PR to try to resolve the bug concerning the repeated observation space

Hello, everyone. I noticed an issue preventing training on environments that use the Repeated space, and I wrote a wrapper that lets me work around this so that I could finish working on my project, which requires one. Along the way, I realized that there were other people that might be able to make use of it, so I submitted it as a PR. I also included a script in /examples/ that demonstrates use of the Repeated space, since I couldn’t find one when I was getting started and thought it might be helpful to new users.

I’ve been interested in RLlib for a while, and I’d like to write code that’s useful to the community. If anyone sees any stylistic (I’ve already gone through lint and cleared out the obvious ones) or functional issues that I should be aware of, please let me know.

1 Like

Hi @MCW_Lad ! Thank you so much for submitting a PR! :slight_smile: It’s really awesome to see open source contributors feel inspired to submit code to the Ray project. Keep up the great work and let me know if you have any questions with your PR.

1 Like

@christina Thanks - I’ve been looking at creating another PR in the next week or so to try to resolve some issues with the integration of centralized modules into multi-agent systems, but hadn’t gotten any feedback on this one, and got a message that it was going stale. Wanted to double check that I’m doing things properly.

The DCO check is failing, but I believe I fixed the issue with a rebase. At the very least, the commits it links don’t contain the email address in the error message.