add info on the application startup to catch errors

app.py

@@ -11,7 +11,7 @@ from langchain.chat_models import init_chat_model
 llm = init_chat_model("gpt-5-nano", 
                       model_provider="openai",
                       api_key=os.environ['OPENAI_API_KEY'])
-
+print("LLM Init.")
 
 # load retreiver
 import os
@@ -33,19 +33,20 @@ def load_from_azure(container_name, local_dir="./index"):
             file.write(container_client.download_blob(blob).readall())
 
 # Download files from Azure
+print("start download faiss")
 load_from_azure("blobcontaineravatarbot")
-
+print("ok.")
 # Load into FAISS
 # from langchain_community.embeddings import HuggingFaceEmbeddings # deprecated
 from langchain_huggingface import HuggingFaceEmbeddings
-
+print("load embeddings")
 embedding_model = HuggingFaceEmbeddings(
     model_name="intfloat/e5-base-v2",
     # multi_process=True,
     model_kwargs={"device": "cuda"},  # use cuda for faster embeddings on nbidia GPUs
     encode_kwargs={"normalize_embeddings": True},  # Set `True` for cosine similarity
 )
-
+print("load vector store")
 vectorstore = FAISS.load_local("./index", embedding_model, allow_dangerous_deserialization=True)
 
 # Include a rate limiter
@@ -77,7 +78,7 @@ class RateLimiter:
             if now - req_time < self.window
         ]
         return self.max_requests - len(self.requests[identifier])
-
+print("Rate Limit init.")
 limiter = RateLimiter(max_requests=10, window_minutes=60)
 
 # setup chatbot
@@ -181,10 +182,9 @@ os.environ["LANGSMITH_API_KEY"] = os.environ['LANGSMITH_API_KEY']
 
 # lauch gradio app
 import gradio as gr
-
 iface = gr.ChatInterface(
     predict,
     api_name="chat",
 )
-
+print("Launch ...")
 iface.launch(share=True)