Add some more linters

examples/image_dataloading.py

@@ -299,7 +299,7 @@ def entrypoint(args: list[str] | None = None):
     ns, args = _parse_process_args(args)
 
     args_set = [
-        args + [f"--worker-id={i}", f"--num-workers={ns.num_workers}"]
+        [*args, f"--worker-id={i}", f"--num-workers={ns.num_workers}"]
         for i in range(ns.num_workers)
     ]
 

examples/video_dataloading.py

@@ -380,7 +380,7 @@ def entrypoint(args: list[str] | None = None):
     ns, args = _parse_process_args(args)
 
     args_set = [
-        args + [f"--worker-id={i}", f"--num-workers={ns.num_workers}"]
+        [*args, f"--worker-id={i}", f"--num-workers={ns.num_workers}"]
         for i in range(ns.num_workers)
     ]
 

pyproject.toml

@@ -15,3 +15,10 @@ extend-exclude = [
   "third_party",
   "setup.py",
 ]
+
+[tool.ruff.lint]
+extend-select = [
+  "B",    # flake8-bugbear
+  "PGH",  # pygrep-hooks
+  "RUF",  # Ruff-specific
+]

src/spdl/dataloader/__init__.py

@@ -10,7 +10,7 @@
 
 from typing import Any
 
-from . import _builder, _flist, _hook, _pipeline, _utils  # noqa: E402
+from . import _builder, _flist, _hook, _pipeline, _utils
 
 _mods = [
     _builder,
@@ -27,7 +27,7 @@ def __dir__():
 
 
 def __getattr__(name: str) -> Any:
-    for mod in _mods + [_flist]:
+    for mod in [*_mods, _flist]:
         if name in mod.__all__:
             return getattr(mod, name)
 

src/spdl/dataloader/_builder.py

@@ -846,7 +846,7 @@ def _get_desc(self) -> list[str]:
         return parts
 
     def __str__(self) -> str:
-        return "\n".join([repr(self)] + self._get_desc())
+        return "\n".join([repr(self), *self._get_desc()])
 
     def build(self, *, num_threads: int | None = None) -> Pipeline:
         """Build the pipeline.

src/spdl/dataloader/_pipeline.py

@@ -261,7 +261,7 @@ def __init__(
     ):
         self._queues = queues
 
-        self._str = "\n".join([repr(self)] + desc)
+        self._str = "\n".join([repr(self), *desc])
 
         self._output_queue = queues[-1]
         self._event_loop = _EventLoop(coro, num_threads)
@@ -281,14 +281,15 @@ def __del__(self) -> None:
         """Stop the pipeline if running."""
         if _EventLoopState.STARTED <= self._event_loop_state < _EventLoopState.STOPPED:
             warnings.warn(
-                f"Pipeline ({repr(self)}) is running in the background, but "
+                f"Pipeline ({self!r}) is running in the background, but "
                 "there is no valid reference pointing the foreground object. "
                 "Stopping the background thread. "
                 "It is strongly advised to stop the pipeline explicity, "
                 "using the `auto_stop` context manager. "
                 "If you are using a framework and you cannot use the "
                 "context manager, try calling `stop` in done callback and "
-                "error callback."
+                "error callback.",
+                stacklevel=1,
             )
             self.stop()
 

src/spdl/io/__init__.py

@@ -19,15 +19,14 @@
 # (black, flake8 and pyre) and documentation tools, this seems like
 # the simplest solution.
 # This import is just for annotation, so please overlook this one.
-from ._type_stub import *  # noqa
+from ._type_stub import *  # noqa: F403
 
-from . import (  # noqa: E402
+from . import (
     _composite,
     _config,
     _convert,
     _core,
     _preprocessing,
-    _type_stub,
 )
 
 
@@ -37,7 +36,6 @@
     _convert,
     _core,
     _preprocessing,
-    _type_stub,
 ]
 
 

tests/spdl_unittest/cuda/nvdec_async_decoding_test.py

@@ -78,7 +78,7 @@ def test_batch_decode_image(get_samples):
     cmd = "ffmpeg -hide_banner -y -f lavfi -i testsrc -frames:v 10 sample_%03d.jpg"
     samples = get_samples(cmd)
 
-    flist = ["NON_EXISTING_FILE.JPG"] + samples
+    flist = ["NON_EXISTING_FILE.JPG", *samples]
 
     async def _test():
         buffer = await spdl.io.async_load_image_batch_nvdec(

tests/spdl_unittest/dataloader/pipeline_test.py

@@ -57,7 +57,7 @@ def test_async_enqueue_simple():
     coro = _enqueue(src, queue)
     asyncio.run(coro)
     vals = _flush_aqueue(queue)
-    assert vals == src + [_EOF]
+    assert vals == [*src, _EOF]
 
 
 def test_async_enqueue_skip():
@@ -72,7 +72,7 @@ def src():
     coro = _enqueue(src(), queue)
     asyncio.run(coro)
     vals = _flush_aqueue(queue)
-    assert vals == list(range(6)) + [_EOF]
+    assert vals == [*list(range(6)), _EOF]
 
 
 def test_async_enqueue_iterator_failure():
@@ -650,15 +650,15 @@ def test_task_stats():
 
     async def _test():
         async with hook.stage_hook():
-            for i in range(3):
+            for _ in range(3):
                 async with hook.task_hook():
                     await asyncio.sleep(0.5)
 
             assert hook.num_tasks == 3
             assert hook.num_success == 3
             assert 0.3 < hook.ave_time < 0.7
 
-            for i in range(2):
+            for _ in range(2):
                 with pytest.raises(RuntimeError):
                     async with hook.task_hook():
                         await asyncio.sleep(1.0)

tests/spdl_unittest/io/async_test.py

@@ -233,7 +233,7 @@ def test_batch_decode_image(get_samples):
     cmd = "ffmpeg -hide_banner -y -f lavfi -i testsrc -frames:v 250 sample_%03d.jpg"
     samples = get_samples(cmd)
 
-    flist = ["NON_EXISTING_FILE.JPG"] + samples
+    flist = ["NON_EXISTING_FILE.JPG", *samples]
 
     async def _test():
         buffer = await spdl.io.async_load_image_batch(