Merge master to output-tags branch and resolve conflict

[arvados.git] / sdk / cwl / arvados_cwl / __init__.py

diff --git a/sdk/cwl/arvados_cwl/__init__.py b/sdk/cwl/arvados_cwl/__init__.py
index 92be92d6e0469fba63a1504a5f0c834fb4a9b2b7..b3d47dd8d05e5981ae4f645fa9c968ee7707e747 100644 (file)
--- a/sdk/cwl/arvados_cwl/__init__.py
+++ b/sdk/cwl/arvados_cwl/__init__.py
@@ -49,7 +49,7 @@ class ArvCwlRunner(object):
 
     """
 
-    def __init__(self, api_client, work_api=None, keep_client=None, output_name=None):
+    def __init__(self, api_client, work_api=None, keep_client=None, output_name=None, output_tags=None):
         self.api = api_client
         self.processes = {}
         self.lock = threading.Lock()
@@ -64,6 +64,7 @@ class ArvCwlRunner(object):
         self.pipeline = None
         self.final_output_collection = None
         self.output_name = output_name
+        self.output_tags = output_tags
         self.project_uuid = None
 
         if keep_client is not None:
@@ -183,7 +184,7 @@ class ArvCwlRunner(object):
             for v in obj:
                 self.check_writable(v)
 
-    def make_output_collection(self, name, outputObj):
+    def make_output_collection(self, name, tagsString, outputObj):
         outputObj = copy.deepcopy(outputObj)
 
         files = []
@@ -248,6 +249,13 @@ class ArvCwlRunner(object):
                     final.api_response()["name"],
                     final.manifest_locator())
 
+        final_uuid = final.manifest_locator()
+        tags = tagsString.split(',')
+        for tag in tags:
+             self.api.links().create(body={
+                "head_uuid": final_uuid, "link_class": "tag", "name": tag
+                }).execute(num_retries=self.num_retries)
+
         def finalcollection(fileobj):
             fileobj["location"] = "keep:%s/%s" % (final.portable_data_hash(), fileobj["location"])
 
@@ -325,9 +333,9 @@ class ArvCwlRunner(object):
                                          self.output_callback,
                                          **kwargs).next()
                 else:
-                    runnerjob = RunnerContainer(self, tool, job_order, kwargs.get("enable_reuse"), self.output_name)
+                    runnerjob = RunnerContainer(self, tool, job_order, kwargs.get("enable_reuse"), self.output_name, self.output_tags)
             else:
-                runnerjob = RunnerJob(self, tool, job_order, kwargs.get("enable_reuse"), self.output_name)
+                runnerjob = RunnerJob(self, tool, job_order, kwargs.get("enable_reuse"), self.output_name, self.output_tags)
 
         if not kwargs.get("submit") and "cwl_runner_job" not in kwargs and not self.work_api == "containers":
             # Create pipeline for local run
@@ -410,7 +418,9 @@ class ArvCwlRunner(object):
         else:
             if self.output_name is None:
                 self.output_name = "Output of %s" % (shortname(tool.tool["id"]))
-            self.final_output, self.final_output_collection = self.make_output_collection(self.output_name, self.final_output)
+            if self.output_tags is None:
+                self.output_tags = ""
+            self.final_output, self.final_output_collection = self.make_output_collection(self.output_name, self.output_tags, self.final_output)
             self.set_crunch_output()
 
         if self.final_status != "success":
@@ -468,6 +478,7 @@ def arg_parser():  # type: () -> argparse.ArgumentParser
 
     parser.add_argument("--project-uuid", type=str, metavar="UUID", help="Project that will own the workflow jobs, if not provided, will go to home project.")
     parser.add_argument("--output-name", type=str, help="Name to use for collection that stores the final output.", default=None)
+    parser.add_argument("--output-tags", type=str, help="Tags for the final output collection separated by commas, e.g., '--output-tags tag0,tag1,tag2'.", default=None)
     parser.add_argument("--ignore-docker-for-reuse", action="store_true",
                         help="Ignore Docker image version when deciding whether to reuse past jobs.",
                         default=False)
@@ -525,7 +536,7 @@ def main(args, stdout, stderr, api_client=None, keep_client=None):
     try:
         if api_client is None:
             api_client=arvados.api('v1', model=OrderedJsonModel())
-        runner = ArvCwlRunner(api_client, work_api=arvargs.work_api, keep_client=keep_client, output_name=arvargs.output_name)
+        runner = ArvCwlRunner(api_client, work_api=arvargs.work_api, keep_client=keep_client, output_name=arvargs.output_name, output_tags=arvargs.output_tags)
     except Exception as e:
         logger.error(e)
         return 1