FEAT: [UI] add model feature in Launch Model list.

doc/source/models/index.rst

@@ -250,6 +250,7 @@ Model Usage
    model_abilities/index
    builtin/index
    custom
+   model_update
    sources/sources
    virtualenv
    lora

doc/source/models/model_update.rst

@@ -0,0 +1,165 @@
+.. _model_update:
+
+============
+Model Update
+============
+.. versionadded:: v1.13.0
+
+This section briefly introduces two common operations on the "Launch Model" page: updating model lists and adding models. They correspond to the "Type Selection + Update" and "Add Model" buttons at the top of the page, facilitating quick refresh of models of a certain type or adding new models locally.
+
+.. raw:: html
+
+    <img class="align-center" alt="model update interface" src="../_static/model_update.png" style="background-color: transparent", width="95%">
+
+Update Models (Launch Model Page)
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+- Operation Location: "Type Selection" dropdown and "Update" button at the top right of the page.
+- Usage:
+1. Select a model type from the dropdown (such as llm, embedding, rerank, image, audio, video).
+2. Click the "Update" button, the page will send an update request to the backend, then automatically jump to the corresponding Tab and refresh the model list of that type.
+
+Add Model (Launch Model Page)
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+- Operation Location: "Add Model" button at the top right of the page.
+- Usage:
+1. Click "Add Model" to open the add dialog.
+2. Complete the model addition process in the dialog.
+3. After successful addition, the page will jump to the corresponding type Tab to immediately view the latest model list.
+
+Xinference Models Hub User Guide
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Overview
+--------
+
+Xinference Models Hub is a full-stack platform for managing and sharing models.
+It provides a comprehensive solution for model registration, browsing, review workflows, and collaborative model management.
+
+You can visit the Models Hub at: https://model.xinference.io
+
+Quick Start
+-----------
+
+User Registration and Login
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+**Registration**
+
+1. Open the website registration page
+2. Fill in the necessary information and submit
+
+**Login**
+
+1. Open the website login page
+2. After successful login, you will be redirected to the model list page
+
+**Password Reset**
+
+1. Click the "Forgot Password" link on the login page
+2. Follow the instructions in the email to reset your password
+
+**Logout**
+
+1. Click the avatar in the top right corner of the page
+2. Select "Logout" from the dropdown menu
+
+Core Features
+-------------
+
+Browse Models
+^^^^^^^^^^^^^
+
+**Model List (Homepage)**
+
+* **Function:** Browse available models, click any model to view details
+* **Location:** "Models" menu in the website navigation bar
+
+.. note::
+   Some advanced models are only visible to authorized users.
+
+**Model Details and Documentation**
+
+* **Function:** View detailed information about models
+* **Default Display:** "README" tab - view model description, usage instructions, and notes
+* **Other Tabs:** Settings (authorized users), review status
+
+User Center
+^^^^^^^^^^^
+
+* **Function:** View and manage personal information
+* **Location:** Click the avatar in the top right corner, select "User Center"
+* **Content:** Personal profile settings
+
+Model Management (Authorized Users)
+-----------------------------------
+
+Model Registration
+^^^^^^^^^^^^^^^^^^
+
+* **Function:** Submit new models to the platform
+* **Location:** Click the avatar in the top right corner, select "Model Registration"
+* **Required Permissions:**
+
+  * **Private Models:** Model registration permission
+  * **Public Models:** Public model registration permission
+  * **Enterprise Models:** Enterprise model registration permission
+
+**Operation Process:**
+
+1. Fill in basic model information
+2. Fill in Readme (can be automatically obtained by clicking the Get Readme button)
+3. Submit (to register public models, enable the Public Model parameter)
+
+**Notes:**
+
+  * Regular users can only register private models
+  * Public model registration requires review, and can be used publicly after approval (no review needed if you have public model registration permission)
+  * Enterprise model registration requires enabling the Public Model parameter first
+
+My Models
+^^^^^^^^^
+
+* **Function:** View models associated with your account (models you registered)
+* **Location:** Click the avatar in the top right corner, select "My Models"
+* **Required Permissions:**
+
+  * **Private Models:** Model registration permission
+  * **Public Models:** Model registration permission
+  * **Enterprise Models:** Model registration permission
+
+Model Maintenance
+^^^^^^^^^^^^^^^^^
+
+* **Function:** Modify and manage existing models
+* **Location:** Click the "Settings" icon on the model details page
+
+* **Permission Requirements:**
+
+  * **Private Models:** Model ownership or any public model management permission
+  * **Advanced Models:** Advanced model update, delete, or expiration permission
+  * **Public Models:** Public model update, delete, or expiration permission
+
+**Notes:**
+
+  * Updating JSON or modifying expiration attributes of public models will automatically create a PR to the xorbitsai/inference repository
+
+Review Workflow
+^^^^^^^^^^^^^^^
+
+**For Model Submitters:**
+
+1. Submit models for review
+2. Check review status on the model details page
+3. Make modifications based on reviewer feedback if needed
+
+**For Reviewers:**
+
+* **Required Permissions:** Model review list permission, model review permission
+
+**Operation Process:**
+
+1. Enter the review queue page
+2. Evaluate model quality and compliance
+3. Approve or reject and provide feedback

xinference/api/restful_api.py

@@ -198,6 +198,14 @@ class RegisterModelRequest(BaseModel):
     persist: bool
 
 
+class AddModelRequest(BaseModel):
+    model_name: str
+
+
+class UpdateModelRequest(BaseModel):
+    model_type: str
+
+
 class BuildGradioInterfaceRequest(BaseModel):
     model_type: str
     model_name: str
@@ -900,6 +908,26 @@ async def internal_exception_handler(request: Request, exc: Exception):
                 else None
             ),
         )
+        self._router.add_api_route(
+            "/v1/models/add",
+            self.add_model,
+            methods=["POST"],
+            dependencies=(
+                [Security(self._auth_service, scopes=["models:add"])]
+                if self.is_authenticated()
+                else None
+            ),
+        )
+        self._router.add_api_route(
+            "/v1/models/update_type",
+            self.update_model_type,
+            methods=["POST"],
+            dependencies=(
+                [Security(self._auth_service, scopes=["models:add"])]
+                if self.is_authenticated()
+                else None
+            ),
+        )
         self._router.add_api_route(
             "/v1/cache/models",
             self.list_cached_models,
@@ -3123,25 +3151,119 @@ async def unregister_model(self, model_type: str, model_name: str) -> JSONRespon
             raise HTTPException(status_code=500, detail=str(e))
         return JSONResponse(content=None)
 
+    async def add_model(self, request: Request) -> JSONResponse:
+        try:
+            # Parse request
+            raw_json = await request.json()
+            body = AddModelRequest.parse_obj(raw_json)
+            model_name = body.model_name
+
+            supervisor_ref = await self._get_supervisor_ref()
+
+            # Call supervisor with model_name only
+            await supervisor_ref.add_model(model_name)
+
+            # Get model type information to return to frontend
+            import requests
+
+            try:
+                info_url = f"https://model.xinference.io/api/models/{model_name}"
+                info_response = requests.get(info_url, timeout=30)
+                if info_response.status_code == 200:
+                    model_info = info_response.json()
+                    model_type = model_info.get("data", {}).get("model_type")
+                    model_data = model_info.get("data", {}).get("model_data", {})
+                else:
+                    model_type = "unknown"
+                    model_data = {}
+            except Exception:
+                model_type = "unknown"
+                model_data = {}
+
+        except ValueError as re:
+            logger.error(f"ValueError in add_model API: {re}", exc_info=True)
+            logger.error(f"ValueError details: {type(re).__name__}: {re}")
+            raise HTTPException(status_code=400, detail=str(re))
+        except Exception as e:
+            logger.error(f"Unexpected error in add_model API: {e}", exc_info=True)
+            logger.error(f"Error details: {type(e).__name__}: {e}")
+            import traceback
+
+            logger.error(f"Full traceback: {traceback.format_exc()}")
+            raise HTTPException(status_code=500, detail=str(e))
+
+        response_data = {
+            "success": True,
+            "message": f"Model added successfully: {model_name}",
+            "data": {
+                "model_name": model_name,
+                "model_type": model_type,
+                "model_version": model_data.get("version"),
+                "model_ability": model_data.get("model_ability", []),
+                "model_family": model_data.get("model_family"),
+            },
+        }
+
+        return JSONResponse(content=response_data)
+
+    async def update_model_type(self, request: Request) -> JSONResponse:
+        try:
+            # Parse request
+            raw_json = await request.json()
+
+            body = UpdateModelRequest.parse_obj(raw_json)
+            model_type = body.model_type
+
+            # Get supervisor reference
+            supervisor_ref = await self._get_supervisor_ref()
+
+            await supervisor_ref.update_model_type(model_type)
+
+        except ValueError as re:
+            logger.error(f"ValueError in update_model_type API: {re}", exc_info=True)
+            raise HTTPException(status_code=400, detail=str(re))
+        except Exception as e:
+            logger.error(
+                f"Unexpected error in update_model_type API: {e}", exc_info=True
+            )
+            raise HTTPException(status_code=500, detail=str(e))
+
+        return JSONResponse(
+            content={
+                "message": f"Model configurations updated successfully for type: {model_type}"
+            }
+        )
+
     async def list_model_registrations(
         self, model_type: str, detailed: bool = Query(False)
     ) -> JSONResponse:
         try:
+
             data = await (await self._get_supervisor_ref()).list_model_registrations(
                 model_type, detailed=detailed
             )
+
             # Remove duplicate model names.
             model_names = set()
             final_data = []
             for item in data:
                 if item["model_name"] not in model_names:
                     model_names.add(item["model_name"])
                     final_data.append(item)
+
             return JSONResponse(content=final_data)
         except ValueError as re:
+            logger.error(
+                f"ValueError in list_model_registrations: {re}",
+                exc_info=True,
+            )
             logger.error(re, exc_info=True)
             raise HTTPException(status_code=400, detail=str(re))
         except Exception as e:
+            logger.error(
+                f"Unexpected error in list_model_registrations: {e}",
+                exc_info=True,
+            )
             logger.error(e, exc_info=True)
             raise HTTPException(status_code=500, detail=str(e))