fix(lint): resolve all 15 ruff errors across src/ and tests/

- E501: wrap long line in cli.py
- TC003: move Path import to TYPE_CHECKING block in hub.py
- F401: add missing assertions for all imported names in test_init.py
- F841: remove unused variable assignments in test files
- SIM105/SIM117: use contextlib.suppress and combined with statement

Author: sacredvoid

autoresearch-results.tsv

@@ -0,0 +1,18 @@
+# metric_direction: higher_is_better
+iteration	commit	metric	delta	guard	status	description
+0	97c61a0	54	0.0	pass	baseline	initial state — coverage 54%
+1	3663cbf	56	+2.0	pass	keep	add tests for _format_gsm8k_prompt and GRPORunner save/init branches
+2	bc6f2da	56	+0.4	pass	keep	add tests for ensure_chat_template — config.py now 100%
+3	4efbb5c	58	+2.0	pass	keep	add SFTRunner/DPORunner save/init and config branch tests
+4	-	58	0.0	-	discard	CLI tests failed — patched wrong import paths for lazy imports
+5	55c676a	66	+8.0	pass	keep	add CLI tests with correct source-module patching — cli.py 49%->99%
+6	6b52cc2	70	+4.0	pass	keep	add EvalRunner save_results, evaluate_all_stages, evaluate tests — eval.py 57%->85%
+7	2e6ebd1	75	+5.0	pass	keep	add ModelServer dispatch/load/generate tests — inference.py 40%->85%, eval.py->100%
+8	a9538e3	79	+4.0	pass	keep	add demo.py tests with mocked gradio — demo.py 0%->84%
+9	5f5deb5	80	+1.0	pass	keep	add rewards edge case tests — rewards.py 95%->100%
+10	fdd7191	86	+6.0	pass	keep	add mocked SFTRunner train/load/save tests — sft.py 48%->95%
+11	7739774	91	+5.0	pass	keep	add mocked DPORunner tests — dpo.py 50%->100%
+12	da80fa5	97	+6.0	pass	keep	add mocked GRPORunner tests — grpo.py 54%->100%, TOTAL 97%
+13	f172b24	98	+1.0	pass	keep	add _load_unsloth/_generate_unsloth tests — inference.py 85%->100%
+14	df08faf	99	+1.0	pass	keep	cover inner functions — sft.py->100%, demo.py 84%->94%, TOTAL 99%
+15	f7dd8f0	99	+0.2	pass	keep	cover cli.py __main__ via runpy — cli.py->100%

src/alignrl/cli.py

@@ -109,7 +109,9 @@ def main() -> None:
     eval_p.add_argument("--model", default="Qwen/Qwen2.5-3B")
     eval_p.add_argument("--adapter", default=None, help="Path to LoRA adapter")
     eval_p.add_argument("--stage", default="base", help="Stage label (base/sft/grpo/dpo)")
-    eval_p.add_argument("--tasks", default=None, help="Comma-separated task list (overrides preset)")
+    eval_p.add_argument(
+        "--tasks", default=None, help="Comma-separated task list (overrides preset)"
+    )
     eval_p.add_argument(
         "--preset",
         default=None,

src/alignrl/hub.py

@@ -2,7 +2,10 @@
 
 from __future__ import annotations
 
-from pathlib import Path
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from pathlib import Path
 
 
 def push_adapter(

tests/test_cli.py

@@ -126,15 +126,15 @@ def test_eval_creates_output(self, tmp_path) -> None:
 class TestMainEntry:
     def test_main_module_execution(self) -> None:
         """Test if __name__ == '__main__' block in cli.py."""
-        with patch("alignrl.cli.main") as mock_main, \
-             patch("alignrl.cli.__name__", "__main__"):
-            # Simulate running as __main__
-            import runpy
-
-            try:
-                runpy.run_module("alignrl.cli", run_name="__main__", alter_sys=True)
-            except SystemExit:
-                pass  # main() will exit without args
+        import contextlib
+        import runpy
+
+        with (
+            patch("alignrl.cli.main"),
+            patch("alignrl.cli.__name__", "__main__"),
+            contextlib.suppress(SystemExit),
+        ):
+            runpy.run_module("alignrl.cli", run_name="__main__", alter_sys=True)
 
 
 class TestCmdServe:

tests/test_demo.py

@@ -16,7 +16,7 @@ def test_creates_gradio_app(self) -> None:
         mock_server.generate.return_value = "test output"
 
         with patch.dict("sys.modules", {"gradio": mock_gr}), \
-             patch("alignrl.demo.ModelServer", return_value=mock_server) as mock_ms, \
+             patch("alignrl.demo.ModelServer", return_value=mock_server), \
              patch("alignrl.demo.InferenceConfig"):
             mock_server.load = MagicMock()
             app = create_demo(stages={"base": None}, model_name="test-model")

tests/test_dpo.py

@@ -106,7 +106,7 @@ def test_load_dataset(self) -> None:
         mock_datasets.load_dataset.return_value = mock_ds
 
         with patch.dict("sys.modules", {"datasets": mock_datasets}):
-            result = runner._load_dataset()
+            runner._load_dataset()
             mock_ds.select.assert_called_once()
             mock_ds.map.assert_called_once()
 

tests/test_grpo.py

@@ -105,7 +105,7 @@ def test_load_dataset(self) -> None:
         mock_datasets.load_dataset.return_value = mock_ds
 
         with patch.dict("sys.modules", {"datasets": mock_datasets}):
-            result = runner._load_dataset()
+            runner._load_dataset()
             mock_datasets.load_dataset.assert_called_once_with(
                 cfg.dataset_name, cfg.dataset_config, split=cfg.dataset_split
             )

tests/test_hub.py

@@ -91,7 +91,7 @@ def test_dpo_push_adapter(self) -> None:
         runner = DPORunner(cfg)
 
         with patch("alignrl.hub.push_adapter", return_value="https://huggingface.co/u/m") as mock:
-            url = runner.push_to_hub("u/m")
+            runner.push_to_hub("u/m")
             mock.assert_called_once()
 
     def test_grpo_push_adapter(self) -> None:
@@ -101,5 +101,5 @@ def test_grpo_push_adapter(self) -> None:
         runner = GRPORunner(cfg)
 
         with patch("alignrl.hub.push_adapter", return_value="https://huggingface.co/u/m") as mock:
-            url = runner.push_to_hub("u/m")
+            runner.push_to_hub("u/m")
             mock.assert_called_once()

tests/test_inference.py

@@ -252,6 +252,6 @@ def test_generate_unsloth_zero_temp(self) -> None:
         server._model.generate.return_value = MagicMock()
         server._tokenizer.decode.return_value = "deterministic"
 
-        result = server._generate_unsloth([{"role": "user", "content": "hi"}])
+        server._generate_unsloth([{"role": "user", "content": "hi"}])
         call_kwargs = server._model.generate.call_args[1]
         assert call_kwargs["do_sample"] is False

tests/test_init.py

@@ -22,21 +22,25 @@ def test_lazy_import_dpo(self) -> None:
         from alignrl import DPOConfig, DPORunner
 
         assert DPOConfig is not None
+        assert DPORunner is not None
 
     def test_lazy_import_grpo(self) -> None:
         from alignrl import GRPOConfig, GRPORunner
 
         assert GRPOConfig is not None
+        assert GRPORunner is not None
 
     def test_lazy_import_eval(self) -> None:
         from alignrl import EvalConfig, EvalRunner
 
         assert EvalConfig is not None
+        assert EvalRunner is not None
 
     def test_lazy_import_inference(self) -> None:
         from alignrl import InferenceConfig, ModelServer, build_prompt
 
         assert InferenceConfig is not None
+        assert ModelServer is not None
         assert build_prompt is not None
 
     def test_lazy_import_types(self) -> None: