feat: Add VectorSearchWithScoreAsync to expose cosine similarity distance

[Copilot]

DocumentCollection.VectorSearchAsync silently discards the Distance field from VectorSearchResult, making it impossible to threshold or rank results by relevance from the typed API.

Changes

• IDocumentCollection<TId, T> — adds VectorSearchWithScoreAsync returning IAsyncEnumerable<(T Document, float Distance)>
• DocumentCollection<TId, T> — implements the method by forwarding result.Distance alongside the fetched document
• ICollectionIndex<TId, T> / CollectionSecondaryIndex — exposes the raw VectorSearchResult stream needed by the new overload

Usage

await foreach (var (doc, score) in collection.VectorSearchWithScoreAsync("embedding", queryVector, k: 10))
{
    if (score < 0.35f) // cosine distance threshold
        yield return doc;
}

• Fixes DocumentCollection<TId,T>.VectorSearchAsync does not expose similarity score (distance) #118

[Copilot]

Copilot wasn't able to review any files in this pull request.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.