From 40776afe5558796fca41e8753abac5f3de87ef58 Mon Sep 17 00:00:00 2001
From: Michael Thornton <mbthornton@lbl.gov>
Date: Mon, 10 Mar 2025 15:12:10 -0700
Subject: [PATCH] update create_job_rec and add unit test

---
 nmdc_automation/workflow_automation/sched.py |  7 +++++-
 tests/test_sched.py                          | 23 ++++++++++++++++++++
 2 files changed, 29 insertions(+), 1 deletion(-)

diff --git a/nmdc_automation/workflow_automation/sched.py b/nmdc_automation/workflow_automation/sched.py
index e1b30a36..849f3061 100644
--- a/nmdc_automation/workflow_automation/sched.py
+++ b/nmdc_automation/workflow_automation/sched.py
@@ -129,7 +129,12 @@ def create_job_rec(self, job: SchedulerJob):
                         continue
                     raise ValueError(f"Unable to find {do_type} in {do_by_type}")
                 input_data_objects.append(dobj.as_dict())
-                v = dobj["url"]
+
+                if k == "input_files":
+                    v = [dobj["url"]]
+                else:
+                    v = dobj["url"]
+                    
             # TODO: Make this smarter
             elif v == "{was_informed_by}":
                 v = job.informed_by
diff --git a/tests/test_sched.py b/tests/test_sched.py
index a9e46e49..55ba6304 100644
--- a/tests/test_sched.py
+++ b/tests/test_sched.py
@@ -266,4 +266,27 @@ def test_scheduler_find_new_jobs(test_db, mock_api, workflows_config_dir, site_c
     assert job_req["config"]["input_data_objects"]
 
 
+def test_scheduler_create_job_rec_has_input_files_as_array(test_db, mock_api, workflows_config_dir, site_config_file):
+    """
+    Test that the input_data_objects field is an array of strings.
+    """
+    reset_db(test_db)
+    load_fixture(test_db, "data_object_set.json")
+    load_fixture(test_db, "data_generation_set.json")
+    load_fixture(test_db, "read_qc_analysis.json", col="workflow_execution_set")
+
+    jm = Scheduler(
+        test_db, workflow_yaml=workflows_config_dir / "workflows.yaml",
+        site_conf=site_config_file
+        )
+
+    resp = jm.cycle()
+    assemblies = [j for j in resp if j["config"]["activity"]["type"] == "nmdc:MetagenomeAssembly"]
+    assert assemblies
+    assembly = assemblies[0]
+
+    assert isinstance(assembly["config"]["inputs"]["shortRead"], bool)
+    assert assembly["config"]["inputs"]["shortRead"] == False
+    assert isinstance(assembly["config"]["inputs"]["input_files"], list)
+