import unittest import tempfile import os import shutil from datetime import datetime, timedelta from src.log_registry import LogRegistry from src.log_pruner import LogPruner class TestLogPruningHeuristic(unittest.TestCase): def setUp(self) -> None: self.temp_dir = tempfile.TemporaryDirectory() self.registry_path = os.path.join(self.temp_dir.name, "registry.toml") self.logs_dir = os.path.join(self.temp_dir.name, "logs", "sessions") os.makedirs(self.logs_dir, exist_ok=True) self.registry = LogRegistry(self.registry_path) self.pruner = LogPruner(self.registry, self.logs_dir) def tearDown(self) -> None: self.temp_dir.cleanup() def _create_session(self, session_id, start_time, message_count=None, size_kb=None, whitelisted=False): path = os.path.join(self.logs_dir, session_id) os.makedirs(path, exist_ok=True) # Create a dummy file with open(os.path.join(path, "comms.log"), "w") as f: f.write("test content") self.registry.register_session(session_id, path, start_time) if message_count is not None or size_kb is not None or whitelisted: self.registry.update_session_metadata( session_id, message_count=message_count if message_count is not None else 10, errors=0, size_kb=size_kb if size_kb is not None else 10, whitelisted=whitelisted, reason="Test" ) def test_get_old_non_whitelisted_sessions_includes_empty_sessions(self) -> None: now = datetime.now() cutoff_time = now - timedelta(days=7) # 1. Old, not whitelisted (should be included) self._create_session("old_nw", now - timedelta(days=10)) # 2. Recent, not whitelisted, but empty (message_count=0) (SHOULD be included based on new heuristic) self._create_session("recent_empty_msgs", now - timedelta(days=1), message_count=0) # 3. Recent, not whitelisted, but empty (size_kb=0) (SHOULD be included based on new heuristic) self._create_session("recent_empty_size", now - timedelta(days=1), size_kb=0) # 4. Recent, not whitelisted, NOT empty (should NOT be included) self._create_session("recent_not_empty", now - timedelta(days=1), message_count=5, size_kb=5) # 5. Old, whitelisted (should NOT be included) self._create_session("old_w", now - timedelta(days=10), whitelisted=True) sessions = self.registry.get_old_non_whitelisted_sessions(cutoff_time) session_ids = {s['session_id'] for s in sessions} self.assertIn("old_nw", session_ids) self.assertIn("recent_empty_msgs", session_ids) self.assertIn("recent_empty_size", session_ids) self.assertNotIn("recent_not_empty", session_ids) self.assertNotIn("old_w", session_ids) def test_get_old_non_whitelisted_sessions_includes_sessions_without_metadata(self) -> None: now = datetime.now() cutoff_time = now - timedelta(days=7) # Recent, not whitelisted, NO metadata (should be included) # _create_session without message_count/size_kb will leave metadata=None self._create_session("recent_no_metadata", now - timedelta(days=1)) sessions = self.registry.get_old_non_whitelisted_sessions(cutoff_time) session_ids = {s['session_id'] for s in sessions} self.assertIn("recent_no_metadata", session_ids) def test_prune_removes_empty_sessions_regardless_of_age(self) -> None: now = datetime.now() # Create a session that is recent but empty session_id = "recent_empty" session_path = os.path.join(self.logs_dir, session_id) os.makedirs(session_path, exist_ok=True) # Actual file size 0 with open(os.path.join(session_path, "comms.log"), "w") as f: pass self.registry.register_session(session_id, session_path, now - timedelta(hours=1)) self.registry.update_session_metadata(session_id, message_count=0, errors=0, size_kb=0, whitelisted=False, reason="Empty") self.assertTrue(os.path.exists(session_path)) # Prune with max_age_days=30 (so 1 hour old is NOT "old" by age) self.pruner.prune(max_age_days=30, min_size_kb=1) self.assertFalse(os.path.exists(session_path)) self.assertNotIn(session_id, self.registry.data) def test_prune_removes_sessions_without_metadata_regardless_of_age(self) -> None: now = datetime.now() session_id = "recent_no_metadata_to_prune" session_path = os.path.join(self.logs_dir, session_id) os.makedirs(session_path, exist_ok=True) # Actual file size 0 with open(os.path.join(session_path, "comms.log"), "w") as f: pass self.registry.register_session(session_id, session_path, now - timedelta(hours=1)) # NO metadata update self.assertTrue(os.path.exists(session_path)) # Prune with max_age_days=30 self.pruner.prune(max_age_days=30, min_size_kb=1) self.assertFalse(os.path.exists(session_path)) self.assertNotIn(session_id, self.registry.data) def test_prune_handles_relative_paths_starting_with_logs(self) -> None: now = datetime.now() # Create a session with a relative path session_id = "rel_path_session" # In this test, self.logs_dir = temp_dir/logs/sessions # project_root will be resolved to temp_dir # So the relative path starting with 'logs/' should be 'logs/rel_path_session' rel_path = f"logs/{session_id}" abs_path = os.path.join(self.temp_dir.name, "logs", session_id) os.makedirs(abs_path, exist_ok=True) with open(os.path.join(abs_path, "comms.log"), "w") as f: f.write("small") # Tiny file # Register with the RELATIVE path self.registry.register_session(session_id, rel_path, now - timedelta(days=10)) self.registry.update_session_metadata(session_id, message_count=1, errors=0, size_kb=0, whitelisted=False, reason="Test") self.assertTrue(os.path.exists(abs_path)) # Prune. It should resolve 'logs/rel_path_session' relative to temp_dir.name self.pruner.prune(max_age_days=1, min_size_kb=1) self.assertFalse(os.path.exists(abs_path)) self.assertNotIn(session_id, self.registry.data) if __name__ == '__main__': unittest.main()