after bufsize=1 change, exploring word or token level stream

app.py

@@ -46,6 +46,7 @@ def generate_response_token_by_token(user_message):
     process = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True, bufsize=1)
 
     start_time = time.time()
+    alltokens = ""
     token_buffer = ''
     while True:
         # Read one character at a time
@@ -56,7 +57,8 @@ def generate_response_token_by_token(user_message):
             token_buffer += char
             if char == ' ' or char == '\n':  # Token delimiters
                 elapsed_time = time.time() - start_time  # Calculate elapsed time
-                yield f"{token_buffer} [Inference time: {elapsed_time:.2f} seconds]"
+                alltokens += token_buffer
+                yield f"{alltokens} [Inference time: {elapsed_time:.2f} seconds]"
                 token_buffer = ''  # Reset token buffer
     
     # Yield the last token if there is any