = MapperEnt(st.fn, self.collection_pattern % urllib.unquote(st.fn[5:]), "File", True)
+ self._pathmap[src] = MapperEnt(st.fn, self.collection_pattern % urllib.parse.unquote(st.fn[5:]), "File", True)
else:
raise WorkflowException("Input file path '%s' is invalid" % st)
elif src.startswith("_:"):
else:
raise SourceLine(obj, "location", WorkflowException).makeError("Don't know what to do with '%s'" % obj["location"])
+ def needs_new_collection(self, srcobj, prefix=""):
+ """Check if files need to be staged into a new collection.
+
+ If all the files are in the same collection and in the same
+ paths they would be staged to, return False. Otherwise, a new
+ collection is needed with files copied/created in the
+ appropriate places.
+ """
+
+ loc = srcobj["location"]
+ if loc.startswith("_:"):
+ return True
+ if prefix:
+ if loc != prefix+srcobj["basename"]:
+ return True
+ else:
+ i = loc.rfind("/")
+ if i > -1:
+ prefix = loc[:i+1]
+ else:
+ prefix = loc+"/"
+ if srcobj["class"] == "File" and loc not in self._pathmap:
+ return True
+ for s in srcobj.get("secondaryFiles", []):
+ if self.needs_new_collection(s, prefix):
+ return True
+ if srcobj.get("listing"):
+ prefix = "%s%s/" % (prefix, srcobj["basename"])
+ for l in srcobj["listing"]:
+ if self.needs_new_collection(l, prefix):
+ return True
+ return False
+
def setup(self, referenced_files, basedir):
# type: (List[Any], unicode) -> None
uploadfiles = set()
packed=False)
for src, ab, st in uploadfiles:
- self._pathmap[src] = MapperEnt(urllib.quote(st.fn, "/:+@"), self.collection_pattern % st.fn[5:],
+ self._pathmap[src] = MapperEnt(urllib.parse.quote(st.fn, "/:+@"), self.collection_pattern % st.fn[5:],
"Directory" if os.path.isdir(ab) else "File", True)
for srcobj in referenced_files:
for l in srcobj.get("listing", []):
self.addentry(l, c, ".", remap)
- container = get_current_container(self.arvrunner.api, self.arvrunner.num_retries, logger)
- info = get_intermediate_collection_info(None, container, self.arvrunner.intermediate_output_ttl)
+ container = arvados_cwl.util.get_current_container(self.arvrunner.api, self.arvrunner.num_retries, logger)
+ info = arvados_cwl.util.get_intermediate_collection_info(None, container, self.arvrunner.intermediate_output_ttl)
c.save_new(name=info["name"],
owner_uuid=self.arvrunner.project_uuid,
elif srcobj["class"] == "File" and (srcobj.get("secondaryFiles") or
(srcobj["location"].startswith("_:") and "contents" in srcobj)):
+ # If all secondary files/directories are located in
+ # the same collection as the primary file and the
+ # paths and names that are consistent with staging,
+ # don't create a new collection.
+ if not self.needs_new_collection(srcobj):
+ continue
+
c = arvados.collection.Collection(api_client=self.arvrunner.api,
keep_client=self.arvrunner.keep_client,
num_retries=self.arvrunner.num_retries )
self.addentry(srcobj, c, ".", remap)
- container = get_current_container(self.arvrunner.api, self.arvrunner.num_retries, logger)
- info = get_intermediate_collection_info(None, container, self.arvrunner.intermediate_output_ttl)
+ container = arvados_cwl.util.get_current_container(self.arvrunner.api, self.arvrunner.num_retries, logger)
+ info = arvados_cwl.util.get_intermediate_collection_info(None, container, self.arvrunner.intermediate_output_ttl)
c.save_new(name=info["name"],
owner_uuid=self.arvrunner.project_uuid,
ab, "File", True)
if srcobj.get("secondaryFiles"):
ab = self.collection_pattern % c.portable_data_hash()
- self._pathmap["_:" + unicode(uuid.uuid4())] = MapperEnt("keep:"+c.portable_data_hash(), ab, "Directory", True)
+ self._pathmap["_:" + str(uuid.uuid4())] = MapperEnt("keep:"+c.portable_data_hash(), ab, "Directory", True)
if remap:
for loc, sub in remap:
# with any secondary files.
self.visitlisting(referenced_files, self.stagedir, basedir)
- for path, (ab, tgt, type, staged) in self._pathmap.items():
+ for path, (ab, tgt, type, staged) in list(self._pathmap.items()):
if type in ("File", "Directory") and ab.startswith("keep:"):
self._pathmap[path] = MapperEnt("$(task.keep)/%s" % ab[5:], tgt, type, staged)
projects / arvados.git / blobdiff