Move: shared judge test base out of one test file

engine/hooks/llm-judge/judge_test_base.py

@@ -0,0 +1,27 @@
+from __future__ import annotations
+
+import json
+import os
+import tempfile
+import unittest
+from unittest.mock import patch
+
+import judge
+
+
+class JudgeTestCase(unittest.TestCase):
+    def setUp(self):
+        super().setUp()
+        self.state = tempfile.TemporaryDirectory()
+        self.judge_env = patch.dict(os.environ, {judge.STATE_ENV: self.state.name})
+        self.judge_env.start()
+        os.environ.pop(judge.CHILD_ENV, None)
+        os.environ.pop(judge.RUNNERS_ENV, None)
+
+    def tearDown(self):
+        self.judge_env.stop()
+        self.state.cleanup()
+        super().tearDown()
+
+    def use_runners(self, *entries):
+        os.environ[judge.RUNNERS_ENV] = json.dumps(list(entries))

engine/hooks/llm-judge/tests/test_judge.py

@@ -6,7 +6,6 @@
 import json
 import os
 import sys
-import tempfile
 import time
 import unittest
 import warnings
@@ -16,6 +15,7 @@
 sys.path.insert(0, LIB_DIR)
 
 import judge  # noqa: E402
+from judge_test_base import JudgeTestCase  # noqa: E402
 
 PY = sys.executable
 
@@ -31,21 +31,7 @@ def runner(name, script):
 SLOW_MATCH = runner("slow", "import json, sys, time; time.sleep(2); print(json.dumps({'match': True, 'prompt': sys.argv[1]}))")
 
 
-class JudgeTestCase(unittest.TestCase):
-    def setUp(self):
-        self.state = tempfile.TemporaryDirectory()
-        self.env = patch.dict(os.environ, {judge.STATE_ENV: self.state.name})
-        self.env.start()
-        os.environ.pop(judge.CHILD_ENV, None)
-        os.environ.pop(judge.RUNNERS_ENV, None)
-
-    def tearDown(self):
-        self.env.stop()
-        self.state.cleanup()
-
-    def use_runners(self, *entries):
-        os.environ[judge.RUNNERS_ENV] = json.dumps(list(entries))
-
+class JudgeBehaviorTestCase(JudgeTestCase):
     def job(self, **overrides):
         base = {
             "id": "job-1",
@@ -59,7 +45,7 @@ def job(self, **overrides):
         return base
 
 
-class TestAsk(JudgeTestCase):
+class TestAsk(JudgeBehaviorTestCase):
     def test_first_runner_fails_second_answers_and_one_failed_attempt_is_recorded(self):
         self.use_runners(EXIT_NONZERO, ANSWER_MATCH)
         result = judge.ask("hello judge")
@@ -125,7 +111,7 @@ def test_default_runner_order_is_codex_then_claude_then_cursor(self):
         self.assertEqual([name for name, _ in judge.runners()], ["codex", "claude", "cursor"])
 
 
-class TestVerdict(JudgeTestCase):
+class TestVerdict(JudgeBehaviorTestCase):
     def test_hit_when_every_hit_key_is_true(self):
         job = self.job(hit_if_all_true=["match", "sure"])
         result = judge.verdict(job, {"outcome": "answered", "runner": "answers", "answer": {"match": True, "sure": True}, "attempts": []})
@@ -148,7 +134,7 @@ def test_unchecked_when_ask_was_unchecked(self):
         self.assertEqual(result["attempts"], attempts)
 
 
-class TestBackground(JudgeTestCase):
+class TestBackground(JudgeBehaviorTestCase):
     def test_enqueue_as_judge_child_returns_none_and_starts_nothing(self):
         os.environ[judge.CHILD_ENV] = "1"
         self.use_runners(ANSWER_MATCH)