Fix ruff errors and make the ruff step fail the Lint job

.github/workflows/lint.yaml

@@ -23,9 +23,7 @@ jobs:
       - name: Check lockfile
         run: uv lock --check
       - name: Lint with ruff
-        continue-on-error: true
         run: |
-          # Fail if codebase contains any of these issues
           uv run ruff check .
           uv run ruff format --check .
       - name: Static Typechecking with mypy

pyproject.toml

@@ -53,6 +53,14 @@ fallback-version = "0.0.0"
 line-length = 140
 src = ["src", "tests"]
 target-version = "py310"
+extend-exclude = [
+    # Notebook submissions, whose cells start with tags such as "#grade" that the formatter would change.
+    "tests/scenario_root/**/*.ipynb",
+    # These files test how the grader handles syntax errors, so they can't be parsed.
+    "tests/scenario_root/test_helper_modules_syntax_error/helper_function.py",
+    "tests/scenario_root/test_module_sandbox_init_error/student_code.py",
+    "tests/scenario_root/test_syntax_error_ungradable/student_code.py",
+]
 
 
 [tool.ruff.lint]
@@ -61,7 +69,6 @@ ignore = [
     "S101",   # flake8-bandit assert
     "S308",   # flake8-bandit suspicious-mark-safe-usage
     "E501",   # pycodestyle line-too-long
-    "PT004",
     "PT011",
     "PT013",
     "S603",
@@ -88,6 +95,13 @@ select = [
     "W",   # pycodestyle warnings
 ]
 
+[tool.ruff.lint.per-file-ignores]
+# numpy and pandas aren't installed in the lite image, so they're imported only when needed.
+"src/pytest_prairielearn_grader/json_utils.py" = ["PLC0415"]
+# Stand-ins for student code: they use names that the sandbox injects, import inside functions to
+# test import restrictions, and evaluate expressions to test the sandbox.
+"tests/scenario_root/**" = ["F821", "PLC0415", "B018"]
+
 [tool.ruff.lint.flake8-pytest-style]
 fixture-parentheses = false
 mark-parentheses = false

src/pytest_prairielearn_grader/_student_code_runner.py

@@ -268,7 +268,7 @@ def load(self, name: str) -> types.ModuleType:
         # Add it before running it so that helper modules can import each other, as in sys.modules.
         self._loaded[name] = module
         try:
-            exec(compile(source, file_name, "exec"), module.__dict__)
+            exec(compile(source, file_name, "exec"), module.__dict__)  # noqa: S102
         except Exception as e:
             del self._loaded[name]
             raise ImportError(f"Helper module '{name}' could not be loaded: {type(e).__name__}: {e}") from e

src/pytest_prairielearn_grader/config.py

@@ -378,9 +378,7 @@ def __post_init__(self) -> None:
         # Validate output_level
         valid_output_levels = ("none", "message", "traceback", "friendly")
         if self.output_level is not None and self.output_level not in valid_output_levels:
-            raise ValueError(
-                f"output_level must be one of {valid_output_levels} or None, got: {self.output_level!r}"
-            )
+            raise ValueError(f"output_level must be one of {valid_output_levels} or None, got: {self.output_level!r}")
 
     def to_dict(self) -> dict[str, Any]:
         """

src/pytest_prairielearn_grader/fixture.py

@@ -63,7 +63,7 @@ def _extract_notebook_code(notebook_path: Path, cell_tag: str | None = None) ->
         ImportError: If ``nbformat`` is not installed.
     """
     try:
-        import nbformat  # type: ignore[import-untyped]
+        import nbformat  # type: ignore[import-untyped]  # noqa: PLC0415
     except ImportError:
         raise ImportError(
             "nbformat is required for Jupyter notebook grading support. Install it with: pip install 'pytest-prairielearn-grader[notebook]'"
@@ -272,7 +272,8 @@ def try_drop_privileges() -> None:
 
     def _connect_to_server(self, timeout: float) -> None:
         """Read the host/port printed by the subprocess and connect the socket."""
-        assert self.process is not None and self.process.stdout is not None
+        assert self.process is not None
+        assert self.process.stdout is not None
         host, port = self.process.stdout.readline().decode().strip().split(",")
         self.student_socket = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
         self.student_socket.settimeout(timeout)

src/pytest_prairielearn_grader/utils.py

@@ -207,7 +207,7 @@ def deserialize_object_unsafe(base64_string: str) -> object:
     dilled_bytes = base64.b64decode(base64_encoded_bytes)
 
     # 3. Deserialize the object using dill
-    return dill.loads(dilled_bytes)
+    return dill.loads(dilled_bytes)  # noqa: S301
 
 
 def get_builtins(builtin_whitelist: list[str] | None) -> dict[str, Any]:
@@ -331,15 +331,15 @@ def drop_privileges(user_name: str) -> None:
         raise NotImplementedError("Dropping privileges is not supported on Windows.")
 
     # pwd module is only available on Unix systems
-    import pwd
+    import pwd  # noqa: PLC0415
 
     try:
         # Get the UID and GID for the target user (e.g., 'nobody')
         pwnam = pwd.getpwnam(user_name)
         target_uid = pwnam.pw_uid
         target_gid = pwnam.pw_gid
     except KeyError:
-        raise ValueError(f"User '{user_name}' not found.")
+        raise ValueError(f"User '{user_name}' not found.") from None
 
     # 1. Set the real and effective GID
     # NOTE: Set GID before UID, as changing UID may restrict GID changes

tests/scenario_root/test_fixture_exception/scenario.py

@@ -1,4 +1,5 @@
 import pytest
+
 from pytest_prairielearn_grader.fixture import StudentFixture
 
 

tests/scenario_root/test_function_not_found/student_code.py

@@ -23,4 +23,4 @@ def raises_import_error():
 
 
 def raises_name_error():
-    return undefined_variable  # noqa: F821
+    return undefined_variable

tests/scenario_root/test_initialization_timeout/scenario.py

@@ -1,10 +1,10 @@
-# Module-level initialization timeout (used as default)
-initialization_timeout = 0.5
-
 import pytest
 
 from pytest_prairielearn_grader.fixture import StudentFixture
 
+# Module-level initialization timeout (used as default)
+initialization_timeout = 0.5
+
 
 @pytest.mark.grading_data(name="uses_module_timeout", points=2)
 def test_with_module_timeout(sandbox: StudentFixture) -> None:

tests/scenario_root/test_matplotlib_plots/student_code.py

@@ -8,7 +8,7 @@
 def create_line_plot():
     """Create a simple line plot with proper labels and title."""
     serializer = MatplotlibSerializer()
-    fig, ax = serializer.subplots()
+    _fig, ax = serializer.subplots()
     x = np.linspace(0, 2 * np.pi, 100)
     y = np.sin(x)
     ax.plot(x, y, label="sin(x)", color="blue", linewidth=2)
@@ -23,7 +23,7 @@ def create_line_plot():
 def create_scatter_plot():
     """Create a scatter plot with random data."""
     serializer = MatplotlibSerializer()
-    fig, ax = serializer.subplots()
+    _fig, ax = serializer.subplots()
     np.random.seed(42)
     x = np.random.randn(50)
     y = 2 * x + np.random.randn(50) * 0.5
@@ -37,7 +37,7 @@ def create_scatter_plot():
 def create_bar_chart():
     """Create a bar chart."""
     serializer = MatplotlibSerializer()
-    fig, ax = serializer.subplots()
+    _fig, ax = serializer.subplots()
     categories = ["A", "B", "C", "D"]
     values = [23, 45, 56, 78]
     ax.bar(categories, values, color=["red", "green", "blue", "yellow"])

tests/scenario_root/test_quiz_2_1/scenario.py

@@ -67,7 +67,8 @@ def test_0(sandbox: StudentFixture, feedback: FeedbackFixture, data_json: DataFi
 
     assert st_result is not None, "Your function did not return any value. Ensure you have a return statement."
 
-    assert isinstance(st_result, list) and len(st_result) == 3, "Your function should return a tuple with exactly three elements."
+    assert isinstance(st_result, list), "Your function should return a tuple with exactly three elements."
+    assert len(st_result) == 3, "Your function should return a tuple with exactly three elements."
 
     points = 0
 

tests/scenario_root/test_quiz_2_1/setup_code.py

@@ -1,7 +1,8 @@
+import sympy as sp
+
+
 def not_allowed(*args, **kwargs):
     raise RuntimeError("Calling sympy.series is not allowed in this question. You can still use sympy to compute derivatives")
 
 
-import sympy as sp
-
 sp.series = not_allowed

tests/scenario_root/test_quiz_3_1/scenario.py

@@ -31,7 +31,7 @@ def compute_starting_parameters() -> tuple[list[str], list[ParameterDict]]:
             continue
         hours_set.add(hours)
         rng.shuffle(members)
-        sale_metadata.append({"hours": list(zip(members, hours)), "units_sold": rng.integers(100, 2001)})
+        sale_metadata.append({"hours": list(zip(members, hours, strict=True)), "units_sold": rng.integers(100, 2001)})
 
     return members, sale_metadata
 
@@ -64,7 +64,7 @@ def test_0(sandbox: StudentFixture, feedback: FeedbackFixture) -> None:
     ref_member_ratings, _, _ = compute_answers(members, sale_metadata)
 
     if student_member_ratings is not None:
-        if type(student_member_ratings) != dict:
+        if not isinstance(student_member_ratings, dict):
             feedback.add_message("member_ratings is not a dictionary")
         else:
             for member, rating in student_member_ratings.items():

tests/scenario_root/test_quiz_3_1/student_code_0.py

@@ -20,6 +20,6 @@ def solve_ratings(members, sale_metadata):
     ratings = la.solve(theMatrix, sales)
     highest_performing = np.max(ratings)
     lowest_performing = np.min(ratings)
-    member_ratings = dict(zip(members, ratings))
+    member_ratings = dict(zip(members, ratings, strict=True))
 
     return member_ratings, highest_performing, lowest_performing

tests/scenario_root/test_quiz_3_1/student_code_100.py

@@ -13,6 +13,6 @@ def solve_ratings(members, sale_metadata):
     ratings = la.solve(theMatrix, sales)
     highest_performing = members[np.argmax(ratings)]
     lowest_performing = members[np.argmin(ratings)]
-    member_ratings = dict(zip(members, ratings))
+    member_ratings = dict(zip(members, ratings, strict=True))
 
     return member_ratings, highest_performing, lowest_performing

tests/scenario_root/test_quiz_3_1/student_code_66.py

@@ -13,6 +13,6 @@ def solve_ratings(members, sale_metadata):
     ratings = la.solve(theMatrix, sales)
     highest_performing = np.max(ratings)
     lowest_performing = np.min(ratings)
-    member_ratings = dict(zip(members, ratings))
+    member_ratings = dict(zip(members, ratings, strict=True))
 
     return member_ratings, highest_performing, lowest_performing

tests/scenario_root/test_timeout/scenario.py

@@ -1,12 +1,12 @@
-# Module-level initialization timeout (used as default when no marker is present)
-initialization_timeout = 0.5
-
 import pytest
 
 from pytest_prairielearn_grader.fixture import FeedbackFixture
 from pytest_prairielearn_grader.fixture import StudentFixture
 from pytest_prairielearn_grader.utils import FunctionStatusCode
 
+# Module-level initialization timeout (used as default when no marker is present)
+initialization_timeout = 0.5
+
 
 @pytest.mark.grading_data(name="initialization_timeout", points=2)
 @pytest.mark.sandbox_timeout(0.05)

tests/scenario_root/test_workspace_function_not_found/student/calculator.py

@@ -14,4 +14,4 @@ def raises_import_error():
 
 
 def imports_missing_module():
-    import no_such_module_inside_function  # noqa: F401, PLC0415
+    import no_such_module_inside_function  # noqa: F401

tests/scenario_root/test_workspace_function_not_found/student/raises_attribute_error_on_load.py

@@ -1,2 +1,2 @@
 missing = None
-missing.attribute  # noqa: B018
+missing.attribute

tests/test_autograder_scenarios.py

@@ -1,3 +1,4 @@
+import json
 import math
 import platform
 import shutil
@@ -9,7 +10,6 @@
 
 pytest_plugins = ("pytester",)
 platform  # noqa: B018
-import json
 
 # --- Configuration ---
 # The root directory where your scenario folders are located
@@ -116,9 +116,7 @@ def test_autograder_scenario_with_pytester(pytester: pytest.Pytester, scenario_d
     # Check gradable field
     expected_gradable = expected_data_obj.get("gradable", True)
     actual_gradable = results_obj.get("gradable", True)
-    assert expected_gradable == actual_gradable, (
-        f"Gradable mismatch: expected {expected_gradable}, got {actual_gradable}"
-    )
+    assert expected_gradable == actual_gradable, f"Gradable mismatch: expected {expected_gradable}, got {actual_gradable}"
 
     # If submission is ungradable, check format_errors and message but skip score/test checks
     if not expected_gradable:

tests/test_import_blacklist.py

@@ -1,5 +1,6 @@
 """How import_blacklist entries match the module an import would load."""
 
+import re
 import sys
 import types
 
@@ -39,7 +40,7 @@ def test_block_blacklisted_functions(monkeypatch: pytest.MonkeyPatch) -> None:
 
     _block_blacklisted_functions(["os", "fake_linalg.inv", "fake_linalg.EPSILON", "missing_module.f"])
 
-    with pytest.raises(RuntimeError, match="'fake_linalg.inv' is blacklisted and cannot be used."):
+    with pytest.raises(RuntimeError, match=re.escape("'fake_linalg.inv' is blacklisted and cannot be used.")):
         attributes["inv"]([[1.0]])
     assert attributes["solve"]([[1.0]], [2.0]) == [2.0]
     assert attributes["EPSILON"] == 1e-8

tests/test_serialize.py

@@ -1,12 +1,13 @@
-from pathlib import Path
 import subprocess
 import sys
+from pathlib import Path
 from typing import Any
 from typing import cast
 
 import numpy as np
 import pandas as pd
 import pytest
+import tomllib
 
 from pytest_prairielearn_grader.json_utils import from_json
 from pytest_prairielearn_grader.json_utils import from_server_json
@@ -111,9 +112,6 @@ def test_from_server_json_networkx_graph() -> None:
 
 
 def test_optional_dependencies_are_declared() -> None:
-    import tomllib
-    from pathlib import Path
-
     project = tomllib.loads((Path(__file__).parents[1] / "pyproject.toml").read_text())["project"]
     requirements = project["dependencies"]