supervisor: timing based test state logic

distinguish between "tests-pending" and "tests-not-running" based on
a 6-hour timeout from when tests were triggered.

- updated prompt: LLM interprets test_trigger_method from QE database
  and applies 6 hour rule using last_status_transition_timestamp.

Author: panyamkeerthana

supervisor/testing_analyst.py

@@ -51,9 +51,16 @@ def render_prompt(input: InputSchema) -> str:
       NEWA will post a comment to the erratum when it has started tests and when they finish.
       Read the JIRA issue in those comments to find test results.
 
+      Tests can trigger at various points in an issue's lifecycle depending on component
+      configuration, but always by the time the erratum moves to QE status. If the erratum
+      is in QE status, and its last_status_transition_timestamp is more than 6 hours ago,
+      and there's no evidence from erratum comments of tests running or completed, then assume
+      tests will not run automatically and return tests-not-running.
+
       Call the final_answer tool passing in the state and a comment as follows.
       The comment should use JIRA comment syntax.
 
+
       If the tests need to be started manually:
          state: tests-not-running
          comment: [explain what needs to be done to start tests]
@@ -73,6 +80,10 @@ def render_prompt(input: InputSchema) -> str:
       If the tests are currently running:
          state: tests-running
          comment: [Provide a brief description of what tests are running and where the results will be]
+
+      If tests have not started or completed when they should have (as described above):
+         state: tests-not-running
+         comment: [Explain the situation and that manual intervention is needed]
     """
     return PromptTemplate(
         PromptTemplateInput(schema=InputSchema, template=template)