stuff

CLAUDE.md

@@ -3,10 +3,28 @@
 This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
 
 ## Project Context
-This is a NiceGUI-based web frontend for managing an Arch Linux system with AMD GPU and Ollama. The application serves as:
-1. System resource monitor (CPU, GPU, memory usage)
-2. Ollama model manager (list, download, delete models via Ollama API on port 11434)
-3. Platform for additional system tools
+This is a NiceGUI-based web platform for testing and managing AI models through Ollama on Arch Linux systems with GPU support. The application serves as an AI model testing environment featuring:
+
+### Core Purpose
+A streamlined interface for testing AI models locally, managing Ollama models, and running various AI-related testing tools.
+
+### Main Features:
+1. **Model Manager** - Complete Ollama model management interface
+   - Download, delete, create, and test models
+   - Support for Hugging Face models via Ollama pull syntax
+   - Rich model metadata display
+   - Quick in-app chat testing
+
+2. **System Monitoring** - Resource tracking for AI workloads
+   - Real-time GPU monitoring (AMD/NVIDIA) to track model performance
+   - CPU and memory usage during model inference
+   - System metrics dashboard
+
+3. **AI Testing Tools**:
+   - **Censor** - Text content filtering/censoring tool for testing AI outputs
+   - Additional testing tools to be added as needed
+
+4. **Settings** - Application configuration and refresh intervals
 
 ## Development Commands
 
@@ -41,36 +59,212 @@ uv sync
 - **UI Framework**: NiceGUI (async web framework based on FastAPI/Vue.js)
 - **Python Version**: 3.13+
 - **Ollama API**: Running on localhost:11434
+- **Dependencies**:
+  - `nicegui` - Main UI framework
+  - `niceguiasyncelement` - Custom async component framework (from git)
+  - `psutil` - System monitoring
+  - `httpx` - Async HTTP client for Ollama API
+  - `python-dotenv` - Environment configuration
 
 ### Project Structure
-- `src/main.py`: Entry point, configures NiceGUI app with environment variables
-- `src/pages/`: Page components inheriting from NiceGUI elements
-- `src/components/`: Reusable UI components (currently empty, ready for implementation)
-- `src/static/`: Static assets (CSS, images)
-- `.env`: Configuration variables (APP_TITLE, APP_PORT, APP_STORAGE_SECRET, APP_SHOW)
+```
+src/
+├── main.py              # Entry point, NiceGUI app configuration with all routes
+├── pages/               # Page components (inheriting NiceGUI elements)
+│   ├── dashboard.py     # Main dashboard with system metrics
+│   ├── ollama_manager.py # Ollama model management interface (AsyncColumn)
+│   ├── system_overview.py # System information page
+│   └── welcome.py       # Welcome/landing page
+├── components/          # Reusable UI components
+│   ├── circular_progress.py # Circular progress indicators
+│   ├── header.py        # App header with live status
+│   ├── sidebar.py       # Navigation sidebar with updated menu structure
+│   ├── bottom_nav.py    # Mobile bottom navigation
+│   ├── ollama_downloader.py # Ollama model downloader component (AsyncCard)
+│   ├── ollama_model_creation.py # Model creation component (AsyncCard)
+│   └── ollama_quick_test.py # Model testing component (AsyncCard)
+├── utils/               # Utility modules
+│   ├── gpu_monitor.py   # GPU monitoring (AMD/NVIDIA auto-detect)
+│   ├── system_monitor.py # System resource monitoring
+│   ├── ollama_monitor.py # Ollama status monitoring (bindable dataclass)
+│   └── ollama.py        # Ollama API client functions
+└── static/              # Static assets (CSS, images)
+    └── style.css        # Custom dark theme styles
+```
 
 ### Key Design Patterns
-1. **Page Classes**: Pages are implemented as classes inheriting from NiceGUI elements (e.g., `WelcomePage` extends `ui.column`)
-2. **Environment Configuration**: All app settings are managed via `.env` file and loaded with python-dotenv
-3. **Async Support**: Main page handlers use async functions for potential async operations
+1. **Async Components**: Uses custom `niceguiasyncelement` framework for async page/component construction
+   - `AsyncColumn`, `AsyncCard` base classes for complex components
+   - `OllamaManagerPage(AsyncColumn)` for full page async initialization
+   - Async component dialogs with `await component.create()` pattern
+2. **Bindable Dataclasses**: Monitor classes use `@binding.bindable_dataclass` for reactive data binding
+   - `SystemMonitor`, `GPUMonitor`, `OllamaMonitor` for real-time data updates
+3. **Environment Configuration**: All app settings are managed via `.env` file and loaded with python-dotenv
+4. **Centralized Routing**: All routes defined in main.py with layout creation pattern
+5. **Real-time Updates**: Timer-based updates every 2 seconds for all monitor instances
+
+## Component Architecture
+
+### Monitor Classes (Supporting AI Testing)
+- **SystemMonitor**: Tracks system resources during AI model testing
+  - CPU usage during model inference
+  - Memory consumption by loaded models
+  - Disk I/O for model loading
+  - Process statistics for Ollama and GPU processes
+
+- **GPUMonitor**: Critical for AI workload monitoring
+  - Auto-detects AMD/NVIDIA GPUs
+  - Tracks GPU usage during model inference
+  - Memory usage by loaded models
+  - Temperature monitoring during extended testing
+  - Power draw under AI workloads
+
+- **OllamaMonitor**: Core service monitoring
+  - Ollama service status and version
+  - Currently loaded/active models
+  - Real-time model state tracking
+
+### UI Components
+- **MetricCircle**: Small circular progress indicator with icon
+- **LargeMetricCircle**: Large circular progress for primary metrics
+- **ColorfulMetricCard**: Action cards with gradient backgrounds
+- **Sidebar**: Navigation menu with updated structure:
+  - Main: Dashboard, System Overview
+  - Tools: Censor (content filtering)
+  - Bottom: Model Manager, Settings
+- **Header**: Top bar with system status indicators
+
+### Ollama-Specific Components (AsyncCard-based):
+- **OllamaDownloaderComponent**: Model downloading with progress tracking (supports HF models via Ollama's pull syntax)
+- **OllamaModelCreationComponent**: Custom model creation from Modelfile
+- **ModelQuickTestComponent**: Interactive model testing interface
 
 ## Ollama Integration
-The Ollama API is available at `http://localhost:11434/api/`. Key endpoints:
+The Ollama API client (`src/utils/ollama.py`) provides async functions:
+- `status()`: Check if Ollama is online and get version
+- `available_models()`: List installed models with detailed metadata
+- `active_models()`: Get currently loaded/running models
+- `delete_model()`: Remove a model
+- `model_info()`: Get detailed model information and Modelfile
+- `stream_chat()`: Stream chat responses
+
+### AI Model Testing Features:
+- **Model Discovery & Management**:
+  - Browse and pull models from Ollama library
+  - Support for HuggingFace models via Ollama syntax
+  - Rich metadata display (size, quantization, parameters, format)
+  - Time tracking for model versions
+
+- **Testing Capabilities**:
+  - Quick chat interface for immediate model testing
+  - Model information and Modelfile inspection
+  - Custom model creation from Modelfiles
+  - Real-time resource monitoring during inference
+
+- **Testing Tools**:
+  - Censor tool for output filtering analysis
+  - Extensible framework for adding new testing tools
+
+API endpoints at `http://localhost:11434/api/`:
 - `/api/version`: Get Ollama version
 - `/api/tags`: List available models
 - `/api/pull`: Download models
 - `/api/delete`: Remove models
 - `/api/generate`: Generate text
 - `/api/chat`: Chat completion
+- `/api/ps`: List running models
+- `/api/show`: Show model details
 
 ## System Monitoring
-For AMD GPU monitoring on Arch Linux:
-- Use `rocm-smi` for GPU stats (temperature, usage, memory)
-- Use `psutil` for CPU and memory monitoring
-- Consider `py3nvml` or direct sysfs reading for additional GPU metrics
+
+### GPU Monitoring Strategy
+The application uses a hierarchical approach for GPU monitoring:
+
+1. **NVIDIA GPUs** (via `nvidia-smi`):
+   - Temperature, usage, memory, power draw
+   - CUDA version and driver info
+   - Multi-GPU support
+
+2. **AMD GPUs** (multiple fallbacks):
+   - Primary: `rocm-smi` for full metrics
+   - Fallback: `/sys/class/drm` filesystem
+   - Reads hwmon for temperature data
+   - Supports both server and consumer GPUs
+
+### CPU & System Monitoring
+- Real-time CPU usage and per-core statistics
+- Memory (RAM and swap) usage
+- Disk usage and I/O statistics
+- Network traffic monitoring
+- Process tracking with top processes by CPU/memory
+- System uptime and kernel information
+
+## UI/UX Features
+
+### Dark Theme
+Custom dark theme with:
+- Background: `#1a1d2e` (main), `#252837` (sidebar)
+- Card backgrounds: `rgba(26, 29, 46, 0.7)` with backdrop blur
+- Accent colors: Cyan (`#06b6d4`) for primary actions
+- Metric colors: Purple (CPU), Green (Memory), Orange (GPU), Cyan (Temp)
+
+### Responsive Design
+- Desktop: Full sidebar navigation
+- Mobile: Bottom navigation bar
+- Adaptive grid layouts for different screen sizes
+- Viewport-aware content scaling
+
+### Real-time Updates
+- System metrics update every 2 seconds (configurable via `MONITORING_UPDATE_INTERVAL`)
+- Live data binding for all metrics
+- Smooth transitions and animations
 
 ## NiceGUI Patterns
-- Use `ui.dark_mode()` for theme toggling
-- Implement pages as classes extending NiceGUI elements for better organization
-- Use `ui.timer()` for periodic updates (system stats)
-- Leverage `ui.refreshable` decorator for dynamic content updates
+- **Data Binding**: Use `bind_text_from()` and `bind_value_from()` for reactive updates
+- **Page Routing**: Navigation via `ui.navigate.to(route)` with centralized route handling
+- **Async Components**: Custom `niceguiasyncelement` framework for complex async initialization
+  - `AsyncColumn.create()` for async page construction
+  - `AsyncCard.create()` for dialog components
+  - `@ui.refreshable` decorators for dynamic content updates
+- **Timer Updates**: `app.timer()` for periodic data refresh (2-second intervals)
+- **Dialog Patterns**: Modal dialogs with `await dialog` for user interactions
+- **Component Layout**: `create_layout(route)` pattern for consistent page structure
+- **Dark Mode**: Forced dark mode with custom CSS overrides
+
+## Environment Variables
+Configured in `.env`:
+- `MONITORING_UPDATE_INTERVAL`: Update frequency in seconds (default: 2)
+- `APP_PORT`: Web server port (default: 8080, use 8081 for testing)
+- `APP_TITLE`: Application title
+- `APP_STORAGE_SECRET`: Session storage encryption key
+- `APP_SHOW`: Auto-open browser on startup
+
+## Testing & Development
+- Run on port 8081 to avoid conflicts: `APP_PORT=8081 uv run python src/main.py`
+- Monitor GPU detection in console logs
+- Check Ollama connectivity at startup
+- Use browser DevTools for WebSocket debugging
+
+## Current Route Structure
+From main.py routing:
+- `/` - Dashboard (system metrics for monitoring AI workloads)
+- `/system` - System Overview (detailed resource information)
+- `/ollama` - Model Manager (primary interface for AI model testing)
+- `/censor` - Censor tool (AI output filtering/testing)
+- `/settings` - Settings (refresh intervals, app configuration)
+
+### Placeholder Routes (may be repurposed for AI tools):
+- `/processes` - Reserved for future AI tools
+- `/network` - Reserved for future AI tools
+- `/packages` - Reserved for future AI tools
+- `/logs` - Reserved for future AI tools
+- `/info` - Reserved for future AI tools
+
+## Future Enhancements
+- Enhanced model chat interface with conversation history
+- Model performance benchmarking tools
+- Batch testing capabilities for multiple models
+- Output comparison tools between different models
+- Integration with more AI model formats
+- Advanced prompt testing and optimization tools
+- Model fine-tuning interface

src/components/circular_progress.py

@@ -1,4 +1,29 @@
 from nicegui import ui
+from utils import SystemMonitor, GPUMonitor
+from typing import Optional, Literal
+
+
+class MetricCircleAdv:
+    def __init__(self, label: str, monitor: SystemMonitor | GPUMonitor,
+                 target_value: str,
+                 target_max_value: str,
+                 color: str,
+                 formatting: Literal['percent', 'units', 'degree'],
+                 icon: Optional[str] = None):
+        with ui.card().classes('metric-card p-4 text-center'):
+            with ui.column().classes('items-center gap-2'):
+                # Icon at top
+                with ui.row().classes('items-center gap-1'):
+                    if icon:
+                        ui.icon(icon, size='sm', color=color)
+
+                    # Title
+                    ui.label(label).classes('text-sm text-grey-5 font-medium')
+
+                # Circular progress - simplified
+                with ui.circular_progress(size='60px', color=color, show_value=False).bind_value_from(monitor, target_value):
+                    if formatting == 'percent':
+                        ui.label().classes('text-lg font-bold text-white').bind_text_from(monitor, target_value, backward=lambda x: f"{int(x * 100)} %")
 
 
 class MetricCircle:

src/components/header.py

@@ -1,17 +1,23 @@
-from nicegui import ui
-from utils import SystemMonitor, GPUMonitor
+from nicegui import ui, binding
+from utils import SystemMonitor, GPUMonitor, OllamaMonitor
 
 
 class Header(ui.header):
-    def __init__(self, system_monitor: SystemMonitor, gpu_monitor: GPUMonitor):
+    def __init__(self, system_monitor: SystemMonitor, gpu_monitor: GPUMonitor, ollama_monitor: OllamaMonitor):
         super().__init__(fixed=True, elevated=False)
 
         with self.classes('bg-transparent'):
-            with ui.row().classes('w-full items-center justify-between px-6 py-3'):
+            with ui.row().classes('w-full items-center justify-between px-6'):
                 # Left side - minimal branding
                 with ui.row().classes('items-center gap-3'):
                     ui.label('ArchGPU Frontend').classes('text-xl font-bold text-white')
-                    ui.chip('Live', icon='circle', color='green').props('size=sm outline')
+                    chip = ui.chip('Live', icon='circle').props('size=sm outline')
+                    chip.bind_text_from(ollama_monitor, 'status', backward=lambda x: self.update_ollama_running_chip(chip, x))
+
+                with ui.row().classes('items-center gap-4'):
+                    print(ollama_monitor.active_models)
+
+                    ui.label().bind_text_from(ollama_monitor, 'active_models', backward=lambda x: self.update_active_models(x))
 
                 # Right side - system status only
                 with ui.row().classes('items-center gap-4'):
@@ -31,3 +37,14 @@ class Header(ui.header):
                         ui.icon('thermostat', size='sm', color='red')
                         ui.label().classes('text-sm text-white').bind_text_from(gpu_monitor, 'temperature',
                                                                                 lambda x: f'{x:.1f}°C')
+
+    def update_ollama_running_chip(self, obj: ui.chip, state):
+        obj.classes(remove='text-red' if state else 'text-green')
+        obj.classes(add='text-green' if state else 'text-red')
+        return 'Ollama Running' if state else 'Ollama stopped'
+
+    def update_active_models(self, active_models):
+        used_vram = 0
+        for active_model in active_models:
+            used_vram += active_model['size_vram']
+        return f'{len(active_models)} Active Models using {(used_vram / 1024**3):.2f} GB'

src/components/ollama_downloader.py

@@ -0,0 +1,57 @@
+from nicegui import ui
+from niceguiasyncelement import AsyncCard
+from pathlib import Path
+from utils import ollama
+
+
+class OllamaDownloaderComponent(AsyncCard):
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self.is_downloading = False
+        self.download_progress = 0
+        self.download_status = ''
+
+    async def build(self) -> None:
+        with self:
+            with ui.column().classes('w-full gap-4'):
+                ui.label('Model Downloader').classes('text-xl font-bold')
+
+                model_input = ui.input(
+                    'Model ID',
+                    placeholder='e.g., TheBloke/Llama-2-7B-GGUF',
+                    value='qwen2.5:0.5b'
+                ).props('outlined dense').classes('w-full')
+                ui.link('Ollama Library', target='https://ollama.com/library/', new_tab=True)
+                ui.link('Using HF Models', target='https://huggingface.co/docs/hub/en/ollama', new_tab=True)
+
+                with ui.row().classes('items-center gap-2'):
+                    ui.icon('check_circle').props(f'color=positive').bind_visibility_from(self, 'download_status', backward=lambda x: True if x == 'success' else False)
+                    self.status_label = ui.label().bind_text_from(self, 'download_status')
+                ui.linear_progress(value=0, show_value=False).props('buffer=0.0 animation-speed=0').bind_value_from(self, 'download_progress')
+
+                self.download_btn = ui.button(
+                    'Download Model',
+                    on_click=lambda m=model_input: self.download_model(m.value)  # type: ignore
+                ).props('color=primary').classes('w-full').bind_enabled_from(self, 'model_id', backward=lambda x: bool(x) and not self.is_downloading)
+
+    async def download_model(self, model):
+        self.download_btn.set_enabled(False)
+        try:
+            async for chunk in ollama.download_model(model):
+                if chunk.strip():
+                    # Parse the JSON chunk and extract content
+                    import json
+                    try:
+                        chunk_data = json.loads(chunk)
+                        self.download_status = chunk_data['status']
+                        if 'total' in chunk_data and 'completed' in chunk_data:
+                            self.download_progress = chunk_data['completed'] / chunk_data['total']
+                            print(self.download_progress)
+                        else:
+                            self.download_progress = 0
+                    except json.JSONDecodeError:
+                        pass  # Skip malformed chunks
+        except Exception as e:
+            ui.notify(f'Error: {str(e)}', type='negative')
+        finally:
+            self.download_btn.set_enabled(True)

src/components/ollama_model_creation.py

@@ -0,0 +1,85 @@
+from nicegui import ui, binding
+from niceguiasyncelement import AsyncCard
+from pathlib import Path
+from utils import ollama
+from typing import Optional, Dict
+
+modelfile_example = """FROM qwen2.5-coder:7b
+PARAMETER num_ctx 8192
+PARAMETER temperature 0.1
+SYSTEM "Du bist ein Python-Experte."
+"""
+
+
+class OllamaModelCreationComponent(AsyncCard):
+    model_name = binding.BindableProperty()
+    model_from = binding.BindableProperty()
+    parameters = binding.BindableProperty()
+    quantize = binding.BindableProperty()
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self.is_downloading = False
+        self.download_progress = 0
+        self.download_status = ''
+
+    async def build(self) -> None:
+        self.classes('w-full')
+        with self:
+            with ui.column().classes('w-full gap-4'):
+                ui.label('Create Model').classes('text-xl font-bold')
+
+                ui.input('Model Name', value='qwen2.5-coder-32k-python:latest').props('outlined dense').classes('w-full').bind_value(self, 'model_name')
+                ui.input('From', value='qwen2.5-coder:7b').props('outlined dense').classes('w-full').bind_value(self, 'model_from')
+                ui.textarea(placeholder='Parameters').classes('w-full').props('autogrow').bind_value(self, 'parameters')
+
+                ui.select(['q4_K_M', 'q4_K_S', 'q8_0'], label='quantize', clearable=True).props('outlined dense').classes('w-full').bind_value(self, 'quantize')
+
+                with ui.row().classes('items-center gap-2'):
+                    ui.icon('check_circle').props(f'color=positive').bind_visibility_from(self, 'download_status', backward=lambda x: True if x == 'success' else False)
+                    self.status_label = ui.label().bind_text_from(self, 'download_status')
+                ui.linear_progress(value=0, show_value=False).props('buffer=0.0 animation-speed=0').bind_value_from(self, 'download_progress')
+                self.create_btn = ui.button('Create Model', on_click=self.create_model).props('color=primary').classes('w-full').bind_enabled_from(self, 'model_id', backward=lambda x: bool(x) and not self.is_downloading)
+
+    async def create_model(self):
+        self.parameters = self.parameters.strip()
+        model_parameters: Optional[Dict[str, str | int | float]] = None
+        if self.parameters:
+            model_parameters = {}
+            for line in self.parameters.split('\n'):
+                line = line.strip()
+                try:
+                    key, value = line.split(' ')
+                except:
+                    ui.notify(f'Not a valid format. {line}')
+                    return
+                if key in ['num_ctx', 'repeat_last_n', 'seed', 'num_predict', 'top_k']:
+                    model_parameters[key] = int(value)
+                elif key in ['repeat_penalty', 'temperature', 'top_p', 'min_p']:
+                    model_parameters[key] = float(value)
+                elif key == 'stop':
+                    model_parameters[key] = value.strip()
+                else:
+                    ui.notify(f'Unknown parameter: {key}')
+                    return
+
+        self.create_btn.set_enabled(False)
+        try:
+            async for chunk in ollama.create_ollama_model(self.model_name, self.model_from, model_parameters, self.quantize):
+                if chunk.strip():
+                    # Parse the JSON chunk and extract content
+                    import json
+                    try:
+                        chunk_data = json.loads(chunk)
+                        self.download_status = chunk_data['status']
+                        if 'total' in chunk_data and 'completed' in chunk_data:
+                            self.download_progress = chunk_data['completed'] / chunk_data['total']
+                            print(self.download_progress)
+                        else:
+                            self.download_progress = 0
+                    except json.JSONDecodeError:
+                        pass  # Skip malformed chunks
+        except Exception as e:
+            ui.notify(f'Error: {str(e)}', type='negative')
+        finally:
+            self.create_btn.set_enabled(True)

src/components/ollama_quick_test.py

@@ -0,0 +1,65 @@
+from nicegui import ui
+from niceguiasyncelement import AsyncCard
+from pathlib import Path
+from utils import ollama
+from typing import Optional
+
+
+class ModelQuickTestComponent(AsyncCard):
+    model: Optional[str]
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self.model = None
+
+    async def build(self, model: str) -> None:
+        self.model = model
+        with self:
+            with ui.card().classes('w-full'):
+                ui.label(f'Quick Chat Test with {model}').classes('text-h6 font-bold mb-4')
+
+                self.quick_test_textarea = ui.textarea(
+                    label='Prompt',
+                    placeholder='Enter your prompt here...',
+                    value='Hello! Tell me a fun fact about AMD GPUs.'
+                ).classes('w-full').props('autogrow outlined')
+
+                self.quick_test_send = ui.button('Send', icon='send', on_click=self._quick_test).props('color=primary')
+
+                with ui.row():
+                    ui.icon('message', size='sm')
+                    ui.label('Response')
+                self.quick_test_response = ui.label('Response will appear here...').classes('text-grey-7')
+
+    async def _quick_test(self):
+        if not self.model:
+            ui.notify('Select a model first.', type='warning')
+            return
+
+        self.quick_test_response.set_text('')
+        prompt = self.quick_test_textarea.value
+
+        # calling stream_ollama_chat
+        data = {
+            "model": self.model,
+            "messages": [{"role": "user", "content": prompt}],
+            "stream": True
+        }
+        self.quick_test_send.set_enabled(False)
+        try:
+            async for chunk in ollama.stream_chat(data):
+                if chunk.strip():
+                    # Parse the JSON chunk and extract content
+                    import json
+                    try:
+                        chunk_data = json.loads(chunk)
+                        if 'message' in chunk_data and 'content' in chunk_data['message']:
+                            content = chunk_data['message']['content']
+                            current_text = self.quick_test_response.text
+                            self.quick_test_response.set_text(current_text + content)
+                    except json.JSONDecodeError:
+                        pass  # Skip malformed chunks
+        except Exception as e:
+            ui.notify(f'Error: {str(e)}', type='negative')
+        finally:
+            self.quick_test_send.set_enabled(True)

src/components/sidebar.py

@@ -8,29 +8,17 @@ class Sidebar:
 
             with ui.column().classes('w-full h-full p-4'):
 
-                # Navigation sections
-                ui.label('MAIN').classes('text-xs text-grey-5 font-bold tracking-wide mb-2')
-
                 with ui.column().classes('gap-1 mb-6'):
                     self._nav_item('Dashboard', 'dashboard', '/', active=(current_route == '/'))
                     self._nav_item('System Overview', 'monitor', '/system', active=(current_route == '/system'))
 
-                ui.label('MANAGEMENT').classes('text-xs text-grey-5 font-bold tracking-wide mb-2')
-
-                with ui.column().classes('gap-1 mb-6'):
-                    self._nav_item('Ollama Manager', 'smart_toy', '/ollama', active=(current_route == '/ollama'))
-                    self._nav_item('Process Manager', 'terminal', '/processes', active=(current_route == '/processes'))
-                    self._nav_item('Network Monitor', 'router', '/network', active=(current_route == '/network'))
-                    self._nav_item('Package Manager', 'inventory_2', '/packages', active=(current_route == '/packages'))
-
                 ui.label('TOOLS').classes('text-xs text-grey-5 font-bold tracking-wide mb-2')
 
                 with ui.column().classes('gap-1 mb-6'):
-                    self._nav_item('Log Viewer', 'description', '/logs', active=(current_route == '/logs'))
-                    self._nav_item('System Info', 'info', '/info', active=(current_route == '/info'))
+                    self._nav_item('Censor', 'description', '/censor', active=(current_route == '/censor'))
 
                 ui.space()
-
+                self._nav_item('Model Manager', 'view_in_ar', '/ollama', active=(current_route == '/ollama'))
                 # Bottom section
                 ui.separator().classes('my-4')
                 self._nav_item('Settings', 'settings', '/settings', active=(current_route == '/settings'))

src/main.py

@@ -5,7 +5,7 @@ from nicegui import ui, app
 
 from components import Header, Sidebar
 from pages import DashboardPage, OllamaManagerPage
-from utils import GPUMonitor, SystemMonitor
+from utils import GPUMonitor, SystemMonitor, OllamaMonitor
 import logging
 
 logging.basicConfig(
@@ -22,9 +22,11 @@ app.add_static_files('/static', 'src/static')
 # Create monitor instances (bindable dataclasses)
 system_monitor = SystemMonitor()
 gpu_monitor = GPUMonitor()
+ollama_monitor = OllamaMonitor()
 
 app.timer(2.0, system_monitor.update)
 app.timer(2.0, gpu_monitor.update)
+app.timer(2.0, ollama_monitor.update)
 
 
 def create_layout(current_route='/'):
@@ -35,7 +37,7 @@ def create_layout(current_route='/'):
     # Add custom CSS
     ui.add_head_html('<link rel="stylesheet" type="text/css" href="/static/style.css">')
 
-    Header(system_monitor, gpu_monitor)
+    Header(system_monitor, gpu_monitor, ollama_monitor)
     Sidebar(current_route)
 
 

src/pages/dashboard.py

@@ -1,6 +1,6 @@
 from typing import Literal
 from nicegui import ui
-from components.circular_progress import MetricCircle, LargeMetricCircle, ColorfulMetricCard
+from components.circular_progress import MetricCircle, LargeMetricCircle, ColorfulMetricCard, MetricCircleAdv
 from utils import SystemMonitor, GPUMonitor
 
 """
@@ -58,6 +58,8 @@ class DashboardPage(ui.column):
         # Main content area with proper viewport handling
         with self:
             with ui.column().classes('w-full max-w-6xl mx-auto p-6 gap-6'):
+                with ui.grid(columns=4).classes('w-full gap-4'):
+                    MetricCircleAdv('CPU', system_monitor, 'cpu_percent', '', icon='memory', formatting='percent', color='#e879f9')
                 # Top stats grid
                 with ui.grid(columns=4).classes('w-full gap-4'):
                     # CPU metric with binding

src/pages/ollama_manager.py

@@ -3,6 +3,9 @@ from utils import ollama
 from typing import Literal, List, Dict
 from pprint import pprint
 from niceguiasyncelement import AsyncColumn
+from components.ollama_downloader import OllamaDownloaderComponent
+from components.ollama_model_creation import OllamaModelCreationComponent
+from components.ollama_quick_test import ModelQuickTestComponent
 
 
 class OllamaManagerPage(AsyncColumn):
@@ -44,53 +47,32 @@ class OllamaManagerPage(AsyncColumn):
                 with ui.row().classes('w-full items-center mb-4'):
                     ui.label('Installed Models').classes('text-h6 font-bold')
                     ui.space()
-                    ui.button('Create New Model', icon='create', on_click=self._create_model).props('color=primary')
-                    ui.button('Pull New Model', icon='download').props('color=primary')
+                    ui.button('Create New Model', icon='create', on_click=self._create_model_dialog).props('color=primary')
+                    ui.button('Pull New Model', icon='download', on_click=self._download_model_dialog).props('color=primary')
 
                 with ui.column().classes('w-full gap-2'):
                     await self.models_container()  # type: ignore
 
-            # Quick test
-            with ui.card().classes('w-full'):
-                ui.label('Quick Chat Test').classes('text-h6 font-bold mb-4')
+    async def _create_model_dialog(self):
 
-                with ui.row().classes('w-full gap-2 mb-2'):
-                    self.quick_test_select = ui.select(
-                        [],
-                        label='Model'
-                    ).classes('flex-grow').props('outlined')
+        with ui.dialog() as dialog:
+            await OllamaModelCreationComponent.create()
+        await dialog
+        self.models_container.refresh()
 
-                self.quick_test_textarea = ui.textarea(
-                    label='Prompt',
-                    placeholder='Enter your prompt here...',
-                    value='Hello! Tell me a fun fact about AMD GPUs.'
-                ).classes('w-full').props('outlined')
+    async def _download_model_dialog(self):
+        with ui.dialog() as dialog:
+            await OllamaDownloaderComponent.create()
+        await dialog
+        self.models_container.refresh()
 
-                self.quick_test_send = ui.button('Send', icon='send', on_click=self._quick_test).props('color=primary')
-
-                with ui.row():
-                    ui.icon('message', size='sm')
-                    ui.label('Response')
-                self.quick_test_response = ui.label('Response will appear here...').classes('text-grey-7')
-        await self._quick_test_populate_options()
-
-    async def _create_model(self):
-        modelfile = """FROM qwen2.5-coder:7b
-        PARAMETER num_ctx 32768
-        PARAMETER temperature 0.1
-        SYSTEM "Du bist ein Python-Experte."
-        """
-        print('creating model')
-        result = await ollama.create_ollama_model(
-            "qwen2.5-coder-32k-python",
-            modelfile
-        )
-        print('finished.')
-        print(result)
-        await self.models_container.refresh()
+    async def _test_model_dialog(self, model):
+        with ui.dialog() as dialog:
+            await ModelQuickTestComponent.create(model)
+        await dialog
 
     async def _loaded_models(self):
-        loaded = await ollama.loaded_models()
+        loaded = await ollama.active_models()
         print(loaded)
 
     async def _delete_model(self, model):
@@ -114,8 +96,6 @@ class OllamaManagerPage(AsyncColumn):
 
         for model in self.models:
             self._create_model_item(model)
-        if hasattr(self, 'quick_test_select'):
-            await self._quick_test_populate_options()
 
     def _create_model_item(self, model: Dict):
         with ui.card().classes('w-full'):
@@ -169,48 +149,14 @@ class OllamaManagerPage(AsyncColumn):
                 ui.space()
 
                 with ui.row().classes('gap-2'):
+                    ui.button(icon='chat', on_click=lambda m=model['name']: self._test_model_dialog(m)).props('round flat').tooltip('Test Model')
                     ui.button(icon='play_arrow').props('round flat color=primary').tooltip('Run Model')
                     ui.button(icon='info', on_click=lambda m=model['name']: self._print_model_info(m)).props('round flat').tooltip('Model Info')
                     ui.button(icon='delete', on_click=lambda m=model['name']: self._delete_model(m)).props('round flat color=negative').tooltip('Delete Model')
 
     async def _print_model_info(self, model_name):
         result = await ollama.model_info(model_name)
-        print(result)
+        for key, value in result.items():
+            print(key)
 
-    async def _quick_test_populate_options(self):
-        select_options = [model['name'] for model in self.models]
-        self.quick_test_select.set_options(select_options)
-
-    async def _quick_test(self):
-        model = self.quick_test_select.value
-        if not model:
-            ui.notify('Select a model first.', type='warning')
-            return
-
-        self.quick_test_response.set_text('')
-        prompt = self.quick_test_textarea.value
-
-        # calling stream_ollama_chat
-        data = {
-            "model": model,
-            "messages": [{"role": "user", "content": prompt}],
-            "stream": True
-        }
-        self.quick_test_send.set_enabled(False)
-        try:
-            async for chunk in ollama.stream_chat(data):
-                if chunk.strip():
-                    # Parse the JSON chunk and extract content
-                    import json
-                    try:
-                        chunk_data = json.loads(chunk)
-                        if 'message' in chunk_data and 'content' in chunk_data['message']:
-                            content = chunk_data['message']['content']
-                            current_text = self.quick_test_response.text
-                            self.quick_test_response.set_text(current_text + content)
-                    except json.JSONDecodeError:
-                        pass  # Skip malformed chunks
-        except Exception as e:
-            ui.notify(f'Error: {str(e)}', type='negative')
-        finally:
-            self.quick_test_send.set_enabled(True)
+        print(result['modelfile'])

src/static/style.css

@@ -168,7 +168,7 @@ body,
 }
 
 .main-content {
-  height: calc(100vh - 64px) !important;
+  height: calc(100vh - 0px) !important;
   width: calc(100vw - 256px) !important;
   overflow-y: auto !important;
   margin-left: 256px !important;

src/utils/__init__.py

@@ -1,4 +1,5 @@
 from .system_monitor import SystemMonitor
 from .gpu_monitor import GPUMonitor
+from .ollama_monitor import OllamaMonitor
 
-__all__ = ['SystemMonitor', 'GPUMonitor']
+__all__ = ['SystemMonitor', 'GPUMonitor', 'OllamaMonitor']

src/utils/ollama.py

@@ -1,6 +1,6 @@
 import httpx
 from nicegui import ui
-from typing import Tuple
+from typing import Tuple, Dict
 
 
 async def status(url='http://127.0.0.1:11434') -> Tuple[bool, str]:
@@ -24,27 +24,30 @@ async def available_models(url='http://127.0.0.1:11434'):
         return response.json()["models"]
 
 
-async def loaded_models(url='http://127.0.0.1:11434'):
+async def active_models(url='http://127.0.0.1:11434'):
     async with httpx.AsyncClient() as client:
         response = await client.get(f"{url}/api/ps")
         response.raise_for_status()
 
-        return response.json()
+        return response.json()["models"]
 
 
-async def create_ollama_model(name, modelfile_content, url='http://127.0.0.1:11434'):
+async def create_ollama_model(model_name, model_from, parameters=None, quantizie=None, url='http://127.0.0.1:11434'):
     data = {
-        "name": name,
-        "from": "qwen2.5-coder:7b",
-        "modelfile": modelfile_content,
-        "stream": False
+        "model": model_name,
+        "from": model_from,
+        "stream": True
     }
 
+    if parameters:
+        data['parameters'] = parameters
+    if quantizie:
+        data['quantizie'] = quantizie
+
     async with httpx.AsyncClient() as client:
-        response = await client.post(f"{url}/api/create", json=data)
-        response.raise_for_status()
-        print(response.text)
-        return response.json()
+        async with client.stream('POST', f"{url}/api/create", json=data) as response:
+            async for chunk in response.aiter_text():
+                yield chunk
 
 
 async def delete_model(name, url='http://127.0.0.1:11434') -> bool:
@@ -79,3 +82,11 @@ async def stream_chat(data, url='http://127.0.0.1:11434'):
         async with client.stream('POST', f"{url}/api/chat", json=data) as response:
             async for chunk in response.aiter_text():
                 yield chunk
+
+
+async def download_model(model, url='http://127.0.0.1:11434'):
+    data = {'name': model, 'stream': True}
+    async with httpx.AsyncClient() as client:
+        async with client.stream('POST', f"{url}/api/pull", json=data) as response:
+            async for chunk in response.aiter_text():
+                yield chunk

src/utils/ollama_monitor.py

@@ -0,0 +1,21 @@
+import psutil
+import platform
+import time
+import logging
+from dataclasses import dataclass, field
+from datetime import datetime, timedelta
+from typing import Dict, Any, List, Optional, Literal
+from nicegui import binding
+from utils import ollama
+logger = logging.getLogger(__name__)
+
+
+@binding.bindable_dataclass
+class OllamaMonitor:
+    status: bool = False
+    version: str = 'Unknown'
+    active_models: Dict = field(default_factory=dict)
+
+    async def update(self):
+        self.status, self.version = await ollama.status()
+        self.active_models = await ollama.active_models() if self.status else {}