Method FindMatchingPartitions

Namespace: LMKit.Retrieval

Assembly: LM-Kit.NET.dll

FindMatchingPartitions(string, int, float, bool, bool, CancellationToken)

Initiates a search operation to find similarities within the text entries contained in all registered DataSource objects.

public List<PartitionSimilarity> FindMatchingPartitions(string text, int topK = 3, float minScore = 0.5, bool forceUniqueDataSource = false, bool forceUniqueSection = false, CancellationToken cancellationToken = default)

Parameters

text string: The text string used to find similar entries.
topK int: Maximum number of PartitionSimilarity instances to return. Default is 3.
minScore float: Minimum similarity score required (range [0 - 1]). Default is 0.5.
forceUniqueDataSource bool: If true, all returned candidates must come from the same DataSource. Default is false.
forceUniqueSection bool: If true, returned results must come from the same Section. Default false.
cancellationToken CancellationToken: A CancellationToken to cancel the operation.

Returns

List<PartitionSimilarity>: A list of PartitionSimilarity objects meeting the criteria.

Examples

using LMKit.Data;
using LMKit.Model;
using LMKit.Retrieval;
using System;
using System.Collections.Generic;

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

        // Suppose we've already added data sources via ragEngine.AddDataSource(...)

        // We want to find the top 2 partitions matching a query
        List<PartitionSimilarity> similarities = ragEngine.FindMatchingPartitions(
            text: "AI research advancements",
            topK: 2,
            minScore: 0.4f
        );

        Console.WriteLine("Similar partitions found: " + similarities.Count);
    }
}

Exceptions

ArgumentNullException: Thrown if there are no data sources registered.
OperationCanceledException: Thrown if the operation is canceled.

FindMatchingPartitions(Attachment, int, float, bool, bool, CancellationToken)

Finds the top topK PartitionSimilarity instances matching the content of the given Attachment. The attachment's content is converted to embeddings (text or vision), and a similarity search is performed against all registered data sources.

public List<PartitionSimilarity> FindMatchingPartitions(Attachment attachment, int topK = 3, float minScore = 0.5, bool forceUniqueDataSource = false, bool forceUniqueSection = false, CancellationToken cancellationToken = default)

Parameters

attachment Attachment: The Attachment containing the content to query for similar partitions.
topK int: The maximum number of PartitionSimilarity results to return. Default is 3.
minScore float: The minimum similarity score threshold (0 to 1) for returned results. Default is 0.5.
forceUniqueDataSource bool: If true, results will be constrained to a single DataSource. Default is false.
forceUniqueSection bool: If true, results will be constrained to a single Section. Default is false.
cancellationToken CancellationToken: An optional CancellationToken to cancel the operation.

Returns

List<PartitionSimilarity>: A list of PartitionSimilarity objects that meet the specified criteria.

Table of Contents

Method FindMatchingPartitions

FindMatchingPartitions(string, int, float, bool, bool, CancellationToken)

Parameters

Returns

Examples

Exceptions

FindMatchingPartitions(Attachment, int, float, bool, bool, CancellationToken)

Parameters

Returns