Lorapok-Dynamic-Ollama-LLM-Chat-Interface

Lorapok Dynamic Ollama LLM Chat Interface - Usage Guide

Running the Server

1. Start Ollama

   .\scripts\run_server.ps1
   

2. Start Open WebUI (if installed)
   • If using Docker: The container should start automatically
   • If using pip/source: Run the appropriate command
   • Access at http://localhost:3000 or 8080

Lorapok Dynamic Console Chat Interface (Like Gemini/Claude)

The enhanced console interface provides a rich, interactive experience similar to Gemini or Claude, powered by the Lorapok Dynamic Ollama LLM Chat Interface.

Starting the Chat

1. On Windows (PowerShell)

   python src/ollama_client.py
   

2. On Linux/Mac (Terminal)

   chmod +x chat.sh
   ./chat.sh <server_ip>
   

   Or directly:

   python3 src/ollama_client.py <server_ip>
   

3. Cross-platform usage
   • Works on Windows, macOS, and Linux
   • Requires Python 3.8+ and requests, rich packages

Available Commands

Command	Description
/help	Show help message with all commands
/model <name>	Switch to a different model
/models	List all available models with details
/pull <model>	Pull a new model from registry
/remove <model>	Remove an installed model
/history [limit]	Show conversation history (default: 10)
/search <query>	Search conversation history
/clear	Clear conversation history
/save [filename]	Save conversation to JSON file
/load <filename>	Load conversation from JSON file
/export <format>	Export conversation (markdown/text)
/stats	Show conversation and performance statistics
/bench [model]	Benchmark current or specified model
/sysinfo	Show system resource information
/config	Show current configuration settings
/set <key> <value>	Update configuration setting
/reset	Reset configuration to defaults
/exit	Exit the chat

Dynamic Features

The enhanced console interface includes several dynamic features:

🔄 Dynamic Model Management

• Auto-refresh: Models list updates automatically
• Smart search: Find models by partial name matching
• Pull new models: Download models directly from the interface
• Model details: View size, modification date, and status

💬 Advanced Conversation Management

• Persistent history: Conversations maintained across sessions
• Search functionality: Find specific messages in history
• Export options: Save conversations in multiple formats
• Statistics tracking: Monitor message counts and performance

📊 Performance Monitoring

• Real-time metrics: Response times, token counts, throughput
• Benchmarking: Test model performance with standard prompts
• System monitoring: CPU, memory, and disk usage tracking
• Historical data: Performance trends over time

⚙️ Dynamic Configuration

• Runtime settings: Change configuration without restart
• Persistent config: Settings saved to config.json
• Flexible options: Timeout, streaming, history limits, etc.
• Reset capability: Restore default settings

🎨 Rich Console Experience

• Live streaming: Real-time response updates
• Colored output: Syntax highlighting and status indicators
• Interactive tables: Formatted data display
• Progress indicators: Visual feedback for long operations

Example Session

🤖 Local LLM Chat Interface
Connected to: 192.168.0.219
Current Model: qwen2.5-coder:7b-instruct

Type your message and press Enter.
Commands: /help, /model, /history, /clear, /exit

You: Hello, how are you?
Assistant (qwen2.5-coder:7b-instruct)
Hello! I'm doing well, thank you for asking. How can I help you today?

You: /model llama2:7b
✅ Switched to: llama2:7b

You: Tell me a joke
Assistant (llama2:7b)
Why don't scientists trust atoms? Because they make up everything!

Connecting from Another PC

Using with VS Code

1. Install Ollama extension
   • Open VS Code
   • Go to Extensions (Ctrl+Shift+X)
   • Search for “Ollama” and install
2. Configure the extension
   • Set the Ollama server URL to http://<server_ip>:11434
   • Use the chat panel for AI assistance

API Usage

The Ollama API is available at http://localhost:11434

List Models

curl http://localhost:11434/api/tags

Generate Response

curl -X POST http://localhost:11434/api/generate \
  -H "Content-Type: application/json" \
  -d '{"model": "qwen2.5-coder:7b-instruct", "prompt": "Hello"}'

Managing Models

• List installed models: ollama list
• Pull a model: ollama pull <model_name>
• Remove a model: ollama rm <model_name>

Available Models

Popular models to try:

• llama2:7b
• codellama:7b
• mistral:7b
• qwen2.5-coder:7b-instruct (default)