Fix CI lint and remove Python 3.11 workflow

Author: Cody McCodePants

.github/workflows/test.yml

@@ -18,7 +18,7 @@ jobs:
       - name: Set up Python
         uses: actions/setup-python@v5
         with:
-          python-version: "3.11"
+          python-version: "3.12"
 
       - name: Install lint dependencies
         run: |
@@ -39,7 +39,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python-version: ["3.11", "3.12", "3.13"]
+        python-version: ["3.12", "3.13"]
 
     steps:
       - name: Checkout

netopsbench/agents/tracing.py

@@ -6,7 +6,7 @@
 import threading
 import uuid
 from datetime import UTC, datetime
-from typing import Any
+from typing import Any, cast
 
 from netopsbench.agents._trace_utils import jsonable as _jsonable
 
@@ -110,7 +110,9 @@ def on_llm_end(self, response: Any, *, run_id: Any, parent_run_id: Any = None, *
                 del kwargs
                 recorder.record_llm_response(response, run_id=run_id, parent_run_id=parent_run_id)
 
-            def on_llm_error(self, error: BaseException, *, run_id: Any, parent_run_id: Any = None, **kwargs: Any) -> None:
+            def on_llm_error(
+                self, error: BaseException, *, run_id: Any, parent_run_id: Any = None, **kwargs: Any
+            ) -> None:
                 del kwargs
                 recorder.record_error(
                     stage="llm",
@@ -175,7 +177,9 @@ def record_llm_request(
             "provider": provider,
             "extra": {
                 "llm_request": {
-                    "messages": [_message_payload(message, index=index) for index, message in enumerate(messages or [], 1)]
+                    "messages": [
+                        _message_payload(message, index=index) for index, message in enumerate(messages or [], 1)
+                    ]
                 }
             },
         }
@@ -297,7 +301,9 @@ def record_tool_error(self, *, error: BaseException, run_id: Any, parent_run_id:
             step["ended_at"] = _isoformat(datetime.now(UTC))
             step["duration_seconds"] = _duration_seconds(step.get("started_at"), step.get("ended_at"))
 
-    def record_error(self, *, stage: str, error: BaseException | str, run_id: Any = None, parent_run_id: Any = None) -> None:
+    def record_error(
+        self, *, stage: str, error: BaseException | str, run_id: Any = None, parent_run_id: Any = None
+    ) -> None:
         if not self.enabled:
             return
         call_id = str(run_id) if run_id else None
@@ -399,7 +405,11 @@ async def chat(self, messages: list[dict[str, Any]], **kwargs: Any) -> Any:
         )
         client = AsyncOpenAI(api_key=self.api_key, base_url=self.base_url, **self.client_kwargs)
         try:
-            response = await client.chat.completions.create(model=self.model, messages=messages, **kwargs)
+            response = await client.chat.completions.create(
+                model=self.model,
+                messages=cast(Any, messages),
+                **kwargs,
+            )
         except Exception as exc:
             self.recorder.record_error(stage="llm", error=exc, run_id=run_id)
             raise

netopsbench/cli/trace.py

@@ -15,7 +15,9 @@ def add_trace_subparser(subparsers: argparse._SubParsersAction) -> None:
     trace_parser = subparsers.add_parser("trace", help="Inspect and export agent runtime traces")
     trace_sub = trace_parser.add_subparsers(dest="trace_action", required=True)
     trace_list = trace_sub.add_parser("list", help="List runs with trace artifacts")
-    trace_list.add_argument("--limit", type=int, default=20, help="Maximum number of runs to show. Default: %(default)s.")
+    trace_list.add_argument(
+        "--limit", type=int, default=20, help="Maximum number of runs to show. Default: %(default)s."
+    )
     trace_export = trace_sub.add_parser("export", help="Export run traces as a Harbor jobs directory")
     trace_export.add_argument("run_id", help="Run id, for example run-20260605T124040Z")
     trace_export.add_argument("--output", required=True, help="Output Harbor jobs directory")

netopsbench/platform/session/orchestrator.py

@@ -353,7 +353,7 @@ def callback(episode_result: dict) -> dict:
             except Exception as exc:
                 trace_recorder.record_error(stage="agent", error=exc)
                 ended_at = self._timestamp()
-                diagnosis_payload = {
+                diagnosis_payload: dict[str, Any] = {
                     "error": str(exc),
                     "success": False,
                     "time_taken_seconds": max(0.0, (ended_at - start_time).total_seconds()),

netopsbench/platform/session/tracing.py

@@ -680,8 +680,15 @@ def _metrics_payload(
 ) -> dict[str, Any]:
     recorder_metrics = trace_recorder.metrics() if trace_recorder is not None else {}
     payload = {
-        "time_taken_seconds": float(diagnosis_payload.get("time_taken_seconds") or max(0.0, (ended_at - started_at).total_seconds())),
-        "tool_calls_count": len((trace_recorder.tool_calls() if trace_recorder is not None else []) or diagnosis_payload.get("tool_calls") or metadata.get("tool_calls") or []),
+        "time_taken_seconds": float(
+            diagnosis_payload.get("time_taken_seconds") or max(0.0, (ended_at - started_at).total_seconds())
+        ),
+        "tool_calls_count": len(
+            (trace_recorder.tool_calls() if trace_recorder is not None else [])
+            or diagnosis_payload.get("tool_calls")
+            or metadata.get("tool_calls")
+            or []
+        ),
     }
     for key in ("input_tokens", "output_tokens", "total_tokens", "llm_call_count"):
         if recorder_metrics.get(key):
@@ -745,7 +752,9 @@ def _steps_from_tool_calls(tool_calls: Any) -> list[dict[str, Any]]:
             "name": (item if isinstance(item, dict) else {"tool": str(item)}).get("tool")
             or (item if isinstance(item, dict) else {}).get("name")
             or "tool",
-            "args": (item if isinstance(item, dict) else {}).get("args") or (item if isinstance(item, dict) else {}).get("input") or {},
+            "args": (item if isinstance(item, dict) else {}).get("args")
+            or (item if isinstance(item, dict) else {}).get("input")
+            or {},
         }
         for index, item in enumerate(tool_calls, 1)
     ]
@@ -756,9 +765,7 @@ def _final_diagnosis_step(diagnosis_payload: dict[str, Any], *, ended_at: dateti
     verdict = final.get("verdict") or ("error" if final.get("error") else "unknown")
     fault_type = final.get("fault_type")
     location = final.get("location") if isinstance(final.get("location"), dict) else {}
-    location_text = ", ".join(
-        str(value) for value in (location or {}).values() if value not in (None, "")
-    )
+    location_text = ", ".join(str(value) for value in (location or {}).values() if value not in (None, ""))
     parts = [f"Final diagnosis: {verdict}"]
     if fault_type:
         parts.append(f"fault_type={fault_type}")
@@ -810,7 +817,9 @@ def _matching_result_row(result_rows: list[dict[str, Any]], index_row: dict[str,
         if row.get("trace_id") == trace_id:
             return row
     for row in result_rows:
-        if row.get("scenario_id") == index_row.get("scenario_id") and row.get("episode_id") == index_row.get("episode_id"):
+        if row.get("scenario_id") == index_row.get("scenario_id") and row.get("episode_id") == index_row.get(
+            "episode_id"
+        ):
             return row
     return None
 
@@ -863,16 +872,16 @@ def _run_times(run_path: Path, index_rows: list[dict[str, Any]]) -> dict[str, st
     report_path = run_path / "report.json"
     if report_path.exists():
         try:
-            summary = (json.loads(report_path.read_text(encoding="utf-8")).get("summary") or {})
+            summary = json.loads(report_path.read_text(encoding="utf-8")).get("summary") or {}
             if summary.get("started_at") and summary.get("completed_at"):
                 return {
                     "started_at": _normalise_iso_z(summary["started_at"]),
                     "finished_at": _normalise_iso_z(summary["completed_at"]),
                 }
         except Exception:
             pass
-    starts = [row.get("started_at") for row in index_rows if row.get("started_at")]
-    ends = [row.get("ended_at") for row in index_rows if row.get("ended_at")]
+    starts = [str(value) for row in index_rows if (value := row.get("started_at"))]
+    ends = [str(value) for row in index_rows if (value := row.get("ended_at"))]
     now = _isoformat(datetime.now(UTC))
     return {
         "started_at": _normalise_iso_z(min(starts) if starts else now),

tests/test_api_sessions.py

@@ -327,9 +327,7 @@ def diagnose(self, context):
             context.trace.record_llm_request([{"role": "user", "content": "diagnose"}], run_id="llm-1")
             context.trace.record_llm_response(
                 SimpleNamespace(
-                    generations=[
-                        [SimpleNamespace(message=SimpleNamespace(type="ai", content="checking"))]
-                    ]
+                    generations=[[SimpleNamespace(message=SimpleNamespace(type="ai", content="checking"))]]
                 ),
                 run_id="llm-1",
             )

tests/test_commands_cli.py

@@ -357,9 +357,7 @@ def fake_launch(folder, *, host, port):
     assert "synced traces:" in out
     expected = tmp_path / ".netopsbench" / "harbor-jobs"
     assert launched == {"folder": expected, "host": "127.0.0.1", "port": "55668"}
-    assert (
-        expected / "netopsbench-run-20260605T124040Z" / "scenario-1__case-1" / "agent" / "trajectory.json"
-    ).exists()
+    assert (expected / "netopsbench-run-20260605T124040Z" / "scenario-1__case-1" / "agent" / "trajectory.json").exists()
 
 
 def test_cli_trace_list_shows_trace_runs(tmp_path, monkeypatch, capsys):
@@ -406,12 +404,8 @@ def fake_launch(folder, *, host, port):
     expected = tmp_path / ".netopsbench" / "harbor-jobs"
     assert "synced traces:" in out
     assert launched == {"folder": expected, "host": "127.0.0.1", "port": "8080-8089"}
-    assert (
-        expected / "netopsbench-run-20260605T124040Z" / "scenario-1__case-1" / "agent" / "trajectory.json"
-    ).exists()
-    assert (
-        expected / "netopsbench-run-20260605T123000Z" / "scenario-1__case-1" / "agent" / "trajectory.json"
-    ).exists()
+    assert (expected / "netopsbench-run-20260605T124040Z" / "scenario-1__case-1" / "agent" / "trajectory.json").exists()
+    assert (expected / "netopsbench-run-20260605T123000Z" / "scenario-1__case-1" / "agent" / "trajectory.json").exists()
 
 
 def test_cli_trace_view_without_run_id_syncs_all_trace_runs(tmp_path, monkeypatch, capsys):
@@ -441,12 +435,8 @@ def fake_launch(folder, *, host, port):
     expected = tmp_path / ".netopsbench" / "harbor-jobs"
     assert "synced traces:" in out
     assert launched == {"folder": expected, "host": "127.0.0.1", "port": "8080-8089"}
-    assert (
-        expected / "netopsbench-run-20260605T124040Z" / "scenario-1__case-1" / "agent" / "trajectory.json"
-    ).exists()
-    assert (
-        expected / "netopsbench-run-20260605T123000Z" / "scenario-1__case-1" / "agent" / "trajectory.json"
-    ).exists()
+    assert (expected / "netopsbench-run-20260605T124040Z" / "scenario-1__case-1" / "agent" / "trajectory.json").exists()
+    assert (expected / "netopsbench-run-20260605T123000Z" / "scenario-1__case-1" / "agent" / "trajectory.json").exists()
 
 
 def test_cli_benchmark_prepare_runs_topology_then_scenario_generation(tmp_path, monkeypatch, capsys):

tests/test_example_agents.py

@@ -251,7 +251,7 @@ def test_minimal_deepagent_parses_final_json_block(monkeypatch):
                 SimpleNamespace(
                     type="ai",
                     content=(
-                        '```json\n'
+                        "```json\n"
                         '{"verdict":"fault_detected","fault_type":"link_down",'
                         '"location":{"device":"leaf1","interface":"Ethernet8"}}\n'
                         "```"

tests/test_session_tracing.py

@@ -52,7 +52,9 @@ def __init__(self, **kwargs):
 def test_disabled_trace_recorder_preserves_api_without_collecting():
     recorder = AgentTraceRecorder.disabled()
     run_id = recorder.record_llm_request([{"role": "user", "content": "diagnose"}], model="gpt-test")
-    recorder.record_llm_response(SimpleNamespace(generations=[[SimpleNamespace(message=SimpleNamespace(content="ok"))]]), run_id=run_id)
+    recorder.record_llm_response(
+        SimpleNamespace(generations=[[SimpleNamespace(message=SimpleNamespace(content="ok"))]]), run_id=run_id
+    )
     recorder.record_tool_start(name="get_topology", args={"verbose": True}, run_id="tool-1")
     recorder.record_tool_end(output={"ok": True}, run_id="tool-1")
     recorder.record_error(stage="agent", error=RuntimeError("boom"))
@@ -66,7 +68,9 @@ def test_disabled_trace_recorder_preserves_api_without_collecting():
 
 def test_trace_recorder_preserves_tool_call_llm_response_payload():
     recorder = AgentTraceRecorder()
-    run_id = recorder.record_llm_request([{"role": "user", "content": "inspect topology"}], model="deepseek-v4-pro", provider="deepseek")
+    run_id = recorder.record_llm_request(
+        [{"role": "user", "content": "inspect topology"}], model="deepseek-v4-pro", provider="deepseek"
+    )
 
     recorder.record_llm_response(
         SimpleNamespace(
@@ -277,7 +281,13 @@ def test_export_traces_writes_harbor_jobs_directory(tmp_path):
         "episodes": [
             {
                 "episode": {"episode_id": "ep1"},
-                "diagnosis": {"trace": {"trace_id": trace_result.trace_id, "case_id": trace_result.case_id, "atif_path": trace_result.atif_path}},
+                "diagnosis": {
+                    "trace": {
+                        "trace_id": trace_result.trace_id,
+                        "case_id": trace_result.case_id,
+                        "atif_path": trace_result.atif_path,
+                    }
+                },
             }
         ]
     }