X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/419013258ddead1ece05327e96017222f4aa7de8..f72b0e8bcc350966ce54954711bed538c527eb00:/sdk/cwl/arvados_cwl/pathmapper.py diff --git a/sdk/cwl/arvados_cwl/pathmapper.py b/sdk/cwl/arvados_cwl/pathmapper.py index e17264bcdd..bd71a13355 100644 --- a/sdk/cwl/arvados_cwl/pathmapper.py +++ b/sdk/cwl/arvados_cwl/pathmapper.py @@ -6,7 +6,7 @@ import os import arvados.commands.run import arvados.collection -from cwltool.pathmapper import PathMapper, MapperEnt, abspath +from cwltool.pathmapper import PathMapper, MapperEnt, abspath, adjustFileObjs, adjustDirObjs from cwltool.workflow import WorkflowException logger = logging.getLogger('arvados.cwl-runner') @@ -42,8 +42,11 @@ class ArvPathMapper(PathMapper): uploadfiles.add((src, ab, st)) elif isinstance(st, arvados.commands.run.ArvFile): self._pathmap[src] = MapperEnt(ab, st.fn, "File") - elif src.startswith("_:") and "contents" in srcobj: - pass + elif src.startswith("_:"): + if "contents" in srcobj: + pass + else: + raise WorkflowException("File literal '%s' is missing contents" % src) else: raise WorkflowException("Input file path '%s' is invalid" % st) if "secondaryFiles" in srcobj: @@ -52,12 +55,14 @@ class ArvPathMapper(PathMapper): elif srcobj["class"] == "Directory": if isinstance(src, basestring) and ArvPathMapper.pdh_dirpath.match(src): self._pathmap[src] = MapperEnt(src, self.collection_pattern % src[5:], "Directory") - for l in srcobj["listing"]: + for l in srcobj.get("listing", []): self.visit(l, uploadfiles) def addentry(self, obj, c, path, subdirs): if obj["location"] in self._pathmap: src, srcpath = self.arvrunner.fs_access.get_collection(self._pathmap[obj["location"]].resolved) + if srcpath == "": + srcpath = "." c.copy(srcpath, path + "/" + obj["basename"], source_collection=src, overwrite=True) for l in obj.get("secondaryFiles", []): self.addentry(l, c, path, subdirs) @@ -145,11 +150,11 @@ class InitialWorkDirPathMapper(PathMapper): def visit(self, obj, stagedir, basedir, copy=False): # type: (Dict[unicode, Any], unicode, unicode, bool) -> None + loc = obj["location"] if obj["class"] == "Directory": - self._pathmap[obj["location"]] = MapperEnt(obj["location"], stagedir, "Directory") + self._pathmap[loc] = MapperEnt(obj["location"], stagedir, "Directory") self.visitlisting(obj.get("listing", []), stagedir, basedir) elif obj["class"] == "File": - loc = obj["location"] if loc in self._pathmap: return tgt = os.path.join(stagedir, obj["basename"]) @@ -167,10 +172,26 @@ class InitialWorkDirPathMapper(PathMapper): # Go through each file and set the target to its own directory along # with any secondary files. - stagedir = self.stagedir - for fob in referenced_files: - self.visit(fob, stagedir, basedir) + self.visitlisting(referenced_files, self.stagedir, basedir) for path, (ab, tgt, type) in self._pathmap.items(): if type in ("File", "Directory") and ab.startswith("keep:"): self._pathmap[path] = MapperEnt("$(task.keep)/%s" % ab[5:], tgt, type) + + +class FinalOutputPathMapper(PathMapper): + def visit(self, obj, stagedir, basedir, copy=False): + # type: (Dict[unicode, Any], unicode, unicode, bool) -> None + loc = obj["location"] + if obj["class"] == "Directory": + self._pathmap[loc] = MapperEnt(loc, stagedir, "Directory") + elif obj["class"] == "File": + if loc in self._pathmap: + return + tgt = os.path.join(stagedir, obj["basename"]) + self._pathmap[loc] = MapperEnt(loc, tgt, "File") + self.visitlisting(obj.get("secondaryFiles", []), stagedir, basedir) + + def setup(self, referenced_files, basedir): + # type: (List[Any], unicode) -> None + self.visitlisting(referenced_files, self.stagedir, basedir)