[pre-commit.ci] pre-commit autoupdate

.pre-commit-config.yaml

@@ -33,7 +33,7 @@ repos:
 
   # Ruff, the Python auto-correcting linter/formatter written in Rust
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.6.9
+    rev: v0.9.1
     hooks:
       - id: ruff
         args: ["--fix", "--show-fixes"]

examples/user_stories/generate_disclosure_risk_report.py

@@ -43,8 +43,7 @@
             config = yaml.load(handle, Loader=yaml.loader.SafeLoader)
     except AttributeError as error:  # pragma:no cover
         print(
-            "Invalid command. Try --help to get more details"
-            f"error message is {error}"
+            f"Invalid command. Try --help to get more detailserror message is {error}"
         )
 
     user_story = config["user_story"]

examples/user_stories/user_story_1/user_story_1_tre.py

@@ -69,8 +69,7 @@ def run_user_story(release_config: dict):
             config = yaml.load(handle, Loader=yaml.loader.SafeLoader)
     except AttributeError as error:
         print(
-            "Invalid command. Try --help to get more details"
-            f"error message is {error}"
+            f"Invalid command. Try --help to get more detailserror message is {error}"
         )
 
     run_user_story(config)

examples/user_stories/user_story_2/user_story_2_tre.py

@@ -151,8 +151,7 @@ def run_user_story(release_config: dict):
             config = yaml.load(handle, Loader=yaml.loader.SafeLoader)
     except AttributeError as error:
         print(
-            "Invalid command. Try --help to get more details"
-            f"error message is {error}"
+            f"Invalid command. Try --help to get more detailserror message is {error}"
         )
 
     run_user_story(config)

examples/user_stories/user_story_3/user_story_3_tre.py

@@ -156,8 +156,7 @@ def run_user_story(release_config: dict):
             config = yaml.load(handle, Loader=yaml.loader.SafeLoader)
     except AttributeError as error:
         print(
-            "Invalid command. Try --help to get more details"
-            f"error message is {error}"
+            f"Invalid command. Try --help to get more detailserror message is {error}"
         )
 
     run_user_story(config)

examples/user_stories/user_story_4/user_story_4_tre.py

@@ -120,8 +120,7 @@ def run_user_story(release_config: dict):
             config = yaml.load(handle, Loader=yaml.loader.SafeLoader)
     except AttributeError as error:
         print(
-            "Invalid command. Try --help to get more details"
-            f"error message is {error}"
+            f"Invalid command. Try --help to get more detailserror message is {error}"
         )
 
     run_user_story(config)

examples/user_stories/user_story_7/user_story_7_tre.py

@@ -72,8 +72,7 @@ def run_user_story(release_config: dict):
             config = yaml.load(handle, Loader=yaml.loader.SafeLoader)
     except AttributeError as error:
         print(
-            "Invalid command. Try --help to get more details"
-            f"error message is {error}"
+            f"Invalid command. Try --help to get more detailserror message is {error}"
         )
 
     run_user_story(config)

examples/user_stories/user_story_8/user_story_8_tre.py

@@ -72,8 +72,7 @@ def run_user_story(release_config: dict):
             config = yaml.load(handle, Loader=yaml.loader.SafeLoader)
     except AttributeError as error:
         print(
-            "Invalid command. Try --help to get more details"
-            f"error message is {error}"
+            f"Invalid command. Try --help to get more detailserror message is {error}"
         )
 
     run_user_story(config)

sacroml/attacks/worst_case_attack.py

@@ -360,7 +360,7 @@ def _get_global_metrics(self, attack_metrics: list) -> dict:
             )
 
             global_metrics["null_auc_3sd_range"] = (
-                f"{0.5 - 3*auc_std:.4f} -> {0.5 + 3*auc_std:.4f}"
+                f"{0.5 - 3 * auc_std:.4f} -> {0.5 + 3 * auc_std:.4f}"
             )
             global_metrics["n_sig_auc_p_vals"] = self._get_n_significant(
                 auc_p_vals, self.p_thresh

sacroml/safemodel/reporting.py

@@ -146,7 +146,7 @@ def get_reporting_string(**kwargs: dict) -> str:
             f"epochs = {inter_params['epochs']}.\n"
         ),
         "basic_params_differ": (
-            "Warning: basic parameters differ in " f"{inter_params['length']} places:\n"
+            f"Warning: basic parameters differ in {inter_params['length']} places:\n"
         ),
         "param_changed_from_to": (
             f"parameter {inter_params['key']} changed from {inter_params['val']} "
@@ -200,10 +200,10 @@ def get_reporting_string(**kwargs: dict) -> str:
             f"{inter_params['diffs_list']}"
         ),
         "no_dp_gradients_key": (
-            "optimizer does not contain key _was_dp_gradients_called" " so is not DP."
+            "optimizer does not contain key _was_dp_gradients_called so is not DP."
         ),
         "found_dp_gradients_key": (
-            "optimizer does  contain key _was_dp_gradients_called" " so should be DP."
+            "optimizer does  contain key _was_dp_gradients_called so should be DP."
         ),
         "changed_opt_no_fit": (
             "optimizer has been changed but fit() has not been rerun."
@@ -229,8 +229,7 @@ def get_reporting_string(**kwargs: dict) -> str:
             f"file gave this error message:  {inter_params['er']}"
         ),
         "loading_from_unsupported": (
-            f"loading from a {inter_params['suffix']} "
-            "file is currently not supported"
+            f"loading from a {inter_params['suffix']} file is currently not supported"
         ),
         "opt_config_changed": ("Optimizer config has been changed since training."),
         "epsilon_above_normal": (
@@ -239,7 +238,7 @@ def get_reporting_string(**kwargs: dict) -> str:
             "Discussion with researcher needed.\n"
         ),
         "recommend_further_discussion": (
-            f"Recommendation is further discussion needed " f"{inter_params['msg']}.\n"
+            f"Recommendation is further discussion needed {inter_params['msg']}.\n"
         ),
         "recommend_allow_release": ("Recommendation is to allow release.\n"),
         "allow_release_eps_below_max": (

tests/attacks/test_structural_attack.py

@@ -88,12 +88,12 @@ def test_unnecessary_risk():
     ]
     for idx, paramdict in enumerate(risky_param_dicts):
         model = DecisionTreeClassifier(**paramdict)
-        errstr = f" unnecessary risk with rule {idx}" f"params are {model.get_params()}"
+        errstr = f" unnecessary risk with rule {idx}params are {model.get_params()}"
         assert sa.get_unnecessary_risk(model), errstr
     model = DecisionTreeClassifier(max_depth=1, min_samples_leaf=150)
-    assert not sa.get_unnecessary_risk(
-        model
-    ), f"should be non-disclosive with {model.get_params}"
+    assert not sa.get_unnecessary_risk(model), (
+        f"should be non-disclosive with {model.get_params}"
+    )
 
     # now random forest
     risky_param_dicts = [
@@ -114,12 +114,12 @@ def test_unnecessary_risk():
     ]
     for idx, paramdict in enumerate(risky_param_dicts):
         model = RandomForestClassifier(**paramdict)
-        errstr = f" unnecessary risk with rule {idx}" f"params are {model.get_params()}"
+        errstr = f" unnecessary risk with rule {idx}params are {model.get_params()}"
         assert sa.get_unnecessary_risk(model), errstr
     model = RandomForestClassifier(max_depth=1, n_estimators=25, min_samples_leaf=150)
-    assert not sa.get_unnecessary_risk(
-        model
-    ), f"should be non-disclosive with {model.get_params}"
+    assert not sa.get_unnecessary_risk(model), (
+        f"should be non-disclosive with {model.get_params}"
+    )
 
     # finally xgboost
     risky_param_dicts = [
@@ -141,12 +141,12 @@ def test_unnecessary_risk():
     ]
     for idx, paramdict in enumerate(risky_param_dicts):
         model = XGBClassifier(**paramdict)
-        errstr = f" unnecessary risk with rule {idx}" f"params are {model.get_params()}"
+        errstr = f" unnecessary risk with rule {idx}params are {model.get_params()}"
         assert sa.get_unnecessary_risk(model), errstr
     model = XGBClassifier(min_child_weight=10)
-    assert not sa.get_unnecessary_risk(
-        model
-    ), f"should be non-disclosive with {model.get_params}"
+    assert not sa.get_unnecessary_risk(model), (
+        f"should be non-disclosive with {model.get_params}"
+    )
 
 
 def test_non_trees():
@@ -170,12 +170,12 @@ def test_dt():
     myattack = sa.StructuralAttack()
     myattack.attack(target)
     assert not myattack.dof_risk, "should be no DoF risk with decision stump"
-    assert (
-        not myattack.k_anonymity_risk
-    ), "should be no k-anonymity risk with min_samples_leaf 150"
-    assert (
-        not myattack.class_disclosure_risk
-    ), "no class disclosure risk for stump with min samples leaf 150"
+    assert not myattack.k_anonymity_risk, (
+        "should be no k-anonymity risk with min_samples_leaf 150"
+    )
+    assert not myattack.class_disclosure_risk, (
+        "no class disclosure risk for stump with min samples leaf 150"
+    )
     assert not myattack.unnecessary_risk, "not unnecessary risk if max_depth < 3.5"
 
     # highly disclosive
@@ -184,15 +184,15 @@ def test_dt():
     myattack = sa.StructuralAttack()
     myattack.attack(target)
     assert not myattack.dof_risk, "should be no DoF risk with decision stump"
-    assert (
-        myattack.k_anonymity_risk
-    ), "should be  k-anonymity risk with unlimited depth and min_samples_leaf 5"
-    assert (
-        myattack.class_disclosure_risk
-    ), "should be class disclosure risk with unlimited depth and min_samples_leaf 5"
-    assert (
-        myattack.unnecessary_risk
-    ), " unnecessary risk with unlimited depth and min_samples_leaf 5"
+    assert myattack.k_anonymity_risk, (
+        "should be  k-anonymity risk with unlimited depth and min_samples_leaf 5"
+    )
+    assert myattack.class_disclosure_risk, (
+        "should be class disclosure risk with unlimited depth and min_samples_leaf 5"
+    )
+    assert myattack.unnecessary_risk, (
+        " unnecessary risk with unlimited depth and min_samples_leaf 5"
+    )
 
 
 def test_adaboost():
@@ -206,9 +206,9 @@ def test_adaboost():
     myattack.THRESHOLD = 2
     myattack.attack(target)
     assert not myattack.dof_risk, "should be no DoF risk with just 2 decision stumps"
-    assert (
-        not myattack.k_anonymity_risk
-    ), "should be no k-anonymity risk with only 2 stumps"
+    assert not myattack.k_anonymity_risk, (
+        "should be no k-anonymity risk with only 2 stumps"
+    )
     assert not myattack.class_disclosure_risk, "no class disclosure risk for 2 stumps"
     assert not myattack.unnecessary_risk, " unnecessary risk not defined for adaboost"
 
@@ -222,12 +222,12 @@ def test_adaboost():
     myattack2 = sa.StructuralAttack()
     myattack2.attack(target)
     assert myattack2.dof_risk, "should be  DoF risk with adaboost of deep trees"
-    assert (
-        myattack2.k_anonymity_risk
-    ), "should be k-anonymity risk with adaboost unlimited depth and min_samples_leaf 2"
-    assert (
-        myattack2.class_disclosure_risk
-    ), "should be class risk with adaboost unlimited depth and min_samples_leaf 2"
+    assert myattack2.k_anonymity_risk, (
+        "should be k-anonymity risk with adaboost unlimited depth and min_samples_leaf 2"
+    )
+    assert myattack2.class_disclosure_risk, (
+        "should be class risk with adaboost unlimited depth and min_samples_leaf 2"
+    )
     assert not myattack2.unnecessary_risk, " unnecessary risk not define for adaboost"
 
 
@@ -238,15 +238,15 @@ def test_rf():
     target = get_target("rf", **param_dict)
     myattack = sa.StructuralAttack()
     myattack.attack(target)
-    assert (
-        not myattack.dof_risk
-    ), "should be no DoF risk with small forest of decision stumps"
-    assert (
-        not myattack.k_anonymity_risk
-    ), "should be no k-anonymity risk with min_samples_leaf 150"
-    assert (
-        not myattack.class_disclosure_risk
-    ), "no class disclosure risk for stumps with min samples leaf 150"
+    assert not myattack.dof_risk, (
+        "should be no DoF risk with small forest of decision stumps"
+    )
+    assert not myattack.k_anonymity_risk, (
+        "should be no k-anonymity risk with min_samples_leaf 150"
+    )
+    assert not myattack.class_disclosure_risk, (
+        "no class disclosure risk for stumps with min samples leaf 150"
+    )
     assert not myattack.unnecessary_risk, "not unnecessary risk if max_depth < 3.5"
 
     # highly disclosive
@@ -260,15 +260,15 @@ def test_rf():
     myattack = sa.StructuralAttack()
     myattack.attack(target)
     assert myattack.dof_risk, "should be  DoF risk with forest of deep trees"
-    assert (
-        myattack.k_anonymity_risk
-    ), "should be  k-anonymity risk with unlimited depth and min_samples_leaf 5"
-    assert (
-        myattack.class_disclosure_risk
-    ), "should be class disclsoure risk with unlimited depth and min_samples_leaf 5"
-    assert (
-        myattack.unnecessary_risk
-    ), " unnecessary risk with unlimited depth and min_samples_leaf 5"
+    assert myattack.k_anonymity_risk, (
+        "should be  k-anonymity risk with unlimited depth and min_samples_leaf 5"
+    )
+    assert myattack.class_disclosure_risk, (
+        "should be class disclsoure risk with unlimited depth and min_samples_leaf 5"
+    )
+    assert myattack.unnecessary_risk, (
+        " unnecessary risk with unlimited depth and min_samples_leaf 5"
+    )
 
 
 def test_xgb():
@@ -278,15 +278,15 @@ def test_xgb():
     target = get_target("xgb", **param_dict)
     myattack = sa.StructuralAttack()
     myattack.attack(target)
-    assert (
-        not myattack.dof_risk
-    ), "should be no DoF risk with small xgb of decision stumps"
-    assert (
-        not myattack.k_anonymity_risk
-    ), "should be no k-anonymity risk with min_samples_leaf 150"
-    assert (
-        not myattack.class_disclosure_risk
-    ), "no class disclosure risk for stumps with min child weight 50"
+    assert not myattack.dof_risk, (
+        "should be no DoF risk with small xgb of decision stumps"
+    )
+    assert not myattack.k_anonymity_risk, (
+        "should be no k-anonymity risk with min_samples_leaf 150"
+    )
+    assert not myattack.class_disclosure_risk, (
+        "no class disclosure risk for stumps with min child weight 50"
+    )
     assert myattack.unnecessary_risk == 0, "not unnecessary risk if max_depth < 3.5"
 
     # highly disclosive
@@ -295,12 +295,12 @@ def test_xgb():
     myattack2 = sa.StructuralAttack()
     myattack2.attack(target2)
     assert myattack2.dof_risk, "should be  DoF risk with xgb of deep trees"
-    assert (
-        myattack2.k_anonymity_risk
-    ), "should be  k-anonymity risk with depth 50 and min_child_weight 1"
-    assert (
-        myattack2.class_disclosure_risk
-    ), "should be class disclosure risk with xgb lots of deep trees"
+    assert myattack2.k_anonymity_risk, (
+        "should be  k-anonymity risk with depth 50 and min_child_weight 1"
+    )
+    assert myattack2.class_disclosure_risk, (
+        "should be class disclosure risk with xgb lots of deep trees"
+    )
     assert myattack2.unnecessary_risk, " unnecessary risk with these xgb params"
 
 
@@ -319,15 +319,15 @@ def test_sklearnmlp():
     paramstr = ""
     for key, val in safeparams.items():
         paramstr += f"{key}:{val}\n"
-    assert (
-        not myattack.dof_risk
-    ), f"should be no DoF risk with small mlp with params {paramstr}"
-    assert (
-        not myattack.k_anonymity_risk
-    ), f"should be no k-anonymity risk with params {paramstr}"
-    assert (
-        myattack.class_disclosure_risk
-    ), f"should be  class disclosure risk with params {paramstr}"
+    assert not myattack.dof_risk, (
+        f"should be no DoF risk with small mlp with params {paramstr}"
+    )
+    assert not myattack.k_anonymity_risk, (
+        f"should be no k-anonymity risk with params {paramstr}"
+    )
+    assert myattack.class_disclosure_risk, (
+        f"should be  class disclosure risk with params {paramstr}"
+    )
     assert not myattack.unnecessary_risk, "not unnecessary risk for mlps at present"
 
     # highly disclosive
@@ -344,12 +344,12 @@ def test_sklearnmlp():
     myattack2 = sa.StructuralAttack()
     myattack2.attack(target2)
     assert myattack2.dof_risk, f"should be DoF risk with this MLP:\n{uparamstr}"
-    assert (
-        myattack2.k_anonymity_risk
-    ), "559/560 records should be k-anonymity risk with this MLP:\n{uparamstr}"
-    assert (
-        myattack2.class_disclosure_risk
-    ), "should be class disclosure risk with this MLP:\n{uparamstr}"
+    assert myattack2.k_anonymity_risk, (
+        "559/560 records should be k-anonymity risk with this MLP:\n{uparamstr}"
+    )
+    assert myattack2.class_disclosure_risk, (
+        "should be class disclosure risk with this MLP:\n{uparamstr}"
+    )
     assert not myattack2.unnecessary_risk, "no unnecessary risk yet for MLPClassifiers"