I agree that is definitely frustrating that the PsychoPy documentation does not seem to be valid. In my case, the serial port was actually a pstbox, so I was able to utilize ioHub and receive triggers and button responses this way. My code resembles the code from this post on E-prime button boxes (Using an E-Prime Button Box - #7 by richard). I’ve also heard from others at my institution that using the serial python library also works with some specific settings in place, though I have limited experience with this.
Hope this helps (and hope the PsychoPy team can update the fMRI trigger documentation at some point soon)!