[Code scan] GDI sub-DAG polling treats queued runs as failures

[njzjz]

Source: Codex global repository scan of deepmodeling/dpti at commit b719828.
Project: DeepModeling Agent Code Scan

Problem
After triggering a sub-DAG, GDILoopSubDagController.wait_until_end() only treats success and running as non-failure states. Airflow DagRuns can legitimately be queued, scheduled, or briefly not visible immediately after trigger_dag(). The current polling loop raises AirflowFailException for these healthy transient states.

Code references

dpti/dpti/dags/dp_ti_gdi.py

Line 187 in b719828

def wait_until_end(self):

dpti/dpti/dags/dp_ti_gdi.py

Line 193 in b719828

elif dag_run_state == State.RUNNING:

dpti/dpti/dags/dp_ti_gdi.py

Line 197 in b719828

raise AirflowFailException(f"subdag dag_run fail dag_id:{self.dag_name}; run_id:{self.run_id};")

dpti/dpti/dags/dp_ti_gdi.py

Line 208 in b719828

c.trigger_dag(dag_id=self.dag_name, run_id=self.run_id,

Reproduction
Unit-test wait_until_end() with get_dag_run_state() returning State.QUEUED and then State.SUCCESS.

Expected result
Queued/scheduled/non-visible transient states should be polled until they become running/success/failure, instead of being treated as immediate failures.