ollama-d 0.2.0

ollama-d

D language bindings for the Ollama AI API - Seamless integration with local AI models

Features

• Text generation with streaming and non-streaming API
• Chat interactions with local AI models
• OpenAI-compatible endpoints support
• Model management (list, create, show models)
• Configurable timeout settings
• Simple and intuitive API design

Prerequisites

• Ollama server running locally (default: http://127.0.0.1:11434)
• Installed AI model (e.g., "llama3.2")
• Vibe.d library

Quick Examples

import ollama;
import std.stdio;

void main() {
    // Initialize Ollama client
    auto client = new OllamaClient();

    // Text generation
    auto generateResponse = client.generate("llama3.2", "Why is the sky blue?");
    writeln("Generate Response: ", generateResponse["response"].get!string);

    // Chat interaction
    Message[] messages = [Message("user", "Hello, how are you?")];
    auto chatResponse = client.chat("llama3.2", messages);
    writeln("Chat Response: ", chatResponse["message"]["content"].get!string);

    // List available models
    auto models = client.listModels();
    writeln("Available Models: ", models["models"].toString());

    // OpenAI-compatible chat completions
    auto openaiResponse = client.chatCompletions("llama3.2", messages, 50, 0.7);
    writeln("OpenAI-style Response: ", openaiResponse["choices"][0]["message"]["content"].get!string);
}

Additional Methods

• generate(): Text generation with custom options
• chat(): Conversational interactions
• listModels(): Retrieve available models
• showModel(): Get detailed model information
• createModel(): Create custom models
• chatCompletions(): OpenAI-compatible chat endpoint
• completions(): OpenAI-compatible text completion

License

MIT License