Microsoft.Extensions.AI.Ollama 9.0.0-preview.9.24556.5

Microsoft.Extensions.AI.Ollama

Provides an implementation of the IChatClient interface for Ollama.

Install the package

From the command-line:

dotnet add package Microsoft.Extensions.AI.Ollama

Or directly in the C# project file:

<ItemGroup>
  <PackageReference Include="Microsoft.Extensions.AI.Ollama" Version="[CURRENTVERSION]" />
</ItemGroup>

Usage Examples

Chat

using Microsoft.Extensions.AI;

IChatClient client = new OllamaChatClient(new Uri("http://localhost:11434/"), "llama3.1");

Console.WriteLine(await client.CompleteAsync("What is AI?"));

Chat + Conversation History

using Microsoft.Extensions.AI;

IChatClient client = new OllamaChatClient(new Uri("http://localhost:11434/"), "llama3.1");

Console.WriteLine(await client.CompleteAsync(
[
    new ChatMessage(ChatRole.System, "You are a helpful AI assistant"),
    new ChatMessage(ChatRole.User, "What is AI?"),
]));

Chat Streaming

using Microsoft.Extensions.AI;

IChatClient client = new OllamaChatClient(new Uri("http://localhost:11434/"), "llama3.1");

await foreach (var update in client.CompleteStreamingAsync("What is AI?"))
{
    Console.Write(update);
}

Tool Calling

Known limitations:

• Only a subset of models provided by Ollama support tool calling.
• Tool calling is currently not supported with streaming requests.

using System.ComponentModel;
using Microsoft.Extensions.AI;

IChatClient ollamaClient = new OllamaChatClient(new Uri("http://localhost:11434/"), "llama3.1");

IChatClient client = new ChatClientBuilder()
    .UseFunctionInvocation()
    .Use(ollamaClient);

ChatOptions chatOptions = new()
{
    Tools = [AIFunctionFactory.Create(GetWeather)]
};

Console.WriteLine(await client.CompleteAsync("Do I need an umbrella?", chatOptions));

[Description("Gets the weather")]
static string GetWeather() => Random.Shared.NextDouble() > 0.5 ? "It's sunny" : "It's raining";

Caching

using Microsoft.Extensions.AI;
using Microsoft.Extensions.Caching.Distributed;
using Microsoft.Extensions.Caching.Memory;
using Microsoft.Extensions.Options;

IDistributedCache cache = new MemoryDistributedCache(Options.Create(new MemoryDistributedCacheOptions()));

IChatClient ollamaClient = new OllamaChatClient(new Uri("http://localhost:11434/"), "llama3.1");

IChatClient client = new ChatClientBuilder()
    .UseDistributedCache(cache)
    .Use(ollamaClient);

for (int i = 0; i < 3; i++)
{
    await foreach (var message in client.CompleteStreamingAsync("In less than 100 words, what is AI?"))
    {
        Console.Write(message);
    }

    Console.WriteLine();
    Console.WriteLine();
}

Telemetry

using Microsoft.Extensions.AI;
using OpenTelemetry.Trace;

// Configure OpenTelemetry exporter
var sourceName = Guid.NewGuid().ToString();
var tracerProvider = OpenTelemetry.Sdk.CreateTracerProviderBuilder()
    .AddSource(sourceName)
    .AddConsoleExporter()
    .Build();

IChatClient ollamaClient = new OllamaChatClient(new Uri("http://localhost:11434/"), "llama3.1");

IChatClient client = new ChatClientBuilder()
    .UseOpenTelemetry(sourceName, c => c.EnableSensitiveData = true)
    .Use(ollamaClient);

Console.WriteLine(await client.CompleteAsync("What is AI?"));

Telemetry, Caching, and Tool Calling

using System.ComponentModel;
using Microsoft.Extensions.AI;
using Microsoft.Extensions.Caching.Distributed;
using Microsoft.Extensions.Caching.Memory;
using Microsoft.Extensions.Options;
using OpenTelemetry.Trace;

// Configure telemetry
var sourceName = Guid.NewGuid().ToString();
var tracerProvider = OpenTelemetry.Sdk.CreateTracerProviderBuilder()
    .AddSource(sourceName)
    .AddConsoleExporter()
    .Build();

// Configure caching
IDistributedCache cache = new MemoryDistributedCache(Options.Create(new MemoryDistributedCacheOptions()));

// Configure tool calling
var chatOptions = new ChatOptions
{
    Tools = [AIFunctionFactory.Create(GetPersonAge)]
};

IChatClient ollamaClient = new OllamaChatClient(new Uri("http://localhost:11434/"), "llama3.1");

IChatClient client = new ChatClientBuilder()
    .UseDistributedCache(cache)
    .UseFunctionInvocation()
    .UseOpenTelemetry(sourceName, c => c.EnableSensitiveData = true)
    .Use(ollamaClient);

for (int i = 0; i < 3; i++)
{
    Console.WriteLine(await client.CompleteAsync("How much older is Alice than Bob?", chatOptions));
}

[Description("Gets the age of a person specified by name.")]
static int GetPersonAge(string personName) =>
    personName switch
    {
        "Alice" => 42,
        "Bob" => 35,
        _ => 26,
    };

Text embedding generation

using Microsoft.Extensions.AI;

IEmbeddingGenerator<string, Embedding<float>> generator =
    new OllamaEmbeddingGenerator(new Uri("http://localhost:11434/"), "all-minilm");

var embeddings = await generator.GenerateAsync("What is AI?");

Console.WriteLine(string.Join(", ", embeddings[0].Vector.ToArray()));

Text embedding generation with caching

using Microsoft.Extensions.AI;
using Microsoft.Extensions.Caching.Distributed;
using Microsoft.Extensions.Caching.Memory;
using Microsoft.Extensions.Options;

IDistributedCache cache = new MemoryDistributedCache(Options.Create(new MemoryDistributedCacheOptions()));

IEmbeddingGenerator<string, Embedding<float>> ollamaGenerator =
    new OllamaEmbeddingGenerator(new Uri("http://localhost:11434/"), "all-minilm");

IEmbeddingGenerator<string, Embedding<float>> generator = new EmbeddingGeneratorBuilder<string, Embedding<float>>()
    .UseDistributedCache(cache)
    .Use(ollamaGenerator);

foreach (var prompt in new[] { "What is AI?", "What is .NET?", "What is AI?" })
{
    var embeddings = await generator.GenerateAsync(prompt);

    Console.WriteLine(string.Join(", ", embeddings[0].Vector.ToArray()));
}

Dependency Injection

using Microsoft.Extensions.AI;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Hosting;
using Microsoft.Extensions.Logging;

// App Setup
var builder = Host.CreateApplicationBuilder();
builder.Services.AddDistributedMemoryCache();
builder.Services.AddLogging(b => b.AddConsole().SetMinimumLevel(LogLevel.Trace));

builder.Services.AddChatClient(b => b
    .UseDistributedCache()
    .UseLogging()
    .Use(new OllamaChatClient(new Uri("http://localhost:11434/"), "llama3.1")));

var app = builder.Build();

// Elsewhere in the app
var chatClient = app.Services.GetRequiredService<IChatClient>();
Console.WriteLine(await chatClient.CompleteAsync("What is AI?"));

Minimal Web API

using Microsoft.Extensions.AI;

var builder = WebApplication.CreateBuilder(args);

builder.Services.AddChatClient(c =>
    c.Use(new OllamaChatClient(new Uri("http://localhost:11434/"), "llama3.1")));

builder.Services.AddEmbeddingGenerator<string,Embedding<float>>(g =>
    g.Use(new OllamaEmbeddingGenerator(endpoint, "all-minilm")));

var app = builder.Build();

app.MapPost("/chat", async (IChatClient client, string message) =>
{
    var response = await client.CompleteAsync(message, cancellationToken: default);
    return response.Message;
});

app.MapPost("/embedding", async (IEmbeddingGenerator<string,Embedding<float>> client, string message) =>
{
    var response = await client.GenerateAsync(message);
    return response[0].Vector;
});

app.Run();

Feedback & Contributing

We welcome feedback and contributions in our GitHub repo.