Fix model system prompt location (__model__["info"]["system"]) and user prompt priority (chat controls first)

Co-authored-by: owndev <69784886+owndev@users.noreply.github.com>

Author: Copilot

docs/google-gemini-integration.md

@@ -259,13 +259,13 @@ The Google Gemini pipeline supports a hierarchical system prompt configuration t
 
 1. **Default System Prompt** (`GOOGLE_DEFAULT_SYSTEM_PROMPT`): Global default applied to all chats, configurable via environment variable or Admin UI valves.
 
-2. **Model System Prompt**: The system prompt configured in the model settings (Admin > Models > Select Model > System Prompt). This is passed via `__metadata__["chat"]["params"]["system"]`.
+2. **Model System Prompt**: The system prompt configured in the model settings (Admin > Models > Select Model > System Prompt). This is accessed via `__model__["info"]["system"]`.
 
 3. **User System Prompt**: The user's personalized system prompt from either:
-   - **User Settings** (Settings > Personalization): Stored in `settings.ui.system`
    - **Chat Controls**: The system message passed with individual chat messages
+   - **User Settings** (Settings > Personalization): Stored in `settings.ui.system`
 
-   Note: User settings take precedence over chat controls if both are set.
+   Note: Chat controls take precedence over user settings if both are set.
 
 ### How It Works
 
@@ -310,7 +310,7 @@ Users can set their personalized system prompt in Open WebUI:
 2. Enter your preferred system prompt in the "System Prompt" field
 3. Save settings
 
-This prompt will be automatically applied to all your Gemini chats, combined with any default and model prompts.
+Alternatively, users can override the system prompt per-chat using chat controls. If both are set, the chat controls value takes precedence.
 
 ### Example
 
@@ -326,7 +326,7 @@ You are a helpful AI assistant.
 You specialize in Python programming.
 ```
 
-**User system prompt (Settings > Personalization OR chat controls):**
+**User system prompt (chat controls OR Settings > Personalization):**
 ```
 My name is John. I prefer detailed explanations.
 ```

pipelines/google/google_gemini.py

@@ -321,31 +321,29 @@ def _get_user_personalization_prompt(
         return None
 
     def _get_model_system_prompt(
-        self, __metadata__: Optional[dict] = None
+        self, __model__: Optional[dict] = None
     ) -> Optional[str]:
         """Get the model's system prompt from model settings.
 
         In Open WebUI, each model can have its own system prompt configured
-        in Admin > Models > Select Model > System Prompt. This is typically
-        passed via __metadata__["chat"]["params"]["system"].
+        in Admin > Models > Select Model > System Prompt. This is accessed
+        via __model__["info"]["system"].
 
         Args:
-            __metadata__: The metadata dict passed to the pipe method
+            __model__: The model dict passed to the pipe method
 
         Returns:
             The model's system prompt or None if not set
         """
-        if __metadata__ is None:
+        if __model__ is None:
             return None
 
         try:
-            chat = __metadata__.get("chat")
-            if chat and isinstance(chat, dict):
-                params = chat.get("params")
-                if params and isinstance(params, dict):
-                    system_prompt = params.get("system")
-                    if system_prompt and isinstance(system_prompt, str):
-                        return system_prompt.strip() or None
+            info = __model__.get("info")
+            if info and isinstance(info, dict):
+                system_prompt = info.get("system")
+                if system_prompt and isinstance(system_prompt, str):
+                    return system_prompt.strip() or None
         except Exception as e:
             self.log.debug(f"Could not retrieve model system prompt: {e}")
 
@@ -355,30 +353,30 @@ def _combine_system_prompts(
         self,
         chat_system_prompt: Optional[str],
         __user__: Optional[dict] = None,
-        __metadata__: Optional[dict] = None,
+        __model__: Optional[dict] = None,
     ) -> Optional[str]:
         """Combine default, model, and user system prompts.
 
         Prompt hierarchy (all prompts are combined if set):
         1. DEFAULT_SYSTEM_PROMPT (environment/valve setting)
-        2. Model system prompt (from model settings - __metadata__["chat"]["params"]["system"])
-        3. User system prompt (from user settings OR chat controls/messages)
+        2. Model system prompt (from model settings - __model__["info"]["system"])
+        3. User system prompt (from chat controls OR user settings - chat controls take precedence)
 
         Args:
             chat_system_prompt: The chat-level system prompt from messages (may be None)
             __user__: The user dict passed to the pipe method
-            __metadata__: The metadata dict passed to the pipe method
+            __model__: The model dict passed to the pipe method
 
         Returns:
             Combined system prompt or None if none are set
         """
         default_prompt = self.valves.DEFAULT_SYSTEM_PROMPT.strip() or None
-        model_prompt = self._get_model_system_prompt(__metadata__)
+        model_prompt = self._get_model_system_prompt(__model__)
         user_personalization = self._get_user_personalization_prompt(__user__)
         chat_prompt = chat_system_prompt.strip() if chat_system_prompt else None
 
-        # User prompt = user personalization OR chat prompt (user settings take precedence)
-        user_prompt = user_personalization or chat_prompt
+        # User prompt = chat controls OR user settings (chat controls take precedence if both are set)
+        user_prompt = chat_prompt or user_personalization
 
         prompts = [p for p in [default_prompt, model_prompt, user_prompt] if p]
 
@@ -485,7 +483,7 @@ async def _build_image_generation_contents(
         messages: List[Dict[str, Any]],
         __event_emitter__: Callable,
         __user__: Optional[dict] = None,
-        __metadata__: Optional[dict] = None,
+        __model__: Optional[dict] = None,
     ) -> Tuple[List[Dict[str, Any]], Optional[str]]:
         """Construct the contents payload for image-capable models.
 
@@ -499,7 +497,7 @@ async def _build_image_generation_contents(
 
         # Combine with default system prompt if configured
         system_instruction = self._combine_system_prompts(
-            user_system_instruction, __user__, __metadata__
+            user_system_instruction, __user__, __model__
         )
 
         last_user_msg = next(
@@ -962,15 +960,15 @@ def _prepare_content(
         self,
         messages: List[Dict[str, Any]],
         __user__: Optional[dict] = None,
-        __metadata__: Optional[dict] = None,
+        __model__: Optional[dict] = None,
     ) -> Tuple[List[Dict[str, Any]], Optional[str]]:
         """
         Prepare messages content for the API and extract system message if present.
 
         Args:
             messages: List of message objects from the request
             __user__: The user dict passed to the pipe method
-            __metadata__: The metadata dict passed to the pipe method
+            __model__: The model dict passed to the pipe method
 
         Returns:
             Tuple of (prepared content list, system message string or None)
@@ -983,7 +981,7 @@ def _prepare_content(
 
         # Combine with default system prompt if configured
         system_message = self._combine_system_prompts(
-            user_system_message, __user__, __metadata__
+            user_system_message, __user__, __model__
         )
 
         # Prepare contents for the API
@@ -2195,6 +2193,7 @@ async def pipe(
         __tools__: dict[str, Any] | None,
         __request__: Optional[Request] = None,
         __user__: Optional[dict] = None,
+        __model__: Optional[dict] = None,
     ) -> Union[str, AsyncIterator[str]]:
         """
         Main method for sending requests to the Google Gemini endpoint.
@@ -2206,6 +2205,7 @@ async def pipe(
             __tools__: Available tools
             __request__: FastAPI request object (for image upload)
             __user__: User information (for image upload)
+            __model__: Model information (for model-specific system prompt)
 
         Returns:
             Response from Google Gemini API, which could be a string or an iterator for streaming.
@@ -2243,7 +2243,7 @@ async def pipe(
                         contents,
                         system_instruction,
                     ) = await self._build_image_generation_contents(
-                        messages, __event_emitter__, __user__, __metadata__
+                        messages, __event_emitter__, __user__, __model__
                     )
                     # For image generation, system_instruction is integrated into the prompt
                     # so it will be None here (this is expected and correct)
@@ -2256,7 +2256,7 @@ async def pipe(
                 # For non-image generation models, use the full conversation history
                 # Prepare content and extract system message normally
                 contents, system_instruction = self._prepare_content(
-                    messages, __user__, __metadata__
+                    messages, __user__, __model__
                 )
                 if not contents:
                     return "Error: No valid message content found"