adds support for o1 (non-preview) model for openai_client, improves h…

…andling of context for codespace-assistant (#308)

NOTE: suggested that you re-run root level `make` if merging this update
into your environment where you already have your packages installed, in
order to cause the updated `openai` and `tiktoken` packages to be
loaded.

assistants/codespace-assistant/assistant/config.py

@@ -126,7 +126,7 @@ class AssistantConfigModel(BaseModel):
         AzureOpenAIClientConfigModel | OpenAIClientConfigModel,
         Field(
             title="OpenAI Reasoning Model Configuration",
-            description="Configuration for the reasoning model, such as o1-preview, o1-mini, etc.",
+            description="Configuration for the reasoning model, such as o1, o1-preview, o1-mini, etc.",
             discriminator="ai_service_type",
             default=AzureOpenAIClientConfigModel.model_construct(),
         ),

assistants/codespace-assistant/assistant/extensions/tools/__mcp_server_configs.py

@@ -21,32 +21,34 @@
         prompt=dedent("""
             Follow these steps for each interaction:
 
-            1. User Identification:
-            - You should assume that you are interacting with default_user
-            - If you have not identified default_user, proactively try to do so.
-
-            2. Memory Retrieval:
-            - Always begin your chat by saying only "Remembering..." and retrieve all relevant information from your knowledge graph
+            1. Memory Retrieval:
+            - Always begin your chat by saying only "Remembering..." and retrieve all relevant information
+              from your knowledge graph
             - Always refer to your knowledge graph as your "memory"
 
-            3. Memory
+            2. Memory
             - While conversing with the user, be attentive to any new information that falls into these categories:
                 a) Basic Identity (age, gender, location, job title, education level, etc.)
                 b) Behaviors (interests, habits, etc.)
                 c) Preferences (communication style, preferred language, etc.)
                 d) Goals (goals, targets, aspirations, etc.)
                 e) Relationships (personal and professional relationships up to 3 degrees of separation)
 
-            4. Memory Update:
+            3. Memory Update:
             - If any new information was gathered during the interaction, update your memory as follows:
                 a) Create entities for recurring organizations, people, and significant events
                 b) Connect them to the current entities using relations
                 b) Store facts about them as observations
         """),
     ),
-    MCPServerConfig(
-        name="Web Research MCP Server",
-        command="npx",
-        args=["-y", "@mzxrai/mcp-webresearch@latest"],
-    ),
+    # MCPServerConfig(
+    #     name="Sequential Thinking MCP Server",
+    #     command="npx",
+    #     args=["-y", "@modelcontextprotocol/server-sequential-thinking"],
+    # ),
+    # MCPServerConfig(
+    #     name="Web Research MCP Server",
+    #     command="npx",
+    #     args=["-y", "@mzxrai/mcp-webresearch@latest"],
+    # ),
 ]

assistants/codespace-assistant/assistant/response/completion_handler.py

@@ -41,7 +41,7 @@ async def handle_completion(
     context: ConversationContext,
     config: AssistantConfigModel,
     silence_token: str,
-    method_metadata_key: str,
+    metadata_key: str,
     response_start_time: float,
 ) -> StepResult:
     # helper function for handling errors
@@ -90,7 +90,7 @@ async def handle_error(error_message: str) -> StepResult:
         step_result.metadata,
         {
             "debug": {
-                method_metadata_key: {
+                metadata_key: {
                     "response": completion.model_dump() if completion else "[no response from openai]",
                 },
             },
@@ -163,7 +163,7 @@ async def handle_error(error_message: str) -> StepResult:
                     mcp_sessions,
                     tool_call,
                     mcp_tools,
-                    f"{method_metadata_key}:request:tool_call_{tool_call_count}",
+                    f"{metadata_key}:request:tool_call_{tool_call_count}",
                 )
             except Exception as e:
                 logger.exception(f"Error handling tool call: {e}")

assistants/codespace-assistant/assistant/response/response.py

@@ -54,6 +54,7 @@ async def respond_to_conversation(
 
         # Initialize a loop control variable
         max_steps = config.extensions_config.tools.max_steps
+        encountered_error = False
         completed_within_max_steps = False
         step_count = 0
 
@@ -69,14 +70,19 @@ async def respond_to_conversation(
                 context=context,
                 config=config,
                 metadata=metadata,
+                metadata_key=f"respond_to_conversation:step_{step_count}",
             )
 
+            if step_result.status == "error":
+                encountered_error = True
+                break
+
             if step_result.status == "final":
                 completed_within_max_steps = True
                 break
 
         # If the conversation did not complete within the maximum number of steps, send a message to the user
-        if not completed_within_max_steps:
+        if not completed_within_max_steps and not encountered_error:
             await context.send_messages(
                 NewConversationMessage(
                     content=config.extensions_config.tools.max_steps_truncation_message,

assistants/codespace-assistant/assistant/response/step_handler.py

@@ -36,11 +36,23 @@ async def next_step(
     context: ConversationContext,
     config: AssistantConfigModel,
     metadata: dict[str, Any],
+    metadata_key: str,
 ) -> StepResult:
     step_result = StepResult(status="continue", metadata=metadata.copy())
 
     # helper function for handling errors
-    async def handle_error(error_message: str) -> StepResult:
+    async def handle_error(error_message: str, error_debug: dict[str, Any] | None = None) -> StepResult:
+        if error_debug is not None:
+            deepmerge.always_merger.merge(
+                step_result.metadata,
+                {
+                    "debug": {
+                        metadata_key: {
+                            "error": error_debug,
+                        },
+                    },
+                },
+            )
         await context.send_messages(
             NewConversationMessage(
                 content=error_message,
@@ -63,9 +75,6 @@ async def handle_error(error_message: str) -> StepResult:
     # )
     # reasoning_request_config = config.reasoning_ai_client_config.request_config
 
-    # Define the metadata key for any metadata created within this method
-    method_metadata_key = "respond_to_conversation"
-
     # Track the start time of the response generation
     response_start_time = time.time()
 
@@ -92,7 +101,7 @@ async def handle_error(error_message: str) -> StepResult:
         step_result.metadata,
         {
             "debug": {
-                method_metadata_key: {
+                metadata_key: {
                     "request": {
                         "model": generative_request_config.model,
                         "messages": chat_message_params,
@@ -111,12 +120,22 @@ async def handle_error(error_message: str) -> StepResult:
 
         except Exception as e:
             logger.exception(f"exception occurred calling openai chat completion: {e}")
+            deepmerge.always_merger.merge(
+                step_result.metadata,
+                {
+                    "debug": {
+                        metadata_key: {
+                            "error": str(e),
+                        },
+                    },
+                },
+            )
             await context.send_messages(
                 NewConversationMessage(
                     content="An error occurred while calling the OpenAI API. Is it configured correctly?"
                     " View the debug inspector for more information.",
                     message_type=MessageType.notice,
-                    metadata={method_metadata_key: {"error": str(e)}},
+                    metadata=step_result.metadata,
                 )
             )
             step_result.status = "error"
@@ -133,7 +152,7 @@ async def handle_error(error_message: str) -> StepResult:
         context,
         config,
         silence_token,
-        method_metadata_key,
+        metadata_key,
         response_start_time,
     )
 

assistants/codespace-assistant/assistant/response/utils/openai_utils.py

@@ -128,6 +128,7 @@ def convert_mcp_tools_to_openai_tools(mcp_tools: List[Tool] | None) -> List[Chat
                             "properties": {
                                 "aiContext": aiContext,
                             },
+                            "required": ["aiContext"],
                         },
                     ),
                 ),

assistants/codespace-assistant/pyproject.toml

@@ -15,8 +15,9 @@ dependencies = [
     "html2docx>=1.6.0",
     "markdown>=3.6",
     "mcp>=1.2.0",
-    "openai>=1.3.9",
+    "openai>=1.60.2",
     "openai-client>=0.1.0",
+    "tiktoken>=0.8.0",
 ]
 
 [tool.hatch.build.targets.wheel]