aiverify-moonshot 0.6.1__py3-none-any.whl → 0.6.3__py3-none-any.whl

moonshot/__main__.py

@@ -116,6 +116,13 @@ def download_nltk_resources() -> None:
             raise
 
 
+def download_spacy_model() -> None:
+    """
+    Downloads the en_core_web_lg model using the spacy module (for entity processor module).
+    """
+    subprocess.run(["python", "-m", "spacy", "download", "en_core_web_lg"])
+
+
 def moonshot_data_installation(unattended: bool, overwrite: bool) -> None:
     """
     Install Moonshot Data from GitHub.
@@ -175,10 +182,12 @@ def moonshot_data_installation(unattended: bool, overwrite: bool) -> None:
     if os.path.exists("requirements.txt"):
         run_subprocess(["pip", "install", "-r", "requirements.txt"], check=True)
         download_nltk_resources()
+        download_spacy_model()
 
     # Change back to the base directory
     os.chdir("..")
 
+
 def check_node() -> bool:
     """
     Check if Node.js is installed on the user's machine.

moonshot/integrations/cli/__main__.py

@@ -1,10 +1,9 @@
-import sys
 import warnings
 
 from moonshot.integrations.cli.cli import CommandLineInterface
 
 
-def start_app(cli_command = None):
+def start_app(cli_command=None):
     """
     Run the Moonshot application
     """
@@ -22,4 +21,3 @@ def start_app(cli_command = None):
     else:
         # Show help if no command is provided
         cli_instance.onecmd("help")
-              

moonshot/integrations/cli/redteam/session.py

@@ -58,7 +58,7 @@ def new_session(args) -> None:
             - context_strategy (str, optional): The context strategy to be used.
             - prompt_template (str, optional): The prompt template to be used.
             - endpoints (str, optional): The list of endpoints for the runner."""
-    global active_session
+    global active_session  # noqa: F824
 
     try:
         required_parameters = [("runner_id", str)]
@@ -136,7 +136,7 @@ def use_session(args) -> None:
     Args:
         args (Namespace): The arguments passed to the function.
     """
-    global active_session
+    global active_session  # noqa: F824
 
     # Load session metadata
     try:
@@ -168,7 +168,7 @@ def show_prompts() -> None:
     """
     Shows the chat table in a session so that users don't have to restart a session to view the chat table
     """
-    global active_session
+    global active_session  # noqa: F824
 
     if not active_session:
         print(ERROR_RED_TEAMING_SHOW_PROMPTS_NO_ACTIVE_SESSION_VALIDATION)
@@ -181,7 +181,7 @@ def end_session() -> None:
     """
     Ends the current session by clearing active_session variable.
     """
-    global active_session
+    global active_session  # noqa: F824
     active_session.clear()
 
 
@@ -249,7 +249,7 @@ def update_chat_display() -> None:
     The table includes columns for the chat ID, prepared prompts, and the prompt/response pairs.
     If there is no active session, a message is printed to the console.
     """
-    global active_session
+    global active_session  # noqa: F824
 
     if active_session:
         list_of_endpoint_chats = api_get_all_chats_from_session(
@@ -311,7 +311,7 @@ def add_bookmark(args) -> None:
 
     If there is no active session, a message is printed to the console and the function returns.
     """
-    global active_session
+    global active_session  # noqa: F824
 
     if active_session:
         try:
@@ -372,7 +372,7 @@ def use_bookmark(args) -> None:
 
     If there is no active session, a message is printed to the console and the function returns.
     """
-    global active_session
+    global active_session  # noqa: F824
     if active_session:
         try:
             bookmark_name = args.bookmark_name
@@ -747,7 +747,7 @@ def _reload_session(runner_id: str) -> None:
     Args:
         runner_id (str): The ID of the runner for which the session metadata needs to be reloaded.
     """
-    global active_session
+    global active_session  # noqa: F824
     try:
         session_metadata = api_load_session(runner_id)
         if not session_metadata:

moonshot/integrations/web_api/app.py

@@ -71,7 +71,7 @@ def create_app(cfg: providers.Configuration) -> CustomFastAPI:
     }
 
     app: CustomFastAPI = CustomFastAPI(
-        title="Project Moonshot", version="0.6.1", **app_kwargs
+        title="Project Moonshot", version="0.6.3", **app_kwargs
     )
 
     if cfg.cors.enabled():

moonshot/integrations/web_api/routes/benchmark_result.py

@@ -131,6 +131,7 @@ async def get_one_results(
                 status_code=500, detail=f"Failed to retrieve result: {e.msg}"
             )
 
+
 @router.delete("/api/v1/benchmarks/results/{result_id}")
 @inject
 def delete_result(

moonshot/integrations/web_api/routes/bookmark.py

@@ -1,4 +1,3 @@
-import os
 from typing import Optional
 
 from dependency_injector.wiring import Provide, inject
@@ -138,6 +137,8 @@ def delete_bookmark(
             raise HTTPException(
                 status_code=500, detail=f"Failed to delete bookmark: {e.msg}"
             )
+
+
 @router.get(
     "/api/v1/bookmarks/export", response_description="Exporting Bookmark to JSON file"
 )
@@ -160,7 +161,9 @@ def export_bookbookmarks(
     """
     try:
         file_path = bookmark_service.export_bookmarks(export_file_name)
-        return FileResponse(file_path, media_type="application/json", filename=export_file_name)
+        return FileResponse(
+            file_path, media_type="application/json", filename=export_file_name
+        )
     except ServiceException as e:
         if e.error_code == "FileNotFound":
             raise HTTPException(

moonshot/integrations/web_api/routes/context_strategy.py

@@ -7,6 +7,7 @@ from ..services.utils.exceptions_handler import ServiceException
 
 router = APIRouter(tags=["Context Strategy"])
 
+
 @router.get("/api/v1/context-strategies")
 @inject
 def get_all_context_strategies(
@@ -46,7 +47,8 @@ def get_all_context_strategies(
                 status_code=500,
                 detail=f"Failed to retrieve context strategies: {e.msg}",
             )
-        
+
+
 @router.get("/api/v1/context-strategies/name")
 @inject
 def get_all_context_strategies_name(

moonshot/integrations/web_api/routes/dataset.py

@@ -1,5 +1,8 @@
+import os
+import tempfile
+
 from dependency_injector.wiring import Provide, inject
-from fastapi import APIRouter, Depends, HTTPException
+from fastapi import APIRouter, Depends, File, Form, HTTPException, UploadFile
 
 from ..container import Container
 from ..schemas.dataset_create_dto import CSV_Dataset_DTO, HF_Dataset_DTO
@@ -10,10 +13,14 @@ from ..services.utils.exceptions_handler import ServiceException
 router = APIRouter(tags=["Datasets"])
 
 
-@router.post("/api/v1/datasets/csv")
+@router.post("/api/v1/datasets/file")
 @inject
-def convert_dataset(
-    dataset_data: CSV_Dataset_DTO,
+async def upload_dataset(
+    file: UploadFile = File(...),
+    name: str = Form(..., min_length=1),
+    description: str = Form(default="", min_length=1),
+    license: str = Form(default=""),
+    reference: str = Form(default=""),
     dataset_service: DatasetService = Depends(Provide[Container.dataset_service]),
 ) -> str:
     """
@@ -32,7 +39,24 @@ def convert_dataset(
                        An error with status code 400 if there is a validation error.
                        An error with status code 500 for any other server-side error.
     """
+
+    # Create a temporary file with a secure random name
+    with tempfile.NamedTemporaryFile(
+        delete=False, suffix=os.path.splitext(file.filename)[1]
+    ) as tmp_file:
+        content = await file.read()
+        tmp_file.write(content)
+        temp_file_path = tmp_file.name
+
     try:
+        # Create the DTO with the form data including optional fields
+        dataset_data = CSV_Dataset_DTO(
+            name=name,
+            description=description,
+            license=license,
+            reference=reference,
+            file_path=temp_file_path,
+        )
         return dataset_service.convert_dataset(dataset_data)
     except ServiceException as e:
         if e.error_code == "FileNotFound":
@@ -47,6 +71,10 @@ def convert_dataset(
             raise HTTPException(
                 status_code=500, detail=f"Failed to convert dataset: {e.msg}"
             )
+    finally:
+        # Clean up the temporary file
+        if os.path.exists(temp_file_path):
+            os.unlink(temp_file_path)
 
 
 @router.post("/api/v1/datasets/hf")

moonshot/integrations/web_api/routes/prompt_template.py

@@ -87,6 +87,7 @@ def get_all_prompt_templates_names(
                 detail=f"Failed to retrieve prompt template names: {e.msg}",
             )
 
+
 @router.delete("/api/v1/prompt-templates/{prompt_template_name}")
 @inject
 def delete_prompt_template(

moonshot/integrations/web_api/schemas/cookbook_create_dto.py

@@ -1,9 +1,11 @@
 from typing import Optional
 
+from pydantic import Field
+
 from moonshot.src.cookbooks.cookbook_arguments import (
     CookbookArguments as CookbookPydanticModel,
 )
-from pydantic import Field
+
 
 class CookbookCreateDTO(CookbookPydanticModel):
     id: Optional[str] = None
@@ -20,4 +22,4 @@ class CookbookUpdateDTO(CookbookPydanticModel):
     description: Optional[str] = Field(default=None, min_length=1)
     tags: Optional[list[str]] = None
     categories: Optional[list[str]] = None
-    recipes: Optional[list[str]] = Field(default=None, min_length=1)
+    recipes: Optional[list[str]] = Field(default=None, min_length=1)

moonshot/integrations/web_api/schemas/dataset_create_dto.py

@@ -8,13 +8,13 @@ from moonshot.src.datasets.dataset_arguments import (
 
 
 class CSV_Dataset_DTO(DatasetPydanticModel):
-    id: Optional[str] = None  # Not a required from user
-    examples: Optional[Any] = None  # Not a required from user
+    id: Optional[str] = None  # Not required from user
+    examples: Optional[Any] = None  # Not required from user
     name: str = Field(..., min_length=1)
     description: str = Field(default="", min_length=1)
     license: Optional[str] = ""
     reference: Optional[str] = ""
-    csv_file_path: str = Field(..., min_length=1)
+    file_path: str = Field(..., min_length=1)
 
 
 class HF_Dataset_DTO(DatasetPydanticModel):

moonshot/integrations/web_api/schemas/prompt_response_model.py

@@ -3,7 +3,6 @@ from pydantic import BaseModel, RootModel
 from ..schemas.session_response_model import ChatRecord
 
 
-
 class PromptInfo(BaseModel):
     current_runner_id: str
     current_chats: dict[str, list[ChatRecord]]

moonshot/integrations/web_api/schemas/recipe_create_dto.py

@@ -1,8 +1,9 @@
 from typing import Optional
 
+from pydantic import Field
+
 from moonshot.src.recipes.recipe_arguments import RecipeArguments as RecipePydanticModel
 
-from pydantic import Field
 
 class RecipeCreateDTO(RecipePydanticModel):
     id: Optional[str] = None

moonshot/integrations/web_api/services/context_strategy_service.py

@@ -1,17 +1,14 @@
-from typing import Any
-
 from .... import api as moonshot_api
 from ..services.utils.exceptions_handler import exception_handler
 from .base_service import BaseService
 
 
 class ContextStrategyService(BaseService):
-
     @exception_handler
     def get_ctx_strategy(self) -> list[dict]:
         strategies = moonshot_api.api_get_all_context_strategy_metadata()
         return strategies
-    
+
     @exception_handler
     def get_ctx_strategy_name(self) -> list[str]:
         strategies = moonshot_api.api_get_all_context_strategies()

moonshot/integrations/web_api/services/cookbook_service.py

@@ -200,8 +200,6 @@ def cookbook_has_categories(categories: str, cookbook: Cookbook) -> bool:
     )
 
 
-
-
 @staticmethod
 def cookbook_metrics_dependency(cookbook: Cookbook) -> dict | None:
     """

moonshot/integrations/web_api/services/dataset_service.py

@@ -4,6 +4,7 @@ from ..schemas.dataset_response_dto import DatasetResponseDTO
 from ..services.base_service import BaseService
 from ..services.utils.exceptions_handler import exception_handler
 from .utils.file_manager import copy_file
+import os
 
 
 class DatasetService(BaseService):
@@ -16,7 +17,7 @@ class DatasetService(BaseService):
             dataset_data (CSV_Dataset_DTO): The data required to convert the dataset.
 
         Returns:
-            str: The path to the newly created dataset.
+            str: The filename of the newly created dataset.
 
         Raises:
             Exception: If an error occurs during dataset conversion.
@@ -27,9 +28,9 @@ class DatasetService(BaseService):
             description=dataset_data.description,
             reference=dataset_data.reference,
             license=dataset_data.license,
-            csv_file_path=dataset_data.csv_file_path,
+            file_path=dataset_data.file_path,
         )
-        return copy_file(new_ds_path)
+        return os.path.splitext(os.path.basename(new_ds_path))[0]
 
     @exception_handler
     def download_dataset(self, dataset_data: HF_Dataset_DTO) -> str:

moonshot/integrations/web_api/services/session_service.py

@@ -100,15 +100,15 @@ class SessionService(BaseService):
         retn_session = []
         runners_with_session = moonshot_api.api_get_all_runner()
         sessions_metadata_dicts = moonshot_api.api_get_all_session_metadata()
-        
-        runners_dict = {runner['id']: runner for runner in runners_with_session}
-        
+
+        runners_dict = {runner["id"]: runner for runner in runners_with_session}
+
         for session in sessions_metadata_dicts:
             sess_id = session.get("session_id")
             if sess_id in runners_dict:
-                session['description'] = runners_dict[sess_id]['description']
+                session["description"] = runners_dict[sess_id]["description"]
                 retn_session.append(session)
-        
+
         return [
             SessionMetadataModel(**metadata) for metadata in sessions_metadata_dicts
         ]

moonshot/integrations/web_api/services/utils/exceptions_handler.py

@@ -1,41 +1,78 @@
 import asyncio
 from typing import Any, Callable
+
 from pydantic import ValidationError
 
+
 class ServiceException(Exception):
     error_code: str
     msg: str
 
-    def __init__(self, msg: str, method_name: str, error_code: str = 'UnknownSessionError'):
+    def __init__(
+        self, msg: str, method_name: str, error_code: str = "UnknownSessionError"
+    ):
         self.error_code = error_code
         message = f"[ServiceException] {error_code} in {method_name} - {msg}"
-        self.msg = message 
+        self.msg = message
         super().__init__(message)
 
+
 def exception_handler(func: Callable[..., Any]) -> Callable[..., Any]:
     if asyncio.iscoroutinefunction(func):
+
         async def async_wrapper(*args: Any, **kwargs: Any) -> Any:
             try:
                 return await func(*args, **kwargs)
             except FileNotFoundError as e:
-                raise ServiceException(f"A file not found error occurred: {e}", func.__name__, "FileNotFound")
+                raise ServiceException(
+                    f"A file not found error occurred: {e}",
+                    func.__name__,
+                    "FileNotFound",
+                )
             except ValidationError as e:
-                raise ServiceException(f"A validation error occurred: {e}", func.__name__, "ValidationError")
+                raise ServiceException(
+                    f"A validation error occurred: {e}",
+                    func.__name__,
+                    "ValidationError",
+                )
             except ValueError as e:
-                raise ServiceException(f"An value error occurred: {e}", func.__name__, "ValueError")
+                raise ServiceException(
+                    f"An value error occurred: {e}", func.__name__, "ValueError"
+                )
             except Exception as e:
-                raise ServiceException(f"An unexpected error occurred: {e}", func.__name__, "UnexpectedError")
+                raise ServiceException(
+                    f"An unexpected error occurred: {e}",
+                    func.__name__,
+                    "UnexpectedError",
+                )
+
         return async_wrapper
     else:
+
         def wrapper(*args: Any, **kwargs: Any) -> Any:
             try:
                 return func(*args, **kwargs)
             except FileNotFoundError as e:
-                raise ServiceException(f"A file not found error occurred: {e}", func.__name__, "FileNotFound")
+                raise ServiceException(
+                    f"A file not found error occurred: {e}",
+                    func.__name__,
+                    "FileNotFound",
+                )
             except ValidationError as e:
-                raise ServiceException(f"A validation error occurred: {e}", func.__name__, "ValidationError")
+                raise ServiceException(
+                    f"A validation error occurred: {e}",
+                    func.__name__,
+                    "ValidationError",
+                )
             except ValueError as e:
-                raise ServiceException(f"An value error occurred: {e}", func.__name__, "ValueError")
+                raise ServiceException(
+                    f"An value error occurred: {e}", func.__name__, "ValueError"
+                )
             except Exception as e:
-                raise ServiceException(f"An unexpected error occurred: {e}", func.__name__, "UnexpectedError")
+                raise ServiceException(
+                    f"An unexpected error occurred: {e}",
+                    func.__name__,
+                    "UnexpectedError",
+                )
+
         return wrapper

moonshot/integrations/web_api/services/utils/results_formatter.py

@@ -1,18 +1,20 @@
-def transform_web_format(original_dict: dict[str, dict[str, list[dict[str, dict[str, float]]]]]) -> dict[str, dict[str, list[dict[str, list[dict[str, list[dict[str, float]]]]]]]]:
+def transform_web_format(
+    original_dict: dict[str, dict[str, list[dict[str, dict[str, float]]]]]
+) -> dict[str, dict[str, list[dict[str, list[dict[str, list[dict[str, float]]]]]]]]:
     transformed = {"metadata": original_dict["metadata"], "results": {"cookbooks": []}}
-    
+
     for cookbook_name, cookbook_data in original_dict.items():
         if cookbook_name == "metadata":
             continue  # Skip the metadata entry
-        
+
         cookbook_entry = {"id": cookbook_name, "recipes": []}
         for recipe_name, recipe_data in cookbook_data.items():
             recipe_entry = {"id": recipe_name, "models": []}
-            
+
             for model_data_str, model_data in recipe_data.items():
                 # Extract model, recipe, dataset, prompt_template from the string
                 model, _, dataset, prompt_template = eval(model_data_str)
-                
+
                 model_entry = {
                     "id": model,
                     "datasets": [
@@ -21,27 +23,34 @@ def transform_web_format(original_dict: dict[str, dict[str, list[dict[str, dict[
                             "prompt_templates": [
                                 {
                                     "id": prompt_template,
-                                    "metrics": model_data["results"]
+                                    "metrics": model_data["results"],
                                 }
-                            ]
+                            ],
                         }
-                    ]
+                    ],
                 }
-                
+
                 # Check if the model already exists in the recipe_entry
-                existing_model = next((m for m in recipe_entry["models"] if m["id"] == model), None)
+                existing_model = next(
+                    (m for m in recipe_entry["models"] if m["id"] == model), None
+                )
                 if existing_model:
                     # Check if the dataset already exists in the model
-                    existing_dataset = next((d for d in existing_model["datasets"] if d["id"] == dataset), None)
+                    existing_dataset = next(
+                        (d for d in existing_model["datasets"] if d["id"] == dataset),
+                        None,
+                    )
                     if existing_dataset:
-                        existing_dataset["prompt_templates"].append(model_entry["datasets"][0]["prompt_templates"][0])
+                        existing_dataset["prompt_templates"].append(
+                            model_entry["datasets"][0]["prompt_templates"][0]
+                        )
                     else:
                         existing_model["datasets"].append(model_entry["datasets"][0])
                 else:
                     recipe_entry["models"].append(model_entry)
-            
+
             cookbook_entry["recipes"].append(recipe_entry)
-        
+
         transformed["results"]["cookbooks"].append(cookbook_entry)
-    
-    return transformed
+
+    return transformed

moonshot/integrations/web_api/status_updater/interface/benchmark_progress_callback.py

@@ -1,14 +1,14 @@
 from abc import ABC, abstractmethod
+
 from ...types.types import TestRunProgress
 
+
 class InterfaceBenchmarkProgressCallback(ABC):
     """
     The interface to implement for handling of benchmark progress updates from moonshot library
 
     """
+
     @abstractmethod
     def on_progress_update(self, progress_data: TestRunProgress) -> None:
         pass
-
-
-

moonshot/integrations/web_api/status_updater/interface/redteam_progress_callback.py

@@ -1,14 +1,14 @@
 from abc import ABC, abstractmethod
+
 from ...types.types import TestRunProgress
 
+
 class InterfaceRedTeamProgressCallback(ABC):
     """
     The interface to implement for handling of redteam progress updates from moonshot library
 
     """
+
     @abstractmethod
     def on_art_progress_update(self, progress_data: TestRunProgress) -> None:
         pass
-
-
-