Fix VectorSet review findings: async pattern, Lease types, missing methods, doc fixes

- Replace ContinueWith(OnlyOnRanToCompletion) with async/await for
  VectorSetLengthAsync and VectorSetDimensionAsync (was hanging on fault)
- Fix doc examples: VectorSetAddRequest.Create() → .Member(),
  VectorSetSimilaritySearchRequest.Create() → .ByVector()
- Add using/dispose pattern for Lease<T> in all doc examples
- Add 5 missing VectorSet methods: RandomMembersAsync,
  GetApproximateVectorAsync, GetLinksAsync, GetLinksWithScoresAsync
- Fix return types to match SE.Redis: Lease<RedisValue>?,
  Lease<VectorSetLink>?, Lease<float>?
- Total: 15 VectorSet methods now covered

Author: imperugo

doc/vectorset.md

@@ -22,30 +22,35 @@ graph LR
 var embedding = await aiModel.GetEmbeddingAsync("Red running shoes, size 42");
 
 await redis.VectorSetAddAsync("products",
-    VectorSetAddRequest.Create("shoe-123", embedding));
+    VectorSetAddRequest.Member("shoe-123", embedding));
 
 // Add with JSON attributes (metadata)
 await redis.VectorSetAddAsync("products",
-    VectorSetAddRequest.Create("shoe-456", embedding)
-        .WithAttributes("""{"category":"shoes","price":79.99,"brand":"Nike"}"""));
+    VectorSetAddRequest.Member("shoe-456", embedding, 
+        attributes: """{"category":"shoes","price":79.99,"brand":"Nike"}"""));
 ```
 
 ## Similarity Search
 
+The search returns a `Lease<T>` which **must be disposed** after use to return pooled memory.
+
 ```csharp
 // Find the 5 most similar items to a query vector
 var queryEmbedding = await aiModel.GetEmbeddingAsync("comfortable sneakers for running");
 
-var results = await redis.VectorSetSimilaritySearchAsync("products",
-    VectorSetSimilaritySearchRequest.Create(queryEmbedding, count: 5));
+using var results = await redis.VectorSetSimilaritySearchAsync("products",
+    VectorSetSimilaritySearchRequest.ByVector(queryEmbedding) with { Count = 5 });
 
-foreach (var result in results)
+if (results is not null)
 {
-    Console.WriteLine($"{result.Member}: score={result.Score:F4}");
-
-    // Get attributes for each result
-    var attrs = await redis.VectorSetGetAttributesJsonAsync("products", result.Member!);
-    Console.WriteLine($"  Attributes: {attrs}");
+    foreach (var result in results.Span)
+    {
+        Console.WriteLine($"{result.Member}: score={result.Score:F4}");
+
+        // Get attributes for each result
+        var attrs = await redis.VectorSetGetAttributesJsonAsync("products", result.Member!);
+        Console.WriteLine($"  Attributes: {attrs}");
+    }
 }
 ```
 
@@ -64,9 +69,19 @@ var dims = await redis.VectorSetDimensionAsync("products");
 // Get a random member
 var random = await redis.VectorSetRandomMemberAsync("products");
 
+// Get multiple random members
+var randoms = await redis.VectorSetRandomMembersAsync("products", 5);
+
 // Get info about the VectorSet
 var info = await redis.VectorSetInfoAsync("products");
 
+// Get the approximate vector for a member
+using var vector = await redis.VectorSetGetApproximateVectorAsync("products", "shoe-123");
+
+// Get HNSW graph neighbors
+var links = await redis.VectorSetGetLinksAsync("products", "shoe-123");
+var linksWithScores = await redis.VectorSetGetLinksWithScoresAsync("products", "shoe-123");
+
 // Remove a member
 await redis.VectorSetRemoveAsync("products", "shoe-123");
 ```
@@ -91,29 +106,32 @@ foreach (var doc in documents)
 {
     var embedding = await aiModel.GetEmbeddingAsync(doc.Content);
     await redis.VectorSetAddAsync("docs",
-        VectorSetAddRequest.Create(doc.Id, embedding)
-            .WithAttributes($"""{{ "title": "{doc.Title}" }}"""));
+        VectorSetAddRequest.Member(doc.Id, embedding,
+            attributes: $"""{{ "title": "{doc.Title}" }}"""));
 }
 
 // Query: find relevant context for a prompt
-var queryEmb = await aiModel.GetEmbeddingAsync(userQuestion);
-var context = await redis.VectorSetSimilaritySearchAsync("docs",
-    VectorSetSimilaritySearchRequest.Create(queryEmb, count: 3));
+using var context = await redis.VectorSetSimilaritySearchAsync("docs",
+    VectorSetSimilaritySearchRequest.ByVector(queryEmb) with { Count = 3 });
 ```
 
 ### Recommendations
 ```csharp
 // Find products similar to what the user just viewed
-var viewedProduct = await redis.VectorSetGetApproximateVectorAsync("products", productId);
-// Use the vector to find similar items
+using var vector = await redis.VectorSetGetApproximateVectorAsync("products", viewedProductId);
+if (vector is not null)
+{
+    using var similar = await redis.VectorSetSimilaritySearchAsync("products",
+        VectorSetSimilaritySearchRequest.ByVector(vector.Span.ToArray()) with { Count = 10 });
+}
 ```
 
 ### Semantic Search
 ```csharp
 // Search by meaning, not keywords
 var searchEmb = await aiModel.GetEmbeddingAsync("something warm for winter");
-var results = await redis.VectorSetSimilaritySearchAsync("clothing",
-    VectorSetSimilaritySearchRequest.Create(searchEmb, count: 20));
+using var results = await redis.VectorSetSimilaritySearchAsync("clothing",
+    VectorSetSimilaritySearchRequest.ByVector(searchEmb) with { Count = 20 });
 ```
 
 ## Performance Notes
@@ -122,3 +140,4 @@ var results = await redis.VectorSetSimilaritySearchAsync("clothing",
 - Approximate nearest neighbor search — extremely fast even with millions of vectors
 - Memory efficient compared to external vector databases
 - Vectors are stored directly in Redis — no external index to maintain
+- `Lease<T>` return types use pooled memory — always dispose after use

src/core/StackExchange.Redis.Extensions.Core/Abstractions/IRedisDatabase.VectorSet.cs

@@ -96,4 +96,40 @@ public partial interface IRedisDatabase
     /// <param name="flag">Behaviour markers associated with a given command.</param>
     /// <returns>A random member, or null if the VectorSet is empty.</returns>
     Task<RedisValue> VectorSetRandomMemberAsync(string key, CommandFlags flag = CommandFlags.None);
+
+    /// <summary>
+    ///     Returns multiple random members from the VectorSet stored at key.
+    /// </summary>
+    /// <param name="key">The key of the VectorSet.</param>
+    /// <param name="count">The number of random members to return.</param>
+    /// <param name="flag">Behaviour markers associated with a given command.</param>
+    /// <returns>An array of random members.</returns>
+    Task<RedisValue[]> VectorSetRandomMembersAsync(string key, long count, CommandFlags flag = CommandFlags.None);
+
+    /// <summary>
+    ///     Returns the approximate vector for a member in the VectorSet.
+    /// </summary>
+    /// <param name="key">The key of the VectorSet.</param>
+    /// <param name="member">The member to retrieve the vector for.</param>
+    /// <param name="flag">Behaviour markers associated with a given command.</param>
+    /// <returns>The approximate vector as a Lease of floats. Must be disposed after use. Null if member not found.</returns>
+    Task<Lease<float>?> VectorSetGetApproximateVectorAsync(string key, string member, CommandFlags flag = CommandFlags.None);
+
+    /// <summary>
+    ///     Returns the links (neighbors) for a member in the VectorSet's HNSW graph.
+    /// </summary>
+    /// <param name="key">The key of the VectorSet.</param>
+    /// <param name="member">The member to retrieve links for.</param>
+    /// <param name="flag">Behaviour markers associated with a given command.</param>
+    /// <returns>The linked member names. The returned Lease must be disposed after use.</returns>
+    Task<Lease<RedisValue>?> VectorSetGetLinksAsync(string key, string member, CommandFlags flag = CommandFlags.None);
+
+    /// <summary>
+    ///     Returns the links (neighbors) with similarity scores for a member in the VectorSet's HNSW graph.
+    /// </summary>
+    /// <param name="key">The key of the VectorSet.</param>
+    /// <param name="member">The member to retrieve links for.</param>
+    /// <param name="flag">Behaviour markers associated with a given command.</param>
+    /// <returns>The links with scores. The returned Lease must be disposed after use.</returns>
+    Task<Lease<VectorSetLink>?> VectorSetGetLinksWithScoresAsync(string key, string member, CommandFlags flag = CommandFlags.None);
 }

src/core/StackExchange.Redis.Extensions.Core/Implementations/RedisDatabase.VectorSet.cs

@@ -1,6 +1,5 @@
 // Copyright (c) Ugo Lattanzi.  All Rights Reserved.  Licensed under the MIT license.  See License.txt in the project root for license information.
 
-using System.Threading;
 using System.Threading.Tasks;
 
 namespace StackExchange.Redis.Extensions.Core.Implementations;
@@ -25,12 +24,16 @@ public Task<bool> VectorSetContainsAsync(string key, string member, CommandFlags
         Database.VectorSetContainsAsync(key, member, flag);
 
     /// <inheritdoc/>
-    public Task<long> VectorSetLengthAsync(string key, CommandFlags flag = CommandFlags.None) =>
-        Database.VectorSetLengthAsync(key, flag).ContinueWith(t => (long)t.Result, CancellationToken.None, TaskContinuationOptions.OnlyOnRanToCompletion, TaskScheduler.Default);
+#pragma warning disable RCS1174 // async/await required for cross-TFM int→long cast
+    public async Task<long> VectorSetLengthAsync(string key, CommandFlags flag = CommandFlags.None) =>
+        await Database.VectorSetLengthAsync(key, flag).ConfigureAwait(false);
+#pragma warning restore RCS1174
 
     /// <inheritdoc/>
-    public Task<long> VectorSetDimensionAsync(string key, CommandFlags flag = CommandFlags.None) =>
-        Database.VectorSetDimensionAsync(key, flag).ContinueWith(t => (long)t.Result, CancellationToken.None, TaskContinuationOptions.OnlyOnRanToCompletion, TaskScheduler.Default);
+#pragma warning disable RCS1174
+    public async Task<long> VectorSetDimensionAsync(string key, CommandFlags flag = CommandFlags.None) =>
+        await Database.VectorSetDimensionAsync(key, flag).ConfigureAwait(false);
+#pragma warning restore RCS1174
 
     /// <inheritdoc/>
     public Task<string?> VectorSetGetAttributesJsonAsync(string key, string member, CommandFlags flag = CommandFlags.None) =>
@@ -47,4 +50,20 @@ public Task<bool> VectorSetSetAttributesJsonAsync(string key, string member, str
     /// <inheritdoc/>
     public Task<RedisValue> VectorSetRandomMemberAsync(string key, CommandFlags flag = CommandFlags.None) =>
         Database.VectorSetRandomMemberAsync(key, flag);
+
+    /// <inheritdoc/>
+    public Task<RedisValue[]> VectorSetRandomMembersAsync(string key, long count, CommandFlags flag = CommandFlags.None) =>
+        Database.VectorSetRandomMembersAsync(key, count, flag);
+
+    /// <inheritdoc/>
+    public Task<Lease<float>?> VectorSetGetApproximateVectorAsync(string key, string member, CommandFlags flag = CommandFlags.None) =>
+        Database.VectorSetGetApproximateVectorAsync(key, member, flag);
+
+    /// <inheritdoc/>
+    public Task<Lease<RedisValue>?> VectorSetGetLinksAsync(string key, string member, CommandFlags flag = CommandFlags.None) =>
+        Database.VectorSetGetLinksAsync(key, member, flag);
+
+    /// <inheritdoc/>
+    public Task<Lease<VectorSetLink>?> VectorSetGetLinksWithScoresAsync(string key, string member, CommandFlags flag = CommandFlags.None) =>
+        Database.VectorSetGetLinksWithScoresAsync(key, member, flag);
 }