Azure OpenAI: Make startGenerateImage LRO internal for clients (#24546)

* Make startGenerateImage LRO internal for clients

* Move client.tsp alongside main.tsp, add @@include

* add newline to EOF

* Missing semicolon (and full run of npx tsp format - thanks, Gerardo)

specification/cognitiveservices/OpenAI.Inference/client.tsp

@@ -0,0 +1,10 @@
+import "@azure-tools/typespec-client-generator-core";
+import "./main.tsp";
+
+using Azure.ClientGenerator.Core;
+
+// Azure-specific long-running operations should be treated as implementation details
+@@internal(Azure.OpenAI.startGenerateImage);
+
+// Some models from client-internal LROs are still desired for custom public surface
+@@include(Azure.OpenAI.ImageGenerationOptions);