Rename completion future to make its meaning clear

Author: vinayakdsci

shortfin/python/shortfin_apps/llm/components/generate.py

@@ -104,11 +104,11 @@ async def run(self):
             # Decode loop.
             await self.token_selector.decode(self.exec_req)
         finally:
-            self.exec_req.completed.set_success()
+            self.exec_req.request_exec_success.set_success()
             self.exec_req.free_cache_pages()
 
     async def await_completion(self):
-        await self.exec_req.completed
+        await self.exec_req.request_exec_success
         return self.index
 
     def results_callback(self, result: int | list[list[int]]):

shortfin/python/shortfin_apps/llm/components/messages.py

@@ -37,7 +37,9 @@ def __init__(
         self.input_token_ids = input_token_ids
         self.prompt_length = len(input_token_ids)
         self.done = sf.VoidFuture()
-        self.completed = sf.VoidFuture()
+        # This is set to success once the request has been decoded successfully
+        # and response is ready to be sent back to the client.
+        self.request_exec_success = sf.VoidFuture()
         self.rid = rid
         # Unique `instance_id` for token selection strategies that may need
         # to differentiate between an original req and a copy of a req.