Merge branch '16139-cwl-secondaryfiles-fix'
authorPeter Amstutz <peter.amstutz@curii.com>
Mon, 10 Feb 2020 15:14:05 +0000 (10:14 -0500)
committerPeter Amstutz <peter.amstutz@curii.com>
Mon, 10 Feb 2020 15:14:05 +0000 (10:14 -0500)
fixes #16139

Arvados-DCO-1.1-Signed-off-by: Peter Amstutz <peter.amstutz@curii.com>

sdk/cwl/arvados_cwl/executor.py
sdk/cwl/arvados_cwl/runner.py
sdk/cwl/setup.py
sdk/cwl/tests/test_submit.py
sdk/cwl/tests/tool/blub.txt.cat [new file with mode: 0644]
sdk/cwl/tests/tool/tool_with_sf.cwl [new file with mode: 0644]
sdk/cwl/tests/tool/tool_with_sf.yml [new file with mode: 0644]
sdk/python/setup.py

index 406ebfd2da064df383105b8e0a7c8f4e7b19a529..99d4c4e9a10a883abc54ce0fe1cbc476af7c7692 100644 (file)
@@ -521,10 +521,10 @@ The 'jobs' API is no longer supported.
             for req in job_reqs:
                 tool.requirements.append(req)
 
-    def arv_executor(self, tool, job_order, runtimeContext, logger=None):
+    def arv_executor(self, updated_tool, job_order, runtimeContext, logger=None):
         self.debug = runtimeContext.debug
 
-        tool.visit(self.check_features)
+        updated_tool.visit(self.check_features)
 
         self.project_uuid = runtimeContext.project_uuid
         self.pipeline = None
@@ -545,16 +545,20 @@ The 'jobs' API is no longer supported.
             raise Exception("--submit-request-uuid requires containers API, but using '{}' api".format(self.work_api))
 
         if not runtimeContext.name:
-            runtimeContext.name = self.name = tool.tool.get("label") or tool.metadata.get("label") or os.path.basename(tool.tool["id"])
+            runtimeContext.name = self.name = updated_tool.tool.get("label") or updated_tool.metadata.get("label") or os.path.basename(updated_tool.tool["id"])
 
         # Upload local file references in the job order.
         job_order = upload_job_order(self, "%s input" % runtimeContext.name,
-                                     tool, job_order)
+                                     updated_tool, job_order)
+
+        # the last clause means: if it is a command line tool, and we
+        # are going to wait for the result, and always_submit_runner
+        # is false, then we don't submit a runner process.
 
         submitting = (runtimeContext.update_workflow or
                       runtimeContext.create_workflow or
                       (runtimeContext.submit and not
-                       (tool.tool["class"] == "CommandLineTool" and
+                       (updated_tool.tool["class"] == "CommandLineTool" and
                         runtimeContext.wait and
                         not runtimeContext.always_submit_runner)))
 
@@ -564,8 +568,11 @@ The 'jobs' API is no longer supported.
         if submitting:
             # Document may have been auto-updated. Reload the original
             # document with updating disabled because we want to
-            # submit the original document, not the auto-updated one.
-            tool = load_tool(tool.tool["id"], loadingContext)
+            # submit the document with its original CWL version, not
+            # the auto-updated one.
+            tool = load_tool(updated_tool.tool["id"], loadingContext)
+        else:
+            tool = updated_tool
 
         # Upload direct dependencies of workflow steps, get back mapping of files to keep references.
         # Also uploads docker images.
@@ -632,22 +639,23 @@ The 'jobs' API is no longer supported.
         if runtimeContext.submit:
             # Submit a runner job to run the workflow for us.
             if self.work_api == "containers":
-                if tool.tool["class"] == "CommandLineTool" and runtimeContext.wait and (not runtimeContext.always_submit_runner):
-                    runtimeContext.runnerjob = tool.tool["id"]
+                if submitting:
+                    tool = RunnerContainer(self, updated_tool,
+                                           tool, loadingContext, runtimeContext.enable_reuse,
+                                           self.output_name,
+                                           self.output_tags,
+                                           submit_runner_ram=runtimeContext.submit_runner_ram,
+                                           name=runtimeContext.name,
+                                           on_error=runtimeContext.on_error,
+                                           submit_runner_image=runtimeContext.submit_runner_image,
+                                           intermediate_output_ttl=runtimeContext.intermediate_output_ttl,
+                                           merged_map=merged_map,
+                                           priority=runtimeContext.priority,
+                                           secret_store=self.secret_store,
+                                           collection_cache_size=runtimeContext.collection_cache_size,
+                                           collection_cache_is_default=self.should_estimate_cache_size)
                 else:
-                    tool = RunnerContainer(self, tool, loadingContext, runtimeContext.enable_reuse,
-                                                self.output_name,
-                                                self.output_tags,
-                                                submit_runner_ram=runtimeContext.submit_runner_ram,
-                                                name=runtimeContext.name,
-                                                on_error=runtimeContext.on_error,
-                                                submit_runner_image=runtimeContext.submit_runner_image,
-                                                intermediate_output_ttl=runtimeContext.intermediate_output_ttl,
-                                                merged_map=merged_map,
-                                                priority=runtimeContext.priority,
-                                                secret_store=self.secret_store,
-                                                collection_cache_size=runtimeContext.collection_cache_size,
-                                                collection_cache_is_default=self.should_estimate_cache_size)
+                    runtimeContext.runnerjob = tool.tool["id"]
 
         if runtimeContext.cwl_runner_job is not None:
             self.uuid = runtimeContext.cwl_runner_job.get('uuid')
index 19a6dd98b332c6dbc8363e989104a075cf90f587..2239e0f9df952b9ab75a7e9a96ab46953ab29f94 100644 (file)
@@ -578,7 +578,8 @@ class Runner(Process):
     """Base class for runner processes, which submit an instance of
     arvados-cwl-runner and wait for the final result."""
 
-    def __init__(self, runner, tool, loadingContext, enable_reuse,
+    def __init__(self, runner, updated_tool,
+                 tool, loadingContext, enable_reuse,
                  output_name, output_tags, submit_runner_ram=0,
                  name=None, on_error=None, submit_runner_image=None,
                  intermediate_output_ttl=0, merged_map=None,
@@ -587,10 +588,9 @@ class Runner(Process):
                  collection_cache_is_default=True):
 
         loadingContext = loadingContext.copy()
-        loadingContext.metadata = loadingContext.metadata.copy()
-        loadingContext.metadata["cwlVersion"] = INTERNAL_VERSION
+        loadingContext.metadata = updated_tool.metadata.copy()
 
-        super(Runner, self).__init__(tool.tool, loadingContext)
+        super(Runner, self).__init__(updated_tool.tool, loadingContext)
 
         self.arvrunner = runner
         self.embedded_tool = tool
index aa68933c6abc17f2d19860c2cd838a3026546257..62ceab2fa11c1895983cd518fb2d4370a49da986 100644 (file)
@@ -60,7 +60,7 @@ setup(name='arvados-cwl-runner',
       ],
       test_suite='tests',
       tests_require=[
-          'mock>=1.0',
+          'mock>=1.0,<4',
           'subprocess32>=3.5.1',
       ],
       zip_safe=True
index 927e43ad76c1fe1e547fd91dc026282726c82a1c..397ae142253d26b41d11aee03586fbc3e352bb31 100644 (file)
@@ -388,6 +388,19 @@ class TestSubmit(unittest.TestCase):
                          stubs.expect_container_request_uuid + '\n')
         self.assertEqual(exited, 0)
 
+
+    @stubs
+    def test_submit_container_tool(self, stubs):
+        # test for issue #16139
+        exited = arvados_cwl.main(
+            ["--submit", "--no-wait", "--api=containers", "--debug",
+                "tests/tool/tool_with_sf.cwl", "tests/tool/tool_with_sf.yml"],
+            stubs.capture_stdout, sys.stderr, api_client=stubs.api, keep_client=stubs.keep_client)
+
+        self.assertEqual(stubs.capture_stdout.getvalue(),
+                         stubs.expect_container_request_uuid + '\n')
+        self.assertEqual(exited, 0)
+
     @stubs
     def test_submit_container_no_reuse(self, stubs):
         exited = arvados_cwl.main(
diff --git a/sdk/cwl/tests/tool/blub.txt.cat b/sdk/cwl/tests/tool/blub.txt.cat
new file mode 100644 (file)
index 0000000..d7c4221
--- /dev/null
@@ -0,0 +1 @@
+clipper clupper
diff --git a/sdk/cwl/tests/tool/tool_with_sf.cwl b/sdk/cwl/tests/tool/tool_with_sf.cwl
new file mode 100644 (file)
index 0000000..0beb7ad
--- /dev/null
@@ -0,0 +1,24 @@
+# Copyright (C) The Arvados Authors. All rights reserved.
+#
+# SPDX-License-Identifier: Apache-2.0
+
+# Test case for arvados-cwl-runner
+#
+# Used to test whether scanning a tool file for dependencies (e.g. default
+# value blub.txt) and uploading to Keep works as intended.
+
+class: CommandLineTool
+cwlVersion: v1.0
+requirements:
+  - class: DockerRequirement
+    dockerPull: debian:8
+inputs:
+  - id: x
+    type: File
+    secondaryFiles:
+      - .cat
+    inputBinding:
+      valueFrom: $(self.path).cat
+      position: 1
+outputs: []
+baseCommand: cat
diff --git a/sdk/cwl/tests/tool/tool_with_sf.yml b/sdk/cwl/tests/tool/tool_with_sf.yml
new file mode 100644 (file)
index 0000000..3f79d57
--- /dev/null
@@ -0,0 +1,3 @@
+x:
+  class: File
+  location: blub.txt
index 87977c2187e014017a2d76c4314ee3b1eede6b9c..ff68e2a7fd8888d10d56b96e13a5a9bf2690a177 100644 (file)
@@ -64,6 +64,6 @@ setup(name='arvados-python-client',
           'Programming Language :: Python :: 3',
       ],
       test_suite='tests',
-      tests_require=['pbr<1.7.0', 'mock>=1.0', 'PyYAML'],
+      tests_require=['pbr<1.7.0', 'mock>=1.0,<4', 'PyYAML'],
       zip_safe=False
       )