No rate limiting

~/Projects/mcat $ cat mcat.py | env/bin/python mcat.py --room "!CJDvntNEPGQmykeIrB:whatthefuck.computer" --stdin -c ~/.mcatrc 
Traceback (most recent call last):
  File "mcat.py", line 55, in <module>
    mcat.f_to_matrix(sys.stdin)
  File "mcat.py", line 24, in f_to_matrix
    self.room.send_text(line)
  File "/home/rrix/Projects/mcat/env/lib/python2.7/site-packages/matrix_client/client.py", line 169, in send_text
    return self.client.api.send_message(self.room_id, text)
  File "/home/rrix/Projects/mcat/env/lib/python2.7/site-packages/matrix_client/api.py", line 194, in send_message
    self.get_text_body(text_content, msgtype)
  File "/home/rrix/Projects/mcat/env/lib/python2.7/site-packages/matrix_client/api.py", line 183, in send_message_event
    return self._send("PUT", path, content)
  File "/home/rrix/Projects/mcat/env/lib/python2.7/site-packages/matrix_client/api.py", line 319, in _send
    code=response.status_code, content=response.text
matrix_client.api.MatrixRequestError: 429: {"errcode":"M_LIMIT_EXCEEDED","error":"Too Many Requests","retry_after_ms":1214}

The javascript SDK implements a fairly complex queuing system, I'm going to strawman something similar to that, shortly.

Start with the send path shown in matrix_client/api.py, especially send_message_event and _send, and compare it with the linked matrix-js-sdk scheduler. The issue's traceback shows Matrix rate-limit responses with retry_after_ms; done would require an agreed queuing or retry design that handles these responses.