[pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Author: pre-commit-ci[bot]

studio/backend/core/inference/llama_cpp.py

@@ -2045,6 +2045,8 @@ def _preempt_ram_disabled_in(args) -> bool:
         elif tok.startswith("--preempt-ram="):
             disabled = tok.split("=", 1)[1].strip() == "0"
     return disabled
+
+
 # Cap tool calls from a single TEXTUAL-fallback turn (mirrors the safetensors
 # loop). Structured delta.tool_calls are grammar-bounded by llama-server; text
 # parsed from content is not, so one runaway turn could fan out unbounded.
@@ -6741,7 +6743,6 @@ def _server_park_grace(self) -> bool:
         """
         try:
             from core.inference.llama_stats import scrape_llama_metrics
-
             metrics = scrape_llama_metrics(self.base_url, timeout_s = 3.0)
         except Exception:
             return False
@@ -6763,9 +6764,7 @@ def _server_park_event(line: str, preempt_policy = None) -> Optional[dict]:
         parked = line == _SERVER_PARKED_COMMENT
         if not parked and line != _SERVER_RESUMED_COMMENT:
             return None
-        hook = getattr(
-            preempt_policy, "on_server_parked" if parked else "on_server_resumed", None
-        )
+        hook = getattr(preempt_policy, "on_server_parked" if parked else "on_server_resumed", None)
         if hook is not None:
             try:
                 hook()

studio/backend/tests/test_server_side_preemption_handoff.py

@@ -142,7 +142,11 @@ def test_an_upstream_build_reports_it_absent(self, monkeypatch, tmp_path):
 # ------------------------------------------------------------------------ the controller
 
 
-def _fill(controller, n = 4, tokens = 2000):
+def _fill(
+    controller,
+    n = 4,
+    tokens = 2000,
+):
     """Register `n` decoding chats. No sweep runs here: `register` never plans, so the
     caller sees the first decision itself."""
     signals = []
@@ -157,7 +161,11 @@ class TestController:
     def test_server_mode_chooses_nobody_and_holds_nothing_back(self):
         controller = get_preemption_controller("server")
         controller.configure(
-            budget = 8192, kv_unified = True, draft_tokens = 2, slots = 4, batch_tokens = 2048,
+            budget = 8192,
+            kv_unified = True,
+            draft_tokens = 2,
+            slots = 4,
+            batch_tokens = 2048,
             server_mode = True,
         )
         assert controller.server_mode is True
@@ -175,7 +183,11 @@ def test_server_mode_chooses_nobody_and_holds_nothing_back(self):
     def test_studio_mode_is_unchanged(self):
         controller = get_preemption_controller("studio")
         controller.configure(
-            budget = 8192, kv_unified = True, draft_tokens = 2, slots = 4, batch_tokens = 2048,
+            budget = 8192,
+            kv_unified = True,
+            draft_tokens = 2,
+            slots = 4,
+            batch_tokens = 2048,
             server_mode = False,
         )
         signals = _fill(controller, n = 4, tokens = 2400)
@@ -227,7 +239,9 @@ def test_the_policy_marks_a_server_park_and_resume(self):
 
 
 def _delta(content: str) -> str:
-    return "data: " + json.dumps({"choices": [{"index": 0, "delta": {"content": content}}]}) + "\n\n"
+    return (
+        "data: " + json.dumps({"choices": [{"index": 0, "delta": {"content": content}}]}) + "\n\n"
+    )
 
 
 def _finish(reason: str = "stop") -> str:
@@ -273,7 +287,14 @@ def __init__(self, monkeypatch, chunks, *, server_preempts):
         recorder = self
 
         @contextlib.contextmanager
-        def fake_stream_with_retry(_client, _url, payload, _cancel_event, headers = None, **_kw):
+        def fake_stream_with_retry(
+            _client,
+            _url,
+            payload,
+            _cancel_event,
+            headers = None,
+            **_kw,
+        ):
             recorder.payloads.append(copy.deepcopy(payload))
             yield _FakeResponse(chunks)
 
@@ -346,9 +367,10 @@ def test_the_pause_is_shown_and_the_text_is_untouched(self, monkeypatch):
         text, marks = _client_view(events)
         assert text == "Once upon a time"
         assert marks == [("paused", len("Once upon")), ("resumed", len("Once upon"))]
-        assert policy.events == ["server-parked", "server-resumed"], (
-            "the ledger is told, and the Studio-side pause handshake never runs"
-        )
+        assert policy.events == [
+            "server-parked",
+            "server-resumed",
+        ], "the ledger is told, and the Studio-side pause handshake never runs"
         assert len(recorder.payloads) == 1, "nothing was re-opened: the server resumed in place"
 
     def test_a_policy_without_the_hooks_is_fine(self, monkeypatch):
@@ -379,7 +401,13 @@ def test_an_upstream_stream_without_comments_is_bytewise_todays(self, monkeypatc
         assert marks == []
 
     def test_a_park_before_the_first_token_is_shown_too(self, monkeypatch):
-        script = [": preempted\n\n", ": resumed\n\n", _delta("Hello"), _finish(), "data: [DONE]\n\n"]
+        script = [
+            ": preempted\n\n",
+            ": resumed\n\n",
+            _delta("Hello"),
+            _finish(),
+            "data: [DONE]\n\n",
+        ]
         recorder = _Recorder(monkeypatch, script, server_preempts = True)
         events = list(
             recorder.backend.generate_chat_completion(