Method QueryPartitionsAsync

Namespace

LMKit.Retrieval

Assembly

LM-Kit.NET.dll

QueryPartitionsAsync(string, IEnumerable<TextPartitionSimilarity>, IConversation, CancellationToken)

Asynchronously generates a text result by querying specified text partitions within the provided chat context.

public Task<TextGenerationResult> QueryPartitionsAsync(string query, IEnumerable<RagEngine.TextPartitionSimilarity> partitionSimilarities, IConversation chat, CancellationToken cancellationToken = default)

Parameters

query string

The query text.

partitionSimilarities IEnumerable<RagEngine.TextPartitionSimilarity>

A collection of RagEngine.TextPartitionSimilarity instances.

chat IConversation

An IConversation representing the chat context.

cancellationToken CancellationToken

A CancellationToken to cancel the operation.

Returns

Task<TextGenerationResult>

A task that returns a TextGenerationResult upon completion.

Examples

using LMKit.Data;
using LMKit.Model;
using LMKit.Retrieval;
using LMKit.TextGeneration;
using LMKit.TextGeneration.Chat;
using System;
using System.Collections.Generic;
using System.Threading.Tasks;

class Example
{
    static async Task Main()
    {
        LM embeddingModel = new LM(new Uri("https://example-embedding-uri.com"));
        RagEngine ragEngine = new RagEngine(embeddingModel);

        // Suppose we have a chat model
        LM chatModel = new LM(new Uri("https://example-chat-model.com"));
        IConversation chat = new SingleTurnConversation(chatModel);

        // Assume we've already imported text, and found some similar partitions
        List<TextPartitionSimilarity> partitions = new List<TextPartitionSimilarity> {
            // ...some placeholder similarities...
        };

        string userQuery = "Summarize the main idea.";

        TextGenerationResult asyncResult = await ragEngine.QueryPartitionsAsync(userQuery, partitions, chat);
        Console.WriteLine("Generated Answer (Async):");
        Console.WriteLine(asyncResult.Text);
    }
}