aiverify-moonshot 0.4.9__py3-none-any.whl → 0.4.11__py3-none-any.whl

{aiverify_moonshot-0.4.9.dist-info → aiverify_moonshot-0.4.11.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: aiverify-moonshot
-Version: 0.4.9
+Version: 0.4.11
 Summary: AI Verify advances Gen AI testing with Project Moonshot.
 Project-URL: Repository, https://github.com/aiverify-foundation/moonshot
 Project-URL: Documentation, https://aiverify-foundation.github.io/moonshot/
@@ -15,38 +15,38 @@ Classifier: License :: OSI Approved :: Apache Software License
 Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.11
 Requires-Python: >=3.11
-Requires-Dist: datasets==2.21.0
-Requires-Dist: ijson==3.3.0
-Requires-Dist: jinja2==3.1.4
-Requires-Dist: numpy==1.26.4
-Requires-Dist: pandas==2.2.2
+Requires-Dist: datasets>=2.21.0
+Requires-Dist: ijson>=3.3.0
+Requires-Dist: jinja2>=3.1.4
+Requires-Dist: numpy>=1.26.4
+Requires-Dist: pandas>=2.2.2
 Requires-Dist: pydantic==2.8.2
-Requires-Dist: pyparsing==3.1.4
-Requires-Dist: python-dotenv==1.0.1
-Requires-Dist: python-slugify==8.0.4
-Requires-Dist: xxhash==3.5.0
+Requires-Dist: pyparsing>=3.1.4
+Requires-Dist: python-dotenv>=1.0.1
+Requires-Dist: python-slugify>=8.0.4
+Requires-Dist: xxhash>=3.5.0
 Provides-Extra: all
-Requires-Dist: cmd2==2.4.3; extra == 'all'
-Requires-Dist: dependency-injector==4.41.0; extra == 'all'
-Requires-Dist: fastapi==0.112.2; extra == 'all'
-Requires-Dist: rich==13.8.0; extra == 'all'
-Requires-Dist: typing-extensions==4.12.2; extra == 'all'
-Requires-Dist: uvicorn==0.30.6; extra == 'all'
+Requires-Dist: cmd2>=2.4.3; extra == 'all'
+Requires-Dist: dependency-injector>=4.41.0; extra == 'all'
+Requires-Dist: fastapi>=0.112.2; extra == 'all'
+Requires-Dist: rich>=13.8.0; extra == 'all'
+Requires-Dist: typing-extensions>=4.12.2; extra == 'all'
+Requires-Dist: uvicorn>=0.30.6; extra == 'all'
 Provides-Extra: cli
-Requires-Dist: cmd2==2.4.3; extra == 'cli'
-Requires-Dist: rich==13.8.0; extra == 'cli'
+Requires-Dist: cmd2>=2.4.3; extra == 'cli'
+Requires-Dist: rich>=13.8.0; extra == 'cli'
 Provides-Extra: web-api
-Requires-Dist: dependency-injector==4.41.0; extra == 'web-api'
-Requires-Dist: fastapi==0.112.2; extra == 'web-api'
-Requires-Dist: typing-extensions==4.12.2; extra == 'web-api'
-Requires-Dist: uvicorn==0.30.6; extra == 'web-api'
+Requires-Dist: dependency-injector>=4.41.0; extra == 'web-api'
+Requires-Dist: fastapi>=0.112.2; extra == 'web-api'
+Requires-Dist: typing-extensions>=4.12.2; extra == 'web-api'
+Requires-Dist: uvicorn>=0.30.6; extra == 'web-api'
 Description-Content-Type: text/markdown
 
 <div align="center">
 
 ![Moonshot Logo](https://github.com/aiverify-foundation/moonshot/raw/main/misc/aiverify-moonshot-logo.png)
 
-**Version 0.4.9**
+**Version 0.4.11**
 
 A simple and modular tool to evaluate any LLM application.
 

{aiverify_moonshot-0.4.9.dist-info → aiverify_moonshot-0.4.11.dist-info}/RECORD

@@ -9,7 +9,7 @@ moonshot/integrations/cli/cli.py,sha256=9tnzcxcSOjblxCUpyh3pK0ke0bLs3s-63OxXtYoZ
 moonshot/integrations/cli/cli_errors.py,sha256=wDwBtxZcHPWIJa6ZVW0MoWm81aeqmhYhi7SLI7TQ4_4,23565
 moonshot/integrations/cli/benchmark/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 moonshot/integrations/cli/benchmark/benchmark.py,sha256=QUxr6DU11-XeH6Y3j1uPsZsotshgy64G_cWNf0Rn2_U,6303
-moonshot/integrations/cli/benchmark/cookbook.py,sha256=h9exPhLKLRa9a1qWB2WptZ35gcVov332tjHDO9ECs0o,29639
+moonshot/integrations/cli/benchmark/cookbook.py,sha256=pgXTKqYAeDj0tLsCN71SCrKM4DDgswwqxJhXaPQmAeU,29518
 moonshot/integrations/cli/benchmark/datasets.py,sha256=Uq5XMNWUp775sz9jCZUZHHmkumPFI7cHVRueHgWm70Q,8965
 moonshot/integrations/cli/benchmark/metrics.py,sha256=SHs-hIa4CIPyOJtxK2U4D6IRHy3ZNsRtZlAMGvF9Qxw,8310
 moonshot/integrations/cli/benchmark/recipe.py,sha256=yVFX3pwNzDL0a95rjlQd4cek06M9blv5jIeYB26jOKk,32481
@@ -34,7 +34,7 @@ moonshot/integrations/cli/utils/process_data.py,sha256=QVL5vp2_8ZgGicmCAdeYEHkeb
 moonshot/integrations/web_api/.env.dev,sha256=0z5_Ut8rF-UqFZtgjkH2qoqORhD5_nSs2w_OeX2SteI,182
 moonshot/integrations/web_api/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 moonshot/integrations/web_api/__main__.py,sha256=MdnLi_ZF-olAAEJwTPU1iGYFYwo-fNWNT2qfchkH3y4,2050
-moonshot/integrations/web_api/app.py,sha256=IucbtkC1n2l0Mx55St0e7FFWa_hV7Mdg64Sv9AZ41Ro,3651
+moonshot/integrations/web_api/app.py,sha256=x9yYCUE78d8uNE6ws87236mHCYTSHqy2KWQdWZey0ns,3652
 moonshot/integrations/web_api/container.py,sha256=DVkJG_qm7ItcG6tgMYOqIj07wpKhPWOOfy6-bEv72y4,5915
 moonshot/integrations/web_api/logging_conf.py,sha256=t3EGRV6tZhV732KXe8_Tiy0fiwVAWxZX5Tt8VTgrrfg,3388
 moonshot/integrations/web_api/log/.gitkeep,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -113,13 +113,14 @@ moonshot/src/api/api_result.py,sha256=M5zKF7ytKp237UZusLSYJ7QVfui85Ys0WEaYySGcAK
 moonshot/src/api/api_run.py,sha256=3PrETAVcFnJ09R0-xhWiFkEfqL6eYj4B2voEGJDPznU,2936
 moonshot/src/api/api_runner.py,sha256=cH0rxWREjc2qKmt4Tuwr-fEMrYDBE_TKRw0jOohNEgU,4179
 moonshot/src/api/api_session.py,sha256=OGH05ZxAwo_hKI-RNaJ-jCp_v-zcTm-9bHUclpq2z4Q,10978
-moonshot/src/bookmark/bookmark.py,sha256=T9dUGtw_l_3PeZT0iO2vFzPiSntX9MFurE9Gcop_JcM,11649
+moonshot/src/bookmark/bookmark.py,sha256=Gf9wZ5wmJUmatHChr0_3tVyUbPMfpCVeEfujQQjIv18,11867
 moonshot/src/bookmark/bookmark_arguments.py,sha256=cB5m2zB8255WVdacmC2-ZYNyaoK4-gOM_Qwb_JDR-34,1449
 moonshot/src/configs/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 moonshot/src/configs/env_variables.py,sha256=eF__UJN37LCzIB4pv_T7G-kQHlOa657QA7IpL1d_0MM,7150
 moonshot/src/connectors/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-moonshot/src/connectors/connector.py,sha256=e0ZCX9m83ezjMiY5H8gbWE64IStsNQqxzVwGtMwShPY,13396
-moonshot/src/connectors/connector_prompt_arguments.py,sha256=cIlAgbFk2g_XUZ0stVM904Ng2g4GYP2LyiAjktmhEQM,470
+moonshot/src/connectors/connector.py,sha256=MkOtC2pr5SwHp8xe5oRhC77BEfNmo883LsBhl4VpKnQ,13555
+moonshot/src/connectors/connector_prompt_arguments.py,sha256=gZ5jKwyZOPhMYQOfRE-Uyo4H4KoP1ngvXIXDcwEkkG0,630
+moonshot/src/connectors/connector_response.py,sha256=UHrX5tTDGrfPUXoIvPeMUOEXqJxlAo8vgGMHBPzDc0k,712
 moonshot/src/connectors_endpoints/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 moonshot/src/connectors_endpoints/connector_endpoint.py,sha256=hrSvq7iw1IigZssKTcdotkcN8stG8Ov7AlJ8xxZcAJo,9526
 moonshot/src/connectors_endpoints/connector_endpoint_arguments.py,sha256=8WyD0EfrlhGM9rqrzoAcF-YWmSWiWdfpO91E7w6HPY0,2318
@@ -139,7 +140,7 @@ moonshot/src/recipes/recipe.py,sha256=PE4mimP7Y7lDanXib2uMrd9KpLWOxRdg0Cps5rLRQx
 moonshot/src/recipes/recipe_arguments.py,sha256=9LdVJidB6Sci941QvnudBznvx9_NVCl8r-HvzSZInlY,3914
 moonshot/src/redteaming/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 moonshot/src/redteaming/attack/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-moonshot/src/redteaming/attack/attack_module.py,sha256=tbVfLs8FkBSQwpbWMkx6Ad7qrQt778boSG6h68lsLQk,26827
+moonshot/src/redteaming/attack/attack_module.py,sha256=R2nfAKiPlEpgtWPHPANyciBxfiKg4YCEusvOg8p-EYM,27905
 moonshot/src/redteaming/attack/attack_module_arguments.py,sha256=L8H6poNj7xNWLUr6jfuSkwE7TNV0aTPaZBal4OvL7IA,1225
 moonshot/src/redteaming/attack/context_strategy.py,sha256=KiAw85Pl8tuTzQ2d3CkZrjn3Ddp7BnlT5Gc67TTnq4M,4821
 moonshot/src/redteaming/context_strategy/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -158,7 +159,7 @@ moonshot/src/runners/runner_arguments.py,sha256=Bg4OPSmgr9jZKNAwPH0T3epEHw-6qGrf
 moonshot/src/runners/runner_type.py,sha256=jOfnAnaCYp-rPTRJXhM8hin_dinlR0sMwmimQXvLcJ0,100
 moonshot/src/runs/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 moonshot/src/runs/run.py,sha256=3Y_256IfOv5zpeujVPsA5YdX7uZjGMRvg0VencUwgLw,14339
-moonshot/src/runs/run_arguments.py,sha256=G043ERvHIU_dd0JghboZgxDWCcjYOaDwue1ieDunDKA,6443
+moonshot/src/runs/run_arguments.py,sha256=rYMyAJrMYNwQq1yPSEXFv7fe6h17ZH37DXKEOixYEko,7450
 moonshot/src/runs/run_progress.py,sha256=d1BcNo6Kp4vA165TDx_xebl8JDo92aV-YutPxsgCOxE,6495
 moonshot/src/runs/run_status.py,sha256=TRtizcDzPxf6aQ2c3OovM6IQKJ0VCBhqDWvn7UBw5Zg,251
 moonshot/src/storage/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -171,9 +172,9 @@ moonshot/src/utils/import_modules.py,sha256=T9zTN59PFnvY2rjyWhSV9KSIAHxWV1pyBemF
 moonshot/src/utils/log.py,sha256=YNgD7Eh2OT36XlmVBKCGUTAh9TRp4Akfe4kDdvHASgs,2502
 moonshot/src/utils/pagination.py,sha256=5seymyRoqyENIhKllAatr1T91kMCGFslcvRnJHyMSvc,814
 moonshot/src/utils/timeit.py,sha256=TvuF0w8KWhp0oZFY0cUU3UY0xlGKjchb0OkfYfgVTlc,866
-aiverify_moonshot-0.4.9.dist-info/METADATA,sha256=vuUKDpxqkTsxqeBCGZ6hAP5LcHl0xyZFklGL1-Fyv7Q,12388
-aiverify_moonshot-0.4.9.dist-info/WHEEL,sha256=1yFddiXMmvYK7QYTqtRNtX66WJ0Mz8PYEiEUoOUUxRY,87
-aiverify_moonshot-0.4.9.dist-info/licenses/AUTHORS.md,sha256=mmAbe3i3sT8JZHJMBhxp3i1xRehV0g7WB4T_eyIBuBs,59
-aiverify_moonshot-0.4.9.dist-info/licenses/LICENSE.md,sha256=53izDRmJZZCjpYGfyLqlxnGQN-aNWBxasuzuMXC5Ias,11347
-aiverify_moonshot-0.4.9.dist-info/licenses/NOTICES.md,sha256=vS1zZYAnGjCJdwQ13xv3b2zc30wOS98ZnCKluT-AhHs,123266
-aiverify_moonshot-0.4.9.dist-info/RECORD,,
+aiverify_moonshot-0.4.11.dist-info/METADATA,sha256=Yf_5kXixSCJzVEeC5vRW3uId1RKe1XLDLnLabE0S_AQ,12390
+aiverify_moonshot-0.4.11.dist-info/WHEEL,sha256=1yFddiXMmvYK7QYTqtRNtX66WJ0Mz8PYEiEUoOUUxRY,87
+aiverify_moonshot-0.4.11.dist-info/licenses/AUTHORS.md,sha256=mmAbe3i3sT8JZHJMBhxp3i1xRehV0g7WB4T_eyIBuBs,59
+aiverify_moonshot-0.4.11.dist-info/licenses/LICENSE.md,sha256=53izDRmJZZCjpYGfyLqlxnGQN-aNWBxasuzuMXC5Ias,11347
+aiverify_moonshot-0.4.11.dist-info/licenses/NOTICES.md,sha256=vS1zZYAnGjCJdwQ13xv3b2zc30wOS98ZnCKluT-AhHs,123266
+aiverify_moonshot-0.4.11.dist-info/RECORD,,

moonshot/integrations/cli/benchmark/cookbook.py

@@ -467,7 +467,10 @@ def _display_view_cookbook(cookbook_info):
     recipes_list = api_read_recipes(recipes)
     if recipes_list:
         table = Table(
-            title="View Cookbook", show_lines=True, expand=True, header_style="bold"
+            title=f'Cookbook "{name}"',
+            show_lines=True,
+            expand=True,
+            header_style="bold",
         )
         table.add_column("No.", width=2)
         table.add_column("Recipe", justify="left", width=78)
@@ -482,7 +485,6 @@ def _display_view_cookbook(cookbook_info):
                 datasets,
                 prompt_templates,
                 metrics,
-                attack_strategies,
                 grading_scale,
                 stats,
             ) = recipe.values()
@@ -494,9 +496,6 @@ def _display_view_cookbook(cookbook_info):
                 "Prompt Templates", prompt_templates
             )
             metrics_info = display_view_list_format("Metrics", metrics)
-            attack_strategies_info = display_view_list_format(
-                "Attack Strategies", attack_strategies
-            )
             grading_scale_info = _display_view_grading_scale_format(
                 "Grading Scale", grading_scale
             )
@@ -506,7 +505,9 @@ def _display_view_cookbook(cookbook_info):
                 f"[red]id: {id}[/red]\n\n[blue]{name}[/blue]\n{description}\n\n"
                 f"{tags_info}\n\n{categories_info}\n\n{grading_scale_info}\n\n{stats_info}"
             )
-            contains_info = f"{datasets_info}\n\n{prompt_templates_info}\n\n{metrics_info}\n\n{attack_strategies_info}"
+            contains_info = (
+                f"{datasets_info}\n\n{prompt_templates_info}\n\n{metrics_info}"
+            )
 
             table.add_section()
             table.add_row(str(recipe_id), recipe_info, contains_info)

moonshot/integrations/web_api/app.py

@@ -71,7 +71,7 @@ def create_app(cfg: providers.Configuration) -> CustomFastAPI:
     }
 
     app: CustomFastAPI = CustomFastAPI(
-        title="Project Moonshot", version="0.4.9", **app_kwargs
+        title="Project Moonshot", version="0.4.11", **app_kwargs
     )
 
     if cfg.cors.enabled():

moonshot/src/bookmark/bookmark.py

@@ -11,9 +11,9 @@ from moonshot.src.messages_constants import (
     BOOKMARK_ADD_BOOKMARK_VALIDATION_ERROR,
     BOOKMARK_DELETE_ALL_BOOKMARK_ERROR,
     BOOKMARK_DELETE_ALL_BOOKMARK_SUCCESS,
-    BOOKMARK_DELETE_BOOKMARK_FAIL,
     BOOKMARK_DELETE_BOOKMARK_ERROR,
     BOOKMARK_DELETE_BOOKMARK_ERROR_1,
+    BOOKMARK_DELETE_BOOKMARK_FAIL,
     BOOKMARK_DELETE_BOOKMARK_SUCCESS,
     BOOKMARK_EXPORT_BOOKMARK_ERROR,
     BOOKMARK_EXPORT_BOOKMARK_VALIDATION_ERROR,
@@ -186,7 +186,9 @@ class Bookmark:
             if (
                 bookmark_info is not None
                 and isinstance(bookmark_info, tuple)
-                and all(isinstance(item, str) for item in bookmark_info)
+                and all(
+                    isinstance(item, str) for item in bookmark_info[1:]
+                )  # Check if the rest are strings besides id
             ):
                 return BookmarkArguments.from_tuple_to_dict(bookmark_info)
             else:
@@ -210,9 +212,11 @@ class Bookmark:
         """
         if isinstance(bookmark_name, str) and bookmark_name:
             try:
-
                 bookmark_info = Storage.read_database_record(
-                self.db_instance, (bookmark_name,), Bookmark.sql_select_bookmark_record)
+                    self.db_instance,
+                    (bookmark_name,),
+                    Bookmark.sql_select_bookmark_record,
+                )
                 if bookmark_info is not None:
                     sql_delete_bookmark_record = textwrap.dedent(
                         f"""
@@ -222,7 +226,10 @@ class Bookmark:
                     Storage.delete_database_record_in_table(
                         self.db_instance, sql_delete_bookmark_record
                     )
-                    return {"success": True, "message": BOOKMARK_DELETE_BOOKMARK_SUCCESS}
+                    return {
+                        "success": True,
+                        "message": BOOKMARK_DELETE_BOOKMARK_SUCCESS,
+                    }
                 else:
                     return {"success": False, "message": BOOKMARK_DELETE_BOOKMARK_FAIL}
             except Exception as e:

moonshot/src/connectors/connector.py

@@ -10,6 +10,7 @@ from typing import Callable
 
 from moonshot.src.configs.env_variables import EnvVariables
 from moonshot.src.connectors.connector_prompt_arguments import ConnectorPromptArguments
+from moonshot.src.connectors.connector_response import ConnectorResponse
 from moonshot.src.connectors_endpoints.connector_endpoint_arguments import (
     ConnectorEndpointArguments,
 )
@@ -168,9 +169,9 @@ class Connector:
         return wrapper
 
     @abstractmethod
-    async def get_response(self, prompt: str) -> str:
+    async def get_response(self, prompt: str) -> ConnectorResponse:
         """
-        Abstract method to be implemented by subclasses to get a response from the connector.
+        Abstract method to be implemented by subclasses to obtain a response from the connector.
 
         This method should asynchronously send a prompt to the connector's API and return the response.
 
@@ -178,7 +179,8 @@ class Connector:
             prompt (str): The input prompt to be sent to the connector.
 
         Returns:
-            str: The response received from the connector.
+            ConnectorResponse: An instance containing the response received from the connector and any
+            additional metadata.
         """
         pass
 
@@ -277,7 +279,7 @@ class Connector:
         object, which is used to generate the prediction. The method also optionally takes a `prompt_callback` function,
         which is called after the prediction is generated.
 
-        The method first prints a message indicating that it is predicting the prompt. It then records the start time
+        The method logs a message indicating that it is predicting the prompt. It then records the start time
         and uses the `connector` to generate a prediction for the `generated_prompt`. The duration of the prediction
         is calculated and stored in the `generated_prompt`.
 

moonshot/src/connectors/connector_prompt_arguments.py

@@ -4,14 +4,21 @@ from typing import Any
 
 from pydantic import BaseModel
 
+from moonshot.src.connectors.connector_response import ConnectorResponse
+
 
 class ConnectorPromptArguments(BaseModel):
+    class Config:
+        arbitrary_types_allowed = True
+
     prompt_index: int  # The index of the prompt in the dataset
 
     prompt: str  # The actual prompt text
 
     target: Any  # The target response for the prompt
 
-    predicted_results: Any = ""  # The predicted results, default is an empty string
+    predicted_results: ConnectorResponse | None = (
+        None  # The predicted results, default is None
+    )
 
     duration: float = 0.0  # The duration it took to get the results, default is 0.0

moonshot/src/connectors/connector_response.py

@@ -0,0 +1,20 @@
+class ConnectorResponse:
+    def __init__(self, response: str, context: list = []):
+        """
+        Initializes a ConnectorResponse instance.
+
+        Args:
+            response (str): The response text from the connector.
+            context (list, optional): Additional context or metadata related to the response. Defaults to an empty list.
+        """
+        self.response = response
+        self.context = context
+
+    def to_dict(self) -> dict:
+        """
+        Converts the ConnectorResponse instance to a dictionary.
+
+        Returns:
+            dict: A dictionary representation of the ConnectorResponse instance.
+        """
+        return {"response": self.response, "context": self.context}

moonshot/src/redteaming/attack/attack_module.py

@@ -31,6 +31,22 @@ class AttackModule:
         INSERT INTO {} (connection_id,context_strategy,prompt_template,attack_module,
         metric,prompt,prepared_prompt,system_prompt,predicted_result,duration,prompt_time)VALUES(?,?,?,?,?,?,?,?,?,?,?)
         """
+    sql_create_chat_history_table = """
+        CREATE TABLE IF NOT EXISTS {} (
+        id INTEGER PRIMARY KEY AUTOINCREMENT,
+        connection_id text NOT NULL,
+        context_strategy text,
+        prompt_template text,
+        attack_module text,
+        metric text,
+        prompt text NOT NULL,
+        prepared_prompt text NOT NULL,
+        system_prompt text,
+        predicted_result text NOT NULL,
+        duration text NOT NULL,
+        prompt_time text NOT NULL
+        );
+    """
 
     def __init__(self, am_id: str, am_arguments: AttackModuleArguments | None = None):
         self.id = am_id
@@ -261,6 +277,15 @@ class AttackModule:
             list: A list of consolidated responses from the specified LLM connector.
         """
         consolidated_responses = []
+
+        # perform a check to see if the target endpoint has its table in the db. if not, create one
+        endpoint_id = target_llm_connector.id.replace("-", "_")
+        if not Storage.check_database_table_exists(self.db_instance, endpoint_id):
+            Storage.create_database_table(
+                self.db_instance,
+                AttackModule.sql_create_chat_history_table.format(endpoint_id),
+            )
+
         for prepared_prompt in list_of_prompts:
             if self.cancel_event.is_set():
                 logger.warning(
@@ -316,6 +341,7 @@ class AttackModule:
             chat_record_tuple (tuple): A tuple containing the chat record information.
             chat_record_id (str): The ID of the chat record.
         """
+
         endpoint_id = chat_record_id.replace("-", "_")
         Storage.create_database_record(
             self.db_instance,
@@ -623,7 +649,8 @@ class RedTeamingPromptArguments(BaseModel):
 
         This method collects all the attributes of the RedTeamingPromptArguments instance and forms a tuple
         with the attribute values in this specific order: conn_id, cs_id, pt_id, am_id, me_id, original_prompt,
-        connector_prompt.prompt, connector_prompt.predicted_results, connector_prompt.duration, start_time.
+        connector_prompt.prompt, system_prompt, connector_prompt.predicted_results.response, 
+        connector_prompt.duration, start_time.
 
         Returns:
             tuple: A tuple representation of the RedTeamingPromptArguments instance.
@@ -637,21 +664,21 @@ class RedTeamingPromptArguments(BaseModel):
             self.original_prompt,
             self.connector_prompt.prompt,
             self.system_prompt,
-            str(self.connector_prompt.predicted_results),
+            self.connector_prompt.predicted_results.response if self.connector_prompt.predicted_results else "",
             str(self.connector_prompt.duration),
             self.start_time,
         )
 
     def to_dict(self) -> dict:
         """
-        Converts the RedTeamingPromptArguments instance into a dict.
+        Converts the RedTeamingPromptArguments instance into a dictionary.
 
-        This method collects all the attributes of the RedTeamingPromptArguments instance and forms a dict
-        with the keys: conn_id, cs_id, pt_id, am_id, me_id, original_prompt, prepared_prompt, system_prompt
+        This method collects all the attributes of the RedTeamingPromptArguments instance and forms a dictionary
+        with the keys: conn_id, cs_id, pt_id, am_id, me_id, original_prompt, system_prompt, prepared_prompt,
         response, duration, start_time.
 
         Returns:
-            dict: A dict representation of the RedTeamingPromptArguments instance.
+            dict: A dictionary representation of the RedTeamingPromptArguments instance.
         """
         return {
             "conn_id": self.conn_id,
@@ -662,7 +689,7 @@ class RedTeamingPromptArguments(BaseModel):
             "original_prompt": self.original_prompt,
             "prepared_prompt": self.connector_prompt.prompt,
             "system_prompt": self.system_prompt,
-            "response": str(self.connector_prompt.predicted_results),
+            "response": self.connector_prompt.predicted_results.response if self.connector_prompt.predicted_results else "",
             "duration": str(self.connector_prompt.duration),
             "start_time": self.start_time,
         }

moonshot/src/runs/run_arguments.py

@@ -1,6 +1,7 @@
 from __future__ import annotations
 
 import ast
+import json
 
 from pydantic import BaseModel
 
@@ -83,20 +84,7 @@ class RunArguments(BaseModel):
         Returns:
             tuple: A tuple containing the run arguments ready for database insertion.
         """
-        return (
-            self.runner_id,
-            self.runner_type.name.lower(),
-            str(self.runner_args),
-            str(self.endpoints),
-            self.results_file,
-            self.start_time,
-            self.end_time,
-            self.duration,
-            str(self.error_messages),
-            str(self.raw_results),
-            str(self.results),
-            self.status.name.lower(),
-        )
+        return self._to_tuple(include_run_id=False)
 
     def to_tuple(self) -> tuple:
         """
@@ -112,7 +100,42 @@ class RunArguments(BaseModel):
         Returns:
             tuple: A tuple containing the serialized attributes of the RunArguments instance.
         """
-        return (
+        return self._to_tuple(include_run_id=True)
+
+    def _to_tuple(self, include_run_id: bool) -> tuple:
+        """
+        Converts the RunArguments object to a tuple format.
+
+        This method serializes the RunArguments instance into a tuple, including or excluding the run_id based on the
+        include_run_id parameter. It ensures that all dictionary keys are converted to strings and JSON-encodes
+        the raw_results and results attributes.
+
+        Args:
+            include_run_id (bool): A flag indicating whether to include the run_id in the resulting tuple.
+
+        Returns:
+            tuple: A tuple containing the serialized attributes of the RunArguments instance.
+        """
+
+        def convert_keys_to_str(d):
+            """
+            Recursively converts all keys in a dictionary to strings.
+
+            Args:
+                d (dict): The dictionary whose keys need to be converted.
+
+            Returns:
+                dict: A new dictionary with all keys converted to strings.
+            """
+
+            def convert_value(v):
+                if isinstance(v, dict):
+                    return {str(k): convert_value(val) for k, val in v.items()}
+                return v
+
+            return {str(k): convert_value(v) for k, v in d.items()}
+
+        base_tuple = (
             self.runner_id,
             self.runner_type.name.lower(),
             str(self.runner_args),
@@ -122,11 +145,11 @@ class RunArguments(BaseModel):
             self.end_time,
             self.duration,
             str(self.error_messages),
-            str(self.raw_results),
-            str(self.results),
+            json.dumps(convert_keys_to_str(self.raw_results)),
+            json.dumps(self.results),
             self.status.name.lower(),
-            self.run_id,
         )
+        return base_tuple + (self.run_id,) if include_run_id else base_tuple
 
     @classmethod
     def from_tuple(cls, run_record: tuple) -> RunArguments:
@@ -156,7 +179,7 @@ class RunArguments(BaseModel):
             end_time=float(run_record[7]),
             duration=run_record[8],
             error_messages=ast.literal_eval(run_record[9]),
-            raw_results=ast.literal_eval(run_record[10]),
-            results=ast.literal_eval(run_record[11]),
+            raw_results=json.loads(run_record[10]),
+            results=json.loads(run_record[11]),
             status=RunStatus(run_record[12]),
         )