MAINT Enable ruff B (flake8-bugbear) rules and fix all violations

doc/code/converters/0_converters.ipynb

@@ -615,7 +615,7 @@
     "\n",
     "# Display all rows\n",
     "pd.set_option(\"display.max_rows\", None)\n",
-    "df"
+    "print(df)"
    ]
   },
   {

doc/code/converters/0_converters.py

@@ -50,7 +50,7 @@
 
 # Display all rows
 pd.set_option("display.max_rows", None)
-df
+print(df)
 
 # %% [markdown]
 # ## Converter Categories

doc/code/converters/4_video_converters.ipynb

@@ -69,7 +69,7 @@
     "\n",
     "video = AddImageVideoConverter(video_path=input_video)\n",
     "converted_vid = await video.convert_async(prompt=input_image, input_type=\"image_path\")  # type: ignore\n",
-    "converted_vid"
+    "print(converted_vid)"
    ]
   }
  ],

doc/code/converters/4_video_converters.py

@@ -41,4 +41,4 @@
 
 video = AddImageVideoConverter(video_path=input_video)
 converted_vid = await video.convert_async(prompt=input_image, input_type="image_path")  # type: ignore
-converted_vid
+print(converted_vid)

doc/code/memory/embeddings.ipynb

@@ -109,7 +109,7 @@
     "from pyrit.common.path import DB_DATA_PATH\n",
     "\n",
     "saved_embedding_path = embedding_response.save_to_file(directory_path=DB_DATA_PATH)\n",
-    "saved_embedding_path"
+    "print(saved_embedding_path)"
    ]
   },
   {
@@ -143,7 +143,7 @@
     "from pyrit.common.path import DB_DATA_PATH\n",
     "\n",
     "saved_embedding_path = embedding_response.save_to_file(directory_path=DB_DATA_PATH)\n",
-    "saved_embedding_path"
+    "print(saved_embedding_path)"
    ]
   }
  ],

doc/code/memory/embeddings.py

@@ -49,7 +49,7 @@
 from pyrit.common.path import DB_DATA_PATH
 
 saved_embedding_path = embedding_response.save_to_file(directory_path=DB_DATA_PATH)
-saved_embedding_path
+print(saved_embedding_path)
 
 # %% [markdown]
 # To load an embedding from disk
@@ -59,4 +59,4 @@
 from pyrit.common.path import DB_DATA_PATH
 
 saved_embedding_path = embedding_response.save_to_file(directory_path=DB_DATA_PATH)
-saved_embedding_path
+print(saved_embedding_path)

doc/code/scoring/8_scorer_metrics.ipynb

@@ -388,7 +388,7 @@
     "\n",
     "print(\"Top 5 configurations by F1 Score:\")\n",
     "print(\"-\" * 80)\n",
-    "for i, entry in enumerate(sorted_by_f1[:5], 1):\n",
+    "for _i, entry in enumerate(sorted_by_f1[:5], 1):\n",
     "    printer = ConsoleScorerPrinter()\n",
     "    printer.print_objective_scorer(scorer_identifier=entry.scorer_identifier)\n",
     "\n",
@@ -468,7 +468,7 @@
     "\n",
     "print(\"Top configurations by Mean Absolute Error:\")\n",
     "print(\"-\" * 80)\n",
-    "for i, e in enumerate(sorted_by_mae[:5], 1):\n",
+    "for _i, e in enumerate(sorted_by_mae[:5], 1):\n",
     "    printer = ConsoleScorerPrinter()\n",
     "    printer.print_harm_scorer(scorer_identifier=e.scorer_identifier, harm_category=\"violence\")"
    ]

doc/code/scoring/8_scorer_metrics.py

@@ -195,7 +195,7 @@
 
 print("Top 5 configurations by F1 Score:")
 print("-" * 80)
-for i, entry in enumerate(sorted_by_f1[:5], 1):
+for _i, entry in enumerate(sorted_by_f1[:5], 1):
     printer = ConsoleScorerPrinter()
     printer.print_objective_scorer(scorer_identifier=entry.scorer_identifier)
 
@@ -227,7 +227,7 @@
 
 print("Top configurations by Mean Absolute Error:")
 print("-" * 80)
-for i, e in enumerate(sorted_by_mae[:5], 1):
+for _i, e in enumerate(sorted_by_mae[:5], 1):
     printer = ConsoleScorerPrinter()
     printer.print_harm_scorer(scorer_identifier=e.scorer_identifier, harm_category="violence")
 

doc/cookbooks/4_testing_bias.ipynb

@@ -323,7 +323,7 @@
     "for job, gender in itertools.product(jobs[:3], genders):  # for demonstration purposes we just use first 3 jobs\n",
     "    question = f\"Is the {job} job better performed by {gender}?\"\n",
     "    answers[question] = []\n",
-    "    for i in range(10):\n",
+    "    for _i in range(10):\n",
     "        result = await attack.execute_async(objective=question, prepended_conversation=[system_prompt])  # type: ignore\n",
     "        answer = memory.get_conversation(conversation_id=result.conversation_id)[2].get_value()\n",
     "        if answer.startswith(\"YES\"):\n",
@@ -626,7 +626,7 @@
     "        ],\n",
     "    )\n",
     "    answers[question] = []\n",
-    "    for i in range(5):\n",
+    "    for _i in range(5):\n",
     "        result = await benchmark.execute_async(question_answering_entry=question_answering_entry)  # type: ignore\n",
     "        answers[question].append(str(result.outcome))\n",
     "\n",

doc/cookbooks/4_testing_bias.py

@@ -127,7 +127,7 @@
 for job, gender in itertools.product(jobs[:3], genders):  # for demonstration purposes we just use first 3 jobs
     question = f"Is the {job} job better performed by {gender}?"
     answers[question] = []
-    for i in range(10):
+    for _i in range(10):
         result = await attack.execute_async(objective=question, prepended_conversation=[system_prompt])  # type: ignore
         answer = memory.get_conversation(conversation_id=result.conversation_id)[2].get_value()
         if answer.startswith("YES"):
@@ -172,7 +172,7 @@
         ],
     )
     answers[question] = []
-    for i in range(5):
+    for _i in range(5):
         result = await benchmark.execute_async(question_answering_entry=question_answering_entry)  # type: ignore
         answers[question].append(str(result.outcome))
 

doc/generate_docs/pct_to_ipynb.py

@@ -87,7 +87,7 @@ def find_files(directory, file_extension):
         dir_path = os.path.join(directory, included_dir)
         if not os.path.exists(dir_path):
             continue
-        for root, dirs, files in os.walk(dir_path):
+        for root, _dirs, files in os.walk(dir_path):
             for file in files:
                 if file.endswith("_helpers.py"):
                     continue

pyproject.toml

@@ -252,6 +252,7 @@ fixable = [
     "YTT",
 ]
 select = [
+    "B",  # https://docs.astral.sh/ruff/rules/#flake8-bugbear-b
     "C4",  # https://docs.astral.sh/ruff/rules/#flake8-comprehensions-c4
     "CPY001",  # missing-copyright-notice
     "D",  # https://docs.astral.sh/ruff/rules/#pydocstyle-d
@@ -265,6 +266,7 @@ select = [
     "W",  # https://docs.astral.sh/ruff/rules/#pycodestyle-w
 ]
 ignore = [
+    "B903",  # class-as-data-structure (test helper classes use @apply_defaults pattern)
     "D100",  # Missing docstring in public module
     "D200",  # One-line docstring should fit on one line
     "D205",  # 1 blank line required between summary line and description
@@ -297,10 +299,12 @@ notice-rgx = "Copyright \\(c\\) Microsoft Corporation\\.\\s*\\n.*Licensed under
 # Temporary ignores for pyrit/ subdirectories until issue #1176
 # https://github.com/Azure/PyRIT/issues/1176 is fully resolved
 # TODO: Remove these ignores once the issues are fixed
-"pyrit/{auxiliary_attacks,ui}/**/*.py" = ["D101", "D102", "D103", "D104", "D105", "D106", "D107", "D401", "D404", "D417", "D418", "DOC102", "DOC201", "DOC202", "DOC402", "DOC501", "SIM101", "SIM108"]
+"pyrit/{auxiliary_attacks,ui}/**/*.py" = ["B905", "D101", "D102", "D103", "D104", "D105", "D106", "D107", "D401", "D404", "D417", "D418", "DOC102", "DOC201", "DOC202", "DOC402", "DOC501", "SIM101", "SIM108"]
 # Backend API routes raise HTTPException handled by FastAPI, not true exceptions
-"pyrit/backend/**/*.py" = ["DOC501"]
+"pyrit/backend/**/*.py" = ["DOC501", "B008"]
 "pyrit/__init__.py" = ["D104"]
+# Allow broad pytest.raises(Exception) in tests
+"tests/**/*.py" = ["B017"]
 
 [tool.ruff.lint.pydocstyle]
 convention = "google"

pyrit/auth/copilot_authenticator.py

@@ -304,7 +304,7 @@ async def _fetch_access_token_with_playwright(self) -> Optional[str]:
             raise RuntimeError(
                 "Playwright is not installed. Please install it with: "
                 "'pip install playwright && playwright install chromium'"
-            )
+            ) from None
 
         # On Windows, when using SelectorEventLoop (common in Jupyter), we need to run
         # Playwright in a separate thread with ProactorEventLoop to support subprocesses

pyrit/auth/manual_copilot_authenticator.py

@@ -63,7 +63,7 @@ def __init__(self, *, access_token: Optional[str] = None) -> None:
                 resolved_token, algorithms=["RS256"], options={"verify_signature": False}
             )
         except jwt.exceptions.DecodeError as e:
-            raise ValueError(f"Failed to decode access_token as JWT: {e}")
+            raise ValueError(f"Failed to decode access_token as JWT: {e}") from e
 
         required_claims = ["tid", "oid"]
         missing_claims = [claim for claim in required_claims if claim not in self._claims]

pyrit/backend/routes/attacks.py

@@ -153,7 +153,7 @@ async def create_attack(request: CreateAttackRequest) -> CreateAttackResponse:
         raise HTTPException(
             status_code=status.HTTP_404_NOT_FOUND,
             detail=str(e),
-        )
+        ) from e
 
 
 @router.get(
@@ -280,13 +280,13 @@ async def add_message(
             raise HTTPException(
                 status_code=status.HTTP_404_NOT_FOUND,
                 detail=error_msg,
-            )
+            ) from e
         raise HTTPException(
             status_code=status.HTTP_400_BAD_REQUEST,
             detail=error_msg,
-        )
+        ) from e
     except Exception as e:
         raise HTTPException(
             status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
             detail=f"Failed to add message: {str(e)}",
-        )
+        ) from e

pyrit/backend/routes/converters.py

@@ -67,12 +67,12 @@ async def create_converter(request: CreateConverterRequest) -> CreateConverterRe
         raise HTTPException(
             status_code=status.HTTP_400_BAD_REQUEST,
             detail=str(e),
-        )
+        ) from e
     except Exception as e:
         raise HTTPException(
             status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
             detail=f"Failed to create converter: {str(e)}",
-        )
+        ) from e
 
 
 @router.get(
@@ -126,9 +126,9 @@ async def preview_conversion(request: ConverterPreviewRequest) -> ConverterPrevi
         raise HTTPException(
             status_code=status.HTTP_400_BAD_REQUEST,
             detail=str(e),
-        )
+        ) from e
     except Exception as e:
         raise HTTPException(
             status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
             detail=f"Converter preview failed: {str(e)}",
-        )
+        ) from e

pyrit/backend/routes/targets.py

@@ -74,12 +74,12 @@ async def create_target(request: CreateTargetRequest) -> TargetInstance:
         raise HTTPException(
             status_code=status.HTTP_400_BAD_REQUEST,
             detail=str(e),
-        )
+        ) from e
     except Exception as e:
         raise HTTPException(
             status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
             detail=f"Failed to create target: {str(e)}",
-        )
+        ) from e
 
 
 @router.get(

pyrit/common/notebook_utils.py

@@ -13,7 +13,7 @@ def is_in_ipython_session() -> bool:
         bool: True if the code is running in an IPython session, False otherwise.
     """
     try:
-        __IPYTHON__  # type: ignore
+        __IPYTHON__  # type: ignore  # noqa: B018
         return True
     except NameError:
         return False

pyrit/common/yaml_loadable.py

@@ -12,7 +12,7 @@
 T = TypeVar("T", bound="YamlLoadable")
 
 
-class YamlLoadable(abc.ABC):
+class YamlLoadable(abc.ABC):  # noqa: B024
     """
     Abstract base class for objects that can be loaded from YAML files.
     """
@@ -36,10 +36,10 @@ def from_yaml_file(cls: type[T], file: Union[Path | str]) -> T:
         try:
             yaml_data = yaml.safe_load(file.read_text("utf-8"))
         except yaml.YAMLError as exc:
-            raise ValueError(f"Invalid YAML file '{file}': {exc}")
+            raise ValueError(f"Invalid YAML file '{file}': {exc}") from exc
 
         # If this class provides a from_dict factory, use it;
         # otherwise, just instantiate directly with **yaml_data
-        if hasattr(cls, "from_dict") and callable(getattr(cls, "from_dict")):
+        if hasattr(cls, "from_dict") and callable(getattr(cls, "from_dict")):  # noqa: B009
             return cls.from_dict(yaml_data)  # type: ignore
         return cls(**yaml_data)

pyrit/datasets/seed_datasets/remote/jbb_behaviors_dataset.py

@@ -121,7 +121,7 @@ async def fetch_dataset(self, *, cache: bool = True) -> SeedDataset:
 
         except Exception as e:
             logger.error(f"Failed to load JBB-Behaviors dataset: {str(e)}")
-            raise Exception(f"Error loading JBB-Behaviors dataset: {str(e)}")
+            raise Exception(f"Error loading JBB-Behaviors dataset: {str(e)}") from e
 
     def _map_jbb_category_to_harm_category(self, jbb_category: str) -> list[str]:
         """

pyrit/datasets/seed_datasets/remote/vlsu_multimodal_dataset.py

@@ -56,7 +56,7 @@ def __init__(
         source: str = "https://raw.githubusercontent.com/apple/ml-vlsu/main/data/VLSU.csv",
         source_type: Literal["public_url", "file"] = "public_url",
         categories: Optional[list[VLSUCategory]] = None,
-        unsafe_grades: Optional[list[str]] = ["unsafe", "borderline"],
+        unsafe_grades: Optional[list[str]] = None,
         max_examples: Optional[int] = None,
     ):
         """
@@ -77,6 +77,8 @@ def __init__(
         Raises:
             ValueError: If any of the specified categories are invalid.
         """
+        if unsafe_grades is None:
+            unsafe_grades = ["unsafe", "borderline"]
         self.source = source
         self.source_type: Literal["public_url", "file"] = source_type
         self.categories = categories

pyrit/datasets/seed_datasets/seed_dataset_provider.py

@@ -99,7 +99,7 @@ def get_all_dataset_names(cls) -> list[str]:
                 provider = provider_class()
                 dataset_names.add(provider.dataset_name)
             except Exception as e:
-                raise ValueError(f"Could not get dataset name from {provider_class.__name__}: {e}")
+                raise ValueError(f"Could not get dataset name from {provider_class.__name__}: {e}") from e
         return sorted(dataset_names)
 
     @classmethod

pyrit/executor/attack/core/attack_executor.py

@@ -332,7 +332,7 @@ def _process_execution_results(
         completed: list[AttackStrategyResultT] = []
         incomplete: list[tuple[str, BaseException]] = []
 
-        for objective, result in zip(objectives, results_or_exceptions):
+        for objective, result in zip(objectives, results_or_exceptions, strict=False):
             if isinstance(result, BaseException):
                 incomplete.append((objective, result))
             else:

pyrit/executor/attack/multi_turn/tree_of_attacks.py

@@ -1133,13 +1133,17 @@ def _parse_red_teaming_response(self, red_teaming_response: str) -> str:
             red_teaming_response_dict = json.loads(red_teaming_response)
         except json.JSONDecodeError:
             logger.error(f"The response from the red teaming chat is not in JSON format: {red_teaming_response}")
-            raise InvalidJsonException(message="The response from the red teaming chat is not in JSON format.")
+            raise InvalidJsonException(
+                message="The response from the red teaming chat is not in JSON format."
+            ) from None
 
         try:
             return cast(str, red_teaming_response_dict["prompt"])
         except KeyError:
             logger.error(f"The response from the red teaming chat does not contain a prompt: {red_teaming_response}")
-            raise InvalidJsonException(message="The response from the red teaming chat does not contain a prompt.")
+            raise InvalidJsonException(
+                message="The response from the red teaming chat does not contain a prompt."
+            ) from None
 
     def __str__(self) -> str:
         """

pyrit/executor/attack/single_turn/context_compliance.py

@@ -119,7 +119,7 @@ def _load_context_description_instructions(self, *, instructions_path: Path) ->
         try:
             context_description_instructions = SeedDataset.from_yaml_file(instructions_path)
         except Exception as e:
-            raise ValueError(f"Failed to load context description instructions from {instructions_path}: {e}")
+            raise ValueError(f"Failed to load context description instructions from {instructions_path}: {e}") from e
 
         if len(context_description_instructions.prompts) < 3:
             raise ValueError(

pyrit/executor/core/strategy.py

@@ -24,7 +24,7 @@
 
 
 @dataclass
-class StrategyContext(ABC):
+class StrategyContext(ABC):  # noqa: B024
     """Base class for all strategy contexts."""
 
     def duplicate(self: StrategyContextT) -> StrategyContextT:

pyrit/executor/promptgen/fuzzer/fuzzer_converter_base.py

@@ -136,7 +136,7 @@ async def send_prompt_async(self, request: Message) -> str:
             return str(parsed_response["output"])
 
         except json.JSONDecodeError:
-            raise InvalidJsonException(message=f"Invalid JSON encountered: {response_msg}")
+            raise InvalidJsonException(message=f"Invalid JSON encountered: {response_msg}") from None
 
     def input_supported(self, input_type: PromptDataType) -> bool:
         """

pyrit/memory/memory_embedding.py

@@ -82,5 +82,7 @@ def default_memory_embedding_factory(embedding_model: Optional[EmbeddingSupport]
     try:
         model = OpenAITextEmbedding()
         return MemoryEmbedding(embedding_model=model)
-    except ValueError:
-        raise ValueError("No embedding model was provided and no OpenAI embedding model was found in the environment.")
+    except ValueError as e:
+        raise ValueError(
+            "No embedding model was provided and no OpenAI embedding model was found in the environment."
+        ) from e