fix(openai): Auto-Context, Tool-Result-Matching, Payload-Preview, leere Bubbles

- Auto-Context (Embedding-Suche) für OpenAI-Provider aktiviert: BuildAutoContextAsync
  auf internal static refactored und in StreamChat + ChatNonStreaming integriert
- Tool-Results mit Role="tool" (OpenAI-Konvention) werden in BuildHistoryFromDb
  jetzt korrekt erkannt (vorher nur Role="user"/Anthropic)
- Payload-Preview zeigt für OpenAI-Provider echtes API-Format (function-Wrapper,
  tool_choice:"auto", max_tokens, stream_options, extra_body)
- Tool-Input (Argumente) wird im SSE tool_end-Event mitgesendet (fehlte im OpenAI-Pfad)
- Leere Bubbles durch Whitespace-only Text nach </think>-Extraktion behoben
  (Trim in DB-Save, BuildHistoryFromDb und Frontend-Guard)
- History-Toggle-Bug: var→const in for-Loop (Block-Scoping für Closures),
  tool_result-Matching per data-tool-call-id statt blindes letztes Element
- Non-Streaming max_tokens für MiniMax auf 32768 + extra_body.reasoning_split

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Endpoints/SessionEndpoint.cs

@@ -317,8 +317,54 @@ public static class SessionEndpoint
         var toolChoice = hasTools || !string.IsNullOrEmpty(agent.WebSearchVersion) ? "any" : null;
         var effectiveChoice = toolChoice == "any" ? "any" : "auto";
 
-        // Payload zusammenbauen (Anthropic-Format)
-        var payload = new Dictionary<string, object>
+        // OpenAI-kompatible Provider: Payload im tatsächlichen API-Format anzeigen
+        if (previewProvider)
+        {
+            var (provider, _) = OpenAiService.ParseModelString(modelId);
+            var maxTokens = provider == "minimax" ? 32768 : 16384;
+
+            // Messages konvertieren (Anthropic → OpenAI-Format)
+            var openAiMessages = OpenAiService.BuildApiMessages(systemPromptText, messages);
+
+            // Tools konvertieren (input_schema → function/parameters)
+            var openAiTools = tools.Count > 0
+                ? OpenAiService.ConvertTo_openAiToolDefs(tools.ToArray())
+                : new List<object>();
+
+            var payload = new Dictionary<string, object>
+            {
+                ["model"] = modelId,
+                ["max_tokens"] = maxTokens,
+                ["messages"] = openAiMessages,
+                ["stream"] = true,
+                ["stream_options"] = new { include_usage = true }
+            };
+
+            if (openAiTools.Count > 0)
+            {
+                payload["tools"] = openAiTools;
+                payload["tool_choice"] = "auto";
+            }
+
+            if (provider == "minimax")
+                payload["extra_body"] = new { reasoning_split = true };
+
+            if (provider == "nvidia")
+            {
+                var nvidiaModelId = modelId.Contains(":") ? modelId.Split(':')[1] : modelId;
+                if (nvidiaModelId.Contains("kimi-k2", StringComparison.OrdinalIgnoreCase)
+                    || nvidiaModelId.Contains("qwq", StringComparison.OrdinalIgnoreCase)
+                    || nvidiaModelId.Contains("deepseek-r1", StringComparison.OrdinalIgnoreCase))
+                    payload["chat_template_kwargs"] = new { thinking = true };
+            }
+
+            ctx.Response.ContentType = "application/json; charset=utf-8";
+            await ctx.Response.WriteAsync(JsonSerializer.Serialize(payload, new JsonSerializerOptions { WriteIndented = true }));
+            return;
+        }
+
+        // Anthropic-Format Payload
+        var anthropicPayload = new Dictionary<string, object>
         {
             ["model"] = modelId,
             ["max_tokens"] = 16384,
@@ -327,16 +373,16 @@ public static class SessionEndpoint
         };
 
         if (systemPrompt.Count > 0)
-            payload["system"] = systemPrompt;
+            anthropicPayload["system"] = systemPrompt;
 
         if (tools.Count > 0)
         {
-            payload["tools"] = tools;
-            payload["tool_choice"] = new { type = effectiveChoice };
+            anthropicPayload["tools"] = tools;
+            anthropicPayload["tool_choice"] = new { type = effectiveChoice };
         }
 
         ctx.Response.ContentType = "application/json; charset=utf-8";
-        await ctx.Response.WriteAsync(JsonSerializer.Serialize(payload, new JsonSerializerOptions { WriteIndented = true }));
+        await ctx.Response.WriteAsync(JsonSerializer.Serialize(anthropicPayload, new JsonSerializerOptions { WriteIndented = true }));
     }
 
     private static async Task HandleStreaming(HttpContext ctx, DatabaseService db, StreamingTracker streamingTracker)