chore: remove pylint

CONTRIBUTING.md

@@ -64,6 +64,11 @@ Then to run on all files in the repository:
 $ pre-commit run -a
 ```
 
+Pre-commit can be configured to automatically run on every `git commit` with:
+```
+$ pre-commit install
+```
+
 ## Documentation
 
 Documentation is hosted here: https://ai-sdc.github.io/SACRO-ML/

examples/user_stories/generate_disclosure_risk_report.py

@@ -49,7 +49,9 @@
     user_story = config["user_story"]
     if user_story == "UNDEFINED":
         print(
-            "User story not selected, please select a user story by referring to user_stories_flow_chart.png and adding the relevant number to the the first line of 'default_config.yaml'"
+            "User story not selected, please select a user story by "
+            "referring to user_stories_flow_chart.png and adding the "
+            "relevant number to the the first line of 'default_config.yaml'"
         )
     elif user_story == 1:
         user_story_1_tre.run_user_story(config)

examples/user_stories/user_story_1/user_story_1_tre.py

@@ -49,7 +49,8 @@ def run_user_story(release_config: dict):
 if __name__ == "__main__":
     parser = argparse.ArgumentParser(
         description=(
-            "Generate a risk report after request_release() has been called by researcher"
+            "Generate a risk report after request_release() "
+            "has been called by researcher"
         )
     )
 

examples/user_stories/user_story_2/user_story_2_tre.py

@@ -49,7 +49,8 @@ def generate_report(
     print()
     print("Acting as TRE...")
     print(
-        "(when instructions on how to recreate the dataset have been provided by the researcher)"
+        "(when instructions on how to recreate the dataset have "
+        "been provided by the researcher)"
     )
     print(directory)
     print()

examples/user_stories/user_story_3/user_story_3_researcher_template.py

@@ -58,7 +58,7 @@ def run_user_story():
 
     X_train, X_test, y_train, y_test = train_test_split(
         feature_dataframe.values,
-        target_dataframe.values.flatten(),
+        target_dataframe.to_numpy().flatten(),
         test_size=0.7,
         random_state=42,
     )

examples/user_stories/user_story_8/data_processing_researcher.py

@@ -1,8 +1,10 @@
 """SUPPORTING FILE FOR USER STORY 2.
 
-This file is an example of a function created by a researcher that will pre-process a dataset
+This file is an example of a function created by a researcher that will
+pre-process a dataset
 
-To use: write a function that will process your input data, and output the processed version
+To use: write a function that will process your input data, and output the
+processed version
 
 NOTE: in order to work, this function needs to:
 
@@ -33,17 +35,17 @@ def process_dataset(data):
 
     row_indices = np.arange(np.shape(X_transformed)[0])
 
-    # This step is not necessary, however it's the simplest way of getting training indices from
-    # the data
-    # Any method of generating indices of samples to be used for training will work here
+    # This step is not necessary, however it's the simplest way of getting
+    # training indices from the data. Any method of generating indices of
+    # samples to be used for training will work here.
     (
         X_train,
         X_test,
         y_train,
         y_test,
         train_indices,
         test_indices,
-    ) = train_test_split(  # pylint: disable=unused-variable
+    ) = train_test_split(
         X_transformed,
         y_transformed,
         row_indices,

pyproject.toml

@@ -84,40 +84,6 @@ packages = {find = {exclude = ["docs*", "examples*", "tests*", "user_stories*"]}
 [tool.setuptools.package-data]
 "sacroml.safemodel" = ["rules.json"]
 
-[tool.pylint]
-master.py-version = "3.9"
-reports.output-format = "colorized"
-disable = [
-    "invalid-name",
-    "no-member",
-    "duplicate-code",
-    "consider-using-min-builtin",
-]
-
-[tool.pylint.messages_control]
-enable = [
-    "useless-suppression",
-]
-
-disable = [
-    "R0917",  # too-many-positional-arguments
-]
-
-[tool.pylint.design]
-max-args = 5  # Maximum number of arguments for function / method (see R0913).
-max-attributes = 7  # Maximum number of attributes for a class (see R0902).
-max-branches = 12  # Maximum number of branch for function / method body (see R0912).
-max-locals = 15  # Maximum number of parents for a class (see R0914).
-max-parents = 7  # Maximum number of parents for a class (see R0901).
-max-public-methods = 20  # Maximum number of public methods for a class (see R0904).
-max-returns = 6  # Max number of return / yield for function / method body (see R0911).
-max-statements = 50  # Max number of statements in function / method body (see R0915).
-min-public-methods = 2  # Minimum number of public methods for a class (see R0903).
-
-[tool.pylint.format]
-max-line-length = 100  # Maximum number of characters on a single line.
-max-module-lines = 1000  # Maximum number of lines in a module.
-
 [tool.ruff]
 indent-width = 4
 line-length = 88
@@ -129,17 +95,17 @@ lint.select = [
   "ARG", # flake8-unused-arguments
   "B",   # flake8-bugbear
   "C4",  # flake8-comprehensions
-#  "C90", # mccabe
+  "C90", # mccabe
   "D",   # pydocstyle
 #  "DTZ", # flake8-datetimez
-#  "E",   # pycodestyle
+  "E",   # pycodestyle
   "EM",  # flake8-errmsg
   "ERA", # eradicate
   "F",   # Pyflakes
   "I",   # isort
   "ICN", # flake8-import-conventions
   "N",   # pep8-naming
-#  "PD",  # pandas-vet
+  "PD",  # pandas-vet
   "PGH", # pygrep-hooks
   "PIE", # flake8-pie
 #  "PL",  # Pylint
@@ -151,7 +117,7 @@ lint.select = [
   "Q",   # flake8-quotes
   "RET", # flake8-return
   "RUF100",  # Ruff-specific
-#  "S",   # flake8-bandit
+  "S",   # flake8-bandit
   "SIM", # flake8-simplify
 #  "T20", # flake8-print
   "TID", # flake8-tidy-imports
@@ -173,8 +139,10 @@ exclude = [
 lint.ignore = [
     "EM101",   # raw-string-in-exception
     "EM102",   # f-string-in-exception
-    "N818",    # error-suffix-on-exception-name
     "PLC0206", # dict-index-missing-items
+    "S101",    # allow asserts
+    "S301",    # allow pickle
+    "PD901",   # allow df for dataframe names
 ]
 
 [tool.ruff.lint.pep8-naming]
@@ -196,6 +164,7 @@ docstring-code-line-length = 80
 "user_stories/**/*" = ["ANN"]
 "tests/**/*" = ["S101", "PLR2004", "ANN"]
 "sacroml/attacks/structural_attack.py" = ["PLR2004"]
+"sacroml/safemodel/classifiers/new_model_template.py" = ["C901"]
 
 [tool.codespell]
 ignore-words-list = [

sacroml/attacks/attack.py

@@ -74,7 +74,7 @@ def _make_report(self, target: Target) -> dict:
         logger.info("Generating report")
         self._construct_metadata()
 
-        if not target.model is None:
+        if target.model is not None:
             self.metadata["target_model"] = target.model.model_name
             self.metadata["target_model_params"] = target.model.model_params
             self.metadata["target_train_params"] = target.model.train_params

sacroml/attacks/attack_report_formatter.py

@@ -110,7 +110,7 @@ def __str__(self) -> str:
         raise NotImplementedError()
 
 
-class FinalRecommendationModule(AnalysisModule):  # pylint: disable=too-many-instance-attributes
+class FinalRecommendationModule(AnalysisModule):
     """Generate the first layer of a recommendation report."""
 
     def __init__(self, report: dict) -> None:
@@ -565,7 +565,7 @@ def process_attack_target_json(
 
         self.text_out.append(bucket_text)
 
-    def export_to_file(  # pylint: disable=too-many-arguments
+    def export_to_file(
         self,
         output_filename: str = "summary.txt",
         move_files: bool = False,

sacroml/attacks/attribute_attack.py

@@ -177,7 +177,7 @@ def _get_unique_features(
     return np.unique(combined_feature)
 
 
-def _get_inference_data(  # pylint: disable=too-many-locals
+def _get_inference_data(
     target: Target, feature_id: int, memberset: bool
 ) -> tuple[np.ndarray, np.ndarray, float]:
     """Return a dataset of each sample with the attributes to test."""
@@ -212,7 +212,7 @@ def _get_inference_data(  # pylint: disable=too-many-locals
     return x_values, y_values, baseline
 
 
-def _infer(  # pylint: disable=too-many-locals
+def _infer(
     target: Target,
     feature_id: int,
     threshold: float,
@@ -329,9 +329,7 @@ def plot_quantitative_risk(res: dict, path: str = "") -> None:
     logger.debug("Saved quantitative risk plot: %s", filename)
 
 
-def plot_categorical_risk(  # pylint: disable=too-many-locals
-    res: dict, path: str = ""
-) -> None:
+def plot_categorical_risk(res: dict, path: str = "") -> None:
     """Generate a bar chart showing categorical risk scores.
 
     Parameters
@@ -376,9 +374,7 @@ def plot_categorical_risk(  # pylint: disable=too-many-locals
     logger.debug("Saved categorical risk plot: %s", filename)
 
 
-def plot_categorical_fraction(  # pylint: disable=too-many-locals
-    res: dict, path: str = ""
-) -> None:
+def plot_categorical_fraction(res: dict, path: str = "") -> None:
     """Generate a bar chart showing fraction of dataset inferred.
 
     Parameters
@@ -472,7 +468,7 @@ def _attack_brute_force(
     ]
 
 
-def _get_bounds_risk_for_sample(  # pylint: disable=too-many-locals,too-many-arguments
+def _get_bounds_risk_for_sample(
     target_model: Model,
     feat_id: int,
     feat_min: float,

sacroml/attacks/data.py

@@ -9,7 +9,7 @@
 from torch.utils.data import DataLoader, Dataset
 
 
-class BaseDataHandler(ABC):  # pylint: disable=too-few-public-methods
+class BaseDataHandler(ABC):
     """Base data handling interface."""
 
     @abstractmethod

sacroml/attacks/likelihood_attack.py

@@ -269,9 +269,9 @@ def _compute_scores(  # pylint: disable=too-many-locals
         global_out_std: float = self._get_global_std(out_conf)
 
         # score each record in the member and non-member sets
-        for i, l in enumerate(combined_y_train):
+        for i, y in enumerate(combined_y_train):
             # get the target model behaviour on the record (handle one-hot or label)
-            label = np.argmax(l) if combined_y_train.ndim > 1 else l
+            label = np.argmax(y) if combined_y_train.ndim > 1 else y
             target_logit: float = _logit(combined_target_preds[i, label])
             # get behaviour observed with the record as a non-member
             out_mean, out_std = self._describe_conf(out_conf[i], global_out_std)

sacroml/attacks/model.py

@@ -11,10 +11,10 @@
 import numpy as np
 
 
-class Model(ABC):  # pylint: disable=too-many-instance-attributes
+class Model(ABC):
     """Interface to a target model."""
 
-    def __init__(  # pylint: disable=too-many-arguments
+    def __init__(
         self,
         model: Any,
         model_path: str = "",
@@ -203,7 +203,7 @@ def save(self, path: str) -> None:
 
     @classmethod
     @abstractmethod
-    def load(  # pylint: disable=too-many-arguments
+    def load(
         cls,
         model_path: str,
         model_module_path: str,

sacroml/attacks/model_pytorch.py

@@ -22,7 +22,7 @@
 class PytorchModel(Model):
     """Interface to Pytorch models."""
 
-    def __init__(  # pylint: disable=too-many-arguments
+    def __init__(
         self,
         model: torch.nn.Module,
         model_path: str = "",
@@ -281,7 +281,7 @@ def save(self, path: str) -> None:
         torch.save(self.model.state_dict(), path)
 
     @classmethod
-    def load(  # pylint: disable=too-many-arguments
+    def load(
         cls,
         model_path: str,
         model_module_path: str,

sacroml/attacks/model_sklearn.py

@@ -18,7 +18,7 @@
 class SklearnModel(Model):
     """Interface to sklearn.base.BaseEstimator models."""
 
-    def __init__(  # pylint: disable=too-many-arguments
+    def __init__(
         self,
         model: sklearn.base.BaseEstimator,
         model_path: str = "",
@@ -217,7 +217,7 @@ def save(self, path: str) -> None:
             raise ValueError(f"Unsupported file format for saving a model: {ext}")
 
     @classmethod
-    def load(  # pylint: disable=too-many-arguments
+    def load(
         cls,
         model_path: str,
         model_module_path: str,

sacroml/attacks/report.py

@@ -172,7 +172,7 @@ def title(
     pdf.ln(h=5)
 
 
-def subtitle(  # pylint: disable = too-many-arguments
+def subtitle(
     pdf: FPDF,
     text: str,
     indent: int = 10,
@@ -186,7 +186,7 @@ def subtitle(  # pylint: disable = too-many-arguments
     pdf.cell(75, 10, text, border, 1)
 
 
-def line(  # pylint: disable = too-many-arguments
+def line(
     pdf: FPDF,
     text: str,
     indent: int = 0,

sacroml/attacks/structural_attack.py

@@ -27,8 +27,6 @@
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger(__name__)
 
-# pylint: disable=chained-comparison
-
 
 def get_unnecessary_risk(model: BaseEstimator) -> bool:
     """Check whether model hyperparameters are in the top 20% most risky.
@@ -226,8 +224,6 @@ def _get_model_param_count_mlp(model: MLPClassifier) -> int:
 class StructuralAttack(Attack):
     """Structural attacks based on the static structure of a model."""
 
-    # pylint: disable=too-many-instance-attributes
-
     def __init__(
         self,
         output_dir: str = "outputs",