Cleanly handle EOFError during interactive CLI commands

When running `vllm chat` or `vllm complete`, handle EOFError (Ctrl+d) and
exit cleanly instead of leaving a traceback behind.

Signed-off-by: Russell Bryant <rbryant@redhat.com>

Author: russellb

vllm/cmd/openai.py

@@ -49,7 +49,10 @@ def chat(system_prompt: Optional[str], model_name: str,
 
     print("Please enter a message for the chat model:")
     while True:
-        input_message = input("> ")
+        try:
+            input_message = input("> ")
+        except EOFError:
+            return
         conversation.append({"role": "user", "content": input_message})
 
         chat_completion = client.chat.completions.create(model=model_name,
@@ -103,7 +106,10 @@ def cmd(args: argparse.Namespace) -> None:
 
         print("Please enter a message for the chat model:")
         while True:
-            input_message = input("> ")
+            try:
+                input_message = input("> ")
+            except EOFError:
+                return
             conversation.append({"role": "user", "content": input_message})
 
             chat_completion = client.chat.completions.create(