Use OpenAI's built-in support for score threshold

Rather than unnecessarily removing the nodes ourselves.
See https://platform.openai.com/docs/guides/retrieval#ranking

Author: kzu

src/dotnet-openai/Vectors/SearchCommand.cs

@@ -44,6 +44,7 @@ public override async Task<int> ExecuteAsync(CommandContext context, SearchSetti
         var content = new Dictionary<string, object>
         {
             ["query"] = settings.Query,
+            ["ranking_options"] = new { score_threshold = settings.Score }
         };
 
         if (settings.Rewrite)
@@ -96,12 +97,6 @@ public override async Task<int> ExecuteAsync(CommandContext context, SearchSetti
             return -1;
         }
 
-        foreach (var result in data.ToList())
-        {
-            if (result!["score"]!.GetValue<double>() < settings.Score)
-                data.Remove(result);
-        }
-
         if (settings.Json)
             return console.RenderJson(node, settings, cts.Token);