Lorapok-Dynamic-Ollama-LLM-Chat-Interface

Lorapok Dynamic Ollama LLM Chat Interface - API Documentation

Ollama API

The Ollama server provides a REST API for interacting with LLMs.

Base URL: http://localhost:11434

Endpoints

GET /api/tags

List available models.

Response:

{
  "models": [
    {
      "name": "qwen2.5-coder:7b-instruct",
      "size": "4.7GB",
      "modified_at": "2024-01-01T00:00:00Z"
    }
  ]
}

POST /api/generate

Generate a response from a prompt.

Request:

{
  "model": "qwen2.5-coder:7b-instruct",
  "prompt": "Write a hello world program in Python",
  "stream": true
}

Response (streaming):

{"response": "Here", "done": false}
{"response": " is", "done": false}
{"response": " a", "done": false}
...
{"response": "!", "done": true}

POST /api/chat

Chat with a model (conversational).

Request:

{
  "model": "qwen2.5-coder:7b-instruct",
  "messages": [
    {"role": "user", "content": "Hello"}
  ],
  "stream": true
}

Python Client

The included src/ollama_client.py provides a Python wrapper for the API in the Lorapok Dynamic Ollama LLM Chat Interface.

Usage

from src.ollama_client import OllamaClient

client = OllamaClient("192.168.1.100")  # Server IP

# List models
models = client.list_models()
print(models)

# Generate response
response = client.generate("Hello, how are you?")
print(response)

# Interactive chat
client.chat()

Class Methods

• __init__(server_ip): Initialize client with server IP
• list_models(): Return list of available models
• generate(prompt, model, stream): Generate response from prompt
• chat(): Start interactive chat session

VS Code Integration

The VS Code Ollama extension provides:

• Code completion
• Chat interface
• Model selection
• Context-aware suggestions

Configure the extension with your server URL for remote access.