Fix CI regressions for cache benchmark work

benchmarks/claude_session_mode_benchmark.py

@@ -11,7 +11,7 @@ import logging
 import os
 from collections import Counter
 from dataclasses import asdict, dataclass, field
-from datetime import UTC, datetime, timedelta
+from datetime import datetime, timedelta, timezone
 from pathlib import Path
 from typing import Any
 
@@ -236,10 +236,10 @@ def decode_project_key(project_key: str) -> str:
 
 def _parse_timestamp(value: str | None) -> datetime:
     if not value:
-        return datetime.min.replace(tzinfo=UTC)
+        return datetime.min.replace(tzinfo=timezone.utc)
     if value.endswith("Z"):
         value = value[:-1] + "+00:00"
-    return datetime.fromisoformat(value).astimezone(UTC)
+    return datetime.fromisoformat(value).astimezone(timezone.utc)
 
 
 def _canonical_block_key(block: Any) -> str:

tests/test_provider_model_fallback.py

@@ -151,16 +151,14 @@ class TestAnthropicConfigLoading:
         """Test loading config from file path in env var."""
         config = {"context_limits": {"file-model": 400000}}
 
-        with tempfile.NamedTemporaryFile(mode="w", suffix=".json", delete=False) as f:
-            json.dump(config, f)
-            f.flush()
+        with tempfile.TemporaryDirectory() as tmpdir:
+            config_path = Path(tmpdir) / "model_limits.json"
+            config_path.write_text(json.dumps(config))
 
-            with patch.dict(os.environ, {"HEADROOM_MODEL_LIMITS": f.name}):
+            with patch.dict(os.environ, {"HEADROOM_MODEL_LIMITS": str(config_path)}):
                 loaded = anthropic_load_config()
                 assert loaded["context_limits"]["file-model"] == 400000
 
-            os.unlink(f.name)
-
     def test_load_from_config_file(self):
         """Test loading from ~/.headroom/models.json."""
         config = {
@@ -291,16 +289,14 @@ class TestOpenAIConfigLoading:
         """Test loading pricing from config."""
         config = {"openai": {"pricing": {"test-model": [5.0, 15.0]}}}
 
-        with tempfile.NamedTemporaryFile(mode="w", suffix=".json", delete=False) as f:
-            json.dump(config, f)
-            f.flush()
+        with tempfile.TemporaryDirectory() as tmpdir:
+            config_path = Path(tmpdir) / "model_limits.json"
+            config_path.write_text(json.dumps(config))
 
-            with patch.dict(os.environ, {"HEADROOM_MODEL_LIMITS": f.name}):
+            with patch.dict(os.environ, {"HEADROOM_MODEL_LIMITS": str(config_path)}):
                 loaded = openai_load_config()
                 assert loaded["pricing"]["test-model"] == [5.0, 15.0]
 
-            os.unlink(f.name)
-
 
 class TestCrossProviderConsistency:
     """Tests for consistency across providers."""
@@ -328,10 +324,8 @@ class TestCrossProviderConsistency:
         anthropic.get_context_limit("claude-future-model-xyz")
         openai.get_context_limit("gpt-future-model-xyz")
 
-    def test_both_providers_warn_for_unknown_models(self, caplog):
+    def test_both_providers_warn_for_unknown_models(self):
         """Test that both providers warn for unknown models."""
-        import logging
-
         # Clear warning caches
         from headroom.providers import anthropic as anthropic_module
         from headroom.providers import openai as openai_module
@@ -339,12 +333,17 @@ class TestCrossProviderConsistency:
         anthropic_module._UNKNOWN_MODEL_WARNINGS.clear()
         openai_module._UNKNOWN_MODEL_WARNINGS.clear()
 
-        with caplog.at_level(logging.WARNING):
+        with (
+            patch.object(anthropic_module.logger, "warning") as anthropic_warning,
+            patch.object(openai_module.logger, "warning") as openai_warning,
+        ):
             anthropic = AnthropicProvider()
             anthropic.get_context_limit("claude-test-unknown-model")
 
             openai = OpenAIProvider()
             openai.get_context_limit("gpt-test-unknown-model")
 
-        assert "claude-test-unknown-model" in caplog.text
-        assert "gpt-test-unknown-model" in caplog.text
+        anthropic_warning.assert_called_once()
+        openai_warning.assert_called_once()
+        assert "claude-test-unknown-model" in anthropic_warning.call_args.args[0]
+        assert "gpt-test-unknown-model" in openai_warning.call_args.args[0]

tests/test_proxy_openai_cache_stability.py

@@ -49,8 +49,8 @@ def test_openai_cache_mode_freezes_previous_turns() -> None:
         proxy.config.mode = "cache"
 
         fake_tracker = _FakePrefixTracker(frozen_count=0)
-        proxy.session_tracker_store.compute_session_id = (
-            lambda request, model, messages: "stable-session"
+        proxy.session_tracker_store.compute_session_id = lambda request, model, messages: (
+            "stable-session"
         )
         proxy.session_tracker_store.get_or_create = lambda session_id, provider: fake_tracker
 
@@ -110,8 +110,8 @@ def test_openai_cache_mode_restores_mutated_frozen_prefix() -> None:
         proxy.config.mode = "cache"
 
         fake_tracker = _FakePrefixTracker(frozen_count=0)
-        proxy.session_tracker_store.compute_session_id = (
-            lambda request, model, messages: "stable-session"
+        proxy.session_tracker_store.compute_session_id = lambda request, model, messages: (
+            "stable-session"
         )
         proxy.session_tracker_store.get_or_create = lambda session_id, provider: fake_tracker