Introduce pytest-raises-without-match rule rule to prevent pytest.raises from being called without match argument

.github/workflows/master.yml

@@ -41,7 +41,10 @@ jobs:
       run: |
         source ./dev/install-common-deps.sh
         pip install -r requirements/lint-requirements.txt
-    - name: Run tests
+    - name: Test custom pylint-plugins
+      run : |
+        pytest tests/pylint_plugins
+    - name: Run lint checks
       run: |
         ./dev/lint.sh
   r:

mlflow/entities/run_info.py

@@ -4,19 +4,22 @@
 from mlflow.exceptions import MlflowException
 
 from mlflow.protos.service_pb2 import RunInfo as ProtoRunInfo
+from mlflow.protos.databricks_pb2 import INVALID_PARAMETER_VALUE
 
 
 def check_run_is_active(run_info):
     if run_info.lifecycle_stage != LifecycleStage.ACTIVE:
         raise MlflowException(
-            "The run {} must be in 'active' lifecycle_stage.".format(run_info.run_id)
+            "The run {} must be in 'active' lifecycle_stage.".format(run_info.run_id),
+            error_code=INVALID_PARAMETER_VALUE,
         )
 
 
 def check_run_is_deleted(run_info):
     if run_info.lifecycle_stage != LifecycleStage.DELETED:
         raise MlflowException(
-            "The run {} must be in 'deleted' lifecycle_stage.".format(run_info.run_id)
+            "The run {} must be in 'deleted' lifecycle_stage.".format(run_info.run_id),
+            error_code=INVALID_PARAMETER_VALUE,
         )
 
 

mlflow/utils/uri.py

@@ -258,7 +258,7 @@ def is_databricks_model_registry_artifacts_uri(artifact_uri):
 def construct_run_url(hostname, experiment_id, run_id, workspace_id=None):
     if not hostname or not experiment_id or not run_id:
         raise MlflowException(
-            "Hostname, experiment ID, and run ID are all required to construct" "a run URL"
+            "Hostname, experiment ID, and run ID are all required to construct a run URL"
         )
     prefix = hostname
     if workspace_id and workspace_id != "0":

pylint_plugins/__init__.py

@@ -0,0 +1,5 @@
+from .pytest_raises_without_match import PytestRaisesWithoutMatch
+
+
+def register(linter):
+    linter.register_checker(PytestRaisesWithoutMatch(linter))

pylint_plugins/pytest_raises_without_match/README.md

@@ -0,0 +1,52 @@
+# `pytest-raises-without-match`
+
+This custom pylint rule disallows calling `pytest.raises` without a `match` argument
+to avoid capturing unintended exceptions and eliminate false-positive tests.
+
+## Example
+
+Suppose we want to test this function throws `Exception("bar")` when `condition2` is satisfied.
+
+```python
+def func():
+    if condition1:
+        raise Exception("foo")
+
+    if condition2:
+        raise Exception("bar")
+```
+
+### Bad
+
+```python
+def test_func():
+    with pytest.raises(Exception):
+        func()
+```
+
+- This test passes when `condition1` is unintentionally satisfied.
+- Future code readers will struggle to identify which exception `pytest.raises` should match.
+
+### Good
+
+```python
+def test_func():
+    with pytest.raises(Exception, match="bar"):
+        func()
+```
+
+- This test fails when `condition1` is unintentionally satisfied.
+- Future code readers can quickly identify which exception `pytest.raises` should match by searching `bar`.
+
+## How to disable this rule
+
+```python
+def test_func():
+    with pytest.raises(Exception):  # pylint: disable=pytest-raises-without-match
+        func()
+```
+
+## References
+
+- https://docs.pytest.org/en/latest/how-to/assert.html#assertions-about-expected-exceptions
+- https://docs.pytest.org/en/latest/reference/reference.html#pytest.raises

pylint_plugins/pytest_raises_without_match/__init__.py

@@ -0,0 +1,38 @@
+import astroid
+from pylint.interfaces import IAstroidChecker
+from pylint.checkers import BaseChecker
+
+
+class PytestRaisesWithoutMatch(BaseChecker):
+    __implements__ = IAstroidChecker
+
+    name = "pytest-raises-without-match"
+    msgs = {
+        "W0001": (
+            "`pytest.raises` must be called with `match` argument` ",
+            name,
+            "Use `pytest.raises(<exception>, match=...)`",
+        ),
+    }
+    priority = -1
+
+    @staticmethod
+    def _is_pytest_raises_call(node: astroid.Call):
+        if not isinstance(node.func, astroid.Attribute) or not isinstance(
+            node.func.expr, astroid.Name
+        ):
+            return False
+        return node.func.expr.name == "pytest" and node.func.attrname == "raises"
+
+    @staticmethod
+    def _called_with_match(node: astroid.Call):
+        # Note `match` is a keyword-only argument:
+        # https://docs.pytest.org/en/latest/reference/reference.html#pytest.raises
+        return any(k.arg == "match" for k in node.keywords)
+
+    def visit_call(self, node: astroid.Call):
+        if not PytestRaisesWithoutMatch._is_pytest_raises_call(node):
+            return
+
+        if not PytestRaisesWithoutMatch._called_with_match(node):
+            self.add_message(self.name, node=node)

pylintrc

@@ -33,7 +33,7 @@ jobs=0
 
 # List of plugins (as comma separated values of python modules names) to load,
 # usually to register additional checkers.
-load-plugins=
+load-plugins=pylint_plugins
 
 # Pickle collected data for later comparisons.
 persistent=yes

tests/autologging/test_autologging_behaviors_unit.py

@@ -224,7 +224,7 @@ def parallel_fn():
         time.sleep(np.random.random())
         patch_destination.fn()
 
-    with pytest.raises(Exception):
+    with pytest.raises(Exception, match="enablement error"):
         test_autolog(silent=True)
 
     with pytest.warns(None):

tests/autologging/test_autologging_client.py

@@ -233,7 +233,7 @@ def test_client_correctly_operates_as_context_manager_for_synchronous_flush():
     assert run_tags_1 == tags_to_log
 
     exc_to_raise = Exception("test exception")
-    with pytest.raises(Exception) as raised_exc_info:
+    with pytest.raises(Exception, match=str(exc_to_raise)) as raised_exc_info:
         with mlflow.start_run(), MlflowAutologgingQueueingClient() as client:
             run_id_2 = mlflow.active_run().info.run_id
             client.log_params(run_id_2, params_to_log)
@@ -264,7 +264,7 @@ def test_logging_failures_are_handled_as_expected():
         client.log_metrics(run_id=pending_run_id, metrics={"a": 1})
         client.set_terminated(run_id=pending_run_id, status="KILLED")
 
-        with pytest.raises(MlflowException) as exc:
+        with pytest.raises(MlflowException, match="Batch logging failed!") as exc:
             client.flush()
 
         runs = mlflow.search_runs(experiment_ids=[experiment_id], output_format="list")

tests/autologging/test_autologging_safety_unit.py

@@ -282,7 +282,7 @@ def patch_impl(original, *args, **kwargs):
 
     safe_patch(test_autologging_integration, patch_destination, "fn", patch_impl)
 
-    with pytest.raises(Exception) as exc:
+    with pytest.raises(Exception, match=str(exc_to_throw)) as exc:
         patch_destination.fn()
 
     assert exc.value == exc_to_throw
@@ -319,7 +319,7 @@ def patch_impl(original, *args, **kwargs):
         raise exc_to_throw
 
     safe_patch(test_autologging_integration, patch_destination, "fn", patch_impl)
-    with pytest.raises(Exception) as exc:
+    with pytest.raises(Exception, match=str(exc_to_throw)) as exc:
         patch_destination.fn()
 
     assert exc.value == exc_to_throw
@@ -860,7 +860,7 @@ def non_throwing_function():
     def throwing_function():
         raise exc_to_throw
 
-    with pytest.raises(Exception) as exc:
+    with pytest.raises(Exception, match=str(exc_to_throw)) as exc:
         throwing_function()
 
     assert exc.value == exc_to_throw
@@ -913,7 +913,7 @@ class ThrowingClass(baseclass, metaclass=metaclass):
         def function(self):
             raise exc_to_throw
 
-    with pytest.raises(Exception) as exc:
+    with pytest.raises(Exception, match=str(exc_to_throw)) as exc:
         ThrowingClass().function()
 
     assert exc.value == exc_to_throw
@@ -998,14 +998,14 @@ def patch_function(original, *args, **kwargs):
 
     patch_function = with_managed_run("test_integration", patch_function)
 
-    with pytest.raises(Exception):
+    with pytest.raises(Exception, match="bad implementation"):
         patch_function(lambda: "foo")
 
     assert patch_function_active_run is not None
     status1 = client.get_run(patch_function_active_run.info.run_id).info.status
     assert RunStatus.from_string(status1) == RunStatus.FAILED
 
-    with mlflow.start_run() as active_run, pytest.raises(Exception):
+    with mlflow.start_run() as active_run, pytest.raises(Exception, match="bad implementation"):
         patch_function(lambda: "foo")
         assert patch_function_active_run == active_run
         # `with_managed_run` should not terminate a preexisting MLflow run,
@@ -1053,14 +1053,14 @@ def _on_exception(self, exception):
 
     TestPatch = with_managed_run("test_integration", TestPatch)
 
-    with pytest.raises(Exception):
+    with pytest.raises(Exception, match="bad implementation"):
         TestPatch.call(lambda: "foo")
 
     assert patch_function_active_run is not None
     status1 = client.get_run(patch_function_active_run.info.run_id).info.status
     assert RunStatus.from_string(status1) == RunStatus.FAILED
 
-    with mlflow.start_run() as active_run, pytest.raises(Exception):
+    with mlflow.start_run() as active_run, pytest.raises(Exception, match="bad implementation"):
         TestPatch.call(lambda: "foo")
         assert patch_function_active_run == active_run
         # `with_managed_run` should not terminate a preexisting MLflow run,
@@ -1108,7 +1108,7 @@ def original():
         "test_integration", lambda original, *args, **kwargs: original(*args, **kwargs)
     )
 
-    with pytest.raises(KeyboardInterrupt):
+    with pytest.raises(KeyboardInterrupt, match=""):
         patch_function_1(original)
 
     assert not mlflow.active_run()
@@ -1124,7 +1124,7 @@ def _on_exception(self, exception):
 
     patch_function_2 = with_managed_run("test_integration", PatchFunction2)
 
-    with pytest.raises(KeyboardInterrupt):
+    with pytest.raises(KeyboardInterrupt, match=""):
 
         patch_function_2.call(original)
 
@@ -1418,8 +1418,8 @@ def patch_fn(original):
 
     # If use safe_patch to patch, exception would not come from original fn and so would be logged
     patch_destination.fn = patch_fn
-    with pytest.raises(Exception):
-        patch_destination.fn()
+    with pytest.raises(Exception, match="Exception that should stop autologging session"):
+        patch_destination.fn(lambda: None)
 
     assert _AutologgingSessionManager.active_session() is None
 
@@ -1573,7 +1573,7 @@ def _predict(self, X, a, b):
         @property
         def predict(self):
             if not self._has_predict:
-                raise AttributeError()
+                raise AttributeError("does not have predict")
             return self._predict
 
     class ExtendedEstimator(BaseEstimator):
@@ -1624,7 +1624,7 @@ def autolog(disable=False, exclusive=False, silent=False):  # pylint: disable=un
 
         bad_estimator = EstimatorCls(has_predict=False)
         assert not hasattr(bad_estimator, "predict")
-        with pytest.raises(AttributeError):
+        with pytest.raises(AttributeError, match="does not have predict"):
             bad_estimator.predict(X=1, a=2, b=3)
 
     autolog(disable=True)

tests/autologging/test_autologging_utils.py

@@ -911,9 +911,9 @@ def f4(self, *args, **kwargs):
     assert 3 == get_instance_method_first_arg_value(Test.f1, [3], {"cd2": 4})
     assert 3 == get_instance_method_first_arg_value(Test.f1, [], {"ab1": 3, "cd2": 4})
     assert 3 == get_instance_method_first_arg_value(Test.f2, [3, 4], {})
-    with pytest.raises(AssertionError):
+    with pytest.raises(AssertionError, match=""):
         get_instance_method_first_arg_value(Test.f3, [], {"ab1": 3, "cd2": 4})
-    with pytest.raises(AssertionError):
+    with pytest.raises(AssertionError, match=""):
         get_instance_method_first_arg_value(Test.f4, [], {"ab1": 3, "cd2": 4})
 
 

tests/azureml/test_deploy.py

@@ -287,7 +287,9 @@ def test_deploy_throws_exception_if_model_does_not_contain_pyfunc_flavor(sklearn
     del model_config.flavors[pyfunc.FLAVOR_NAME]
     model_config.save(model_config_path)
 
-    with AzureMLMocks(), pytest.raises(MlflowException) as exc:
+    with AzureMLMocks(), pytest.raises(
+        MlflowException, match="does not contain the `python_function` flavor"
+    ) as exc:
         workspace = get_azure_workspace()
         mlflow.azureml.deploy(model_uri=model_path, workspace=workspace)
         assert exc.error_code == INVALID_PARAMETER_VALUE
@@ -304,7 +306,7 @@ def test_deploy_throws_exception_if_model_python_version_is_less_than_three(
     model_config.flavors[pyfunc.FLAVOR_NAME][pyfunc.PY_VERSION] = "2.7.6"
     model_config.save(model_config_path)
 
-    with AzureMLMocks(), pytest.raises(MlflowException) as exc:
+    with AzureMLMocks(), pytest.raises(MlflowException, match="Python 3 and above") as exc:
         workspace = get_azure_workspace()
         mlflow.azureml.deploy(model_uri=model_path, workspace=workspace)
         assert exc.error_code == INVALID_PARAMETER_VALUE

tests/azureml/test_image_creation.py

@@ -417,7 +417,9 @@ def test_build_image_throws_exception_if_model_does_not_contain_pyfunc_flavor(
     del model_config.flavors[pyfunc.FLAVOR_NAME]
     model_config.save(model_config_path)
 
-    with AzureMLMocks(), pytest.raises(MlflowException) as exc:
+    with AzureMLMocks(), pytest.raises(
+        MlflowException, match="does not contain the `python_function` flavor"
+    ) as exc:
         workspace = get_azure_workspace()
         mlflow.azureml.build_image(model_uri=model_path, workspace=workspace)
         assert exc.error_code == INVALID_PARAMETER_VALUE
@@ -434,7 +436,7 @@ def test_build_image_throws_exception_if_model_python_version_is_less_than_three
     model_config.flavors[pyfunc.FLAVOR_NAME][pyfunc.PY_VERSION] = "2.7.6"
     model_config.save(model_config_path)
 
-    with AzureMLMocks(), pytest.raises(MlflowException) as exc:
+    with AzureMLMocks(), pytest.raises(MlflowException, match="Python 3 and above") as exc:
         workspace = get_azure_workspace()
         mlflow.azureml.build_image(model_uri=model_path, workspace=workspace)
         assert exc.error_code == INVALID_PARAMETER_VALUE

tests/data/test_data.py

@@ -50,7 +50,7 @@ def test_download_uri():
     # Verify exceptions are thrown when downloading from unsupported/invalid URIs
     invalid_prefixes = ["file://", "/tmp"]
     for prefix in invalid_prefixes:
-        with temp_directory() as dst_dir, pytest.raises(DownloadException):
+        with temp_directory() as dst_dir, pytest.raises(DownloadException, match="`uri` must be"):
             download_uri(
                 uri=os.path.join(prefix, "some/path"), output_path=os.path.join(dst_dir, "tmp-file")
             )

tests/deployments/test_deployments.py

@@ -45,7 +45,9 @@ def test_get_success():
 
 
 def test_wrong_target_name():
-    with pytest.raises(MlflowException):
+    with pytest.raises(
+        MlflowException, match='No plugin found for managing model deployments to "wrong_target"'
+    ):
         deployments.get_deploy_client("wrong_target")
 
 
@@ -56,15 +58,15 @@ class DummyPlugin:
     dummy_plugin = DummyPlugin()
     plugin_manager = DeploymentPlugins()
     plugin_manager.registry["dummy"] = dummy_plugin
-    with pytest.raises(MlflowException):
+    with pytest.raises(MlflowException, match="Plugin registered for the target dummy"):
         plugin_manager["dummy"]  # pylint: disable=pointless-statement
 
 
 def test_plugin_raising_error():
     client = deployments.get_deploy_client(f_target)
     # special case to raise error
     os.environ["raiseError"] = "True"
-    with pytest.raises(RuntimeError):
+    with pytest.raises(RuntimeError, match="Error requested"):
         client.list_deployments()
     os.environ["raiseError"] = "False"
 

tests/entities/test_run.py

@@ -95,6 +95,6 @@ def test_string_repr(self):
 
     def test_creating_run_with_absent_info_throws_exception(self):
         run_data = TestRunData._create()[0]
-        with pytest.raises(MlflowException) as no_info_exc:
+        with pytest.raises(MlflowException, match="run_info cannot be None") as no_info_exc:
             Run(None, run_data)
         assert "run_info cannot be None" in str(no_info_exc)