VectorEnv gym visulization correction

[bridgesign]

This is a solution in context of #535

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[bridgesign]

@btaba Is it possible to add any testing to ensure it does not break when changes are made?

[btaba]

You can add a test in gym_test.py.

[bridgesign]

@btaba I have added tests. In addition, added kwargs in the method to allow width and height which is a part of the render function in gym api.

[btaba]

@bridgesign thanks for the PR, can you fix the failing tests?

[bridgesign]

@btaba I am not sure what the issue is. It works when I try locally. I think it is breaking because render_array is broken #558.

[btaba]

I don't think this is related to #558 . One of the tests you added is failing, maybe lower the batch size

[bridgesign]

@btaba Done

[bridgesign]

Just saw it failed again. I have absolutely no idea why the test is failing. 16 envs should not be too much right?

[btaba]

@bridgesign maybe try a smaller image or just the non-batched test? Would be great if you can narrow it down

[bridgesign]

@btaba I have tried reducing the size of the entire output similar to a single env test case which is passing. Unfortunately I am not able to reproduce the CI error on my end. I read about when this error happens and it is related to low memory. However, this should not be the case as the final size is less than that of a single environment scenario. My guess is that the issue lies with how the rendering process itself handles multiple environments which I do not really know.

When I try to do the same locally, I observe a large delay in processing the states to images even locally. Is there some kind of timeout on the test cases? It might be that is too low but I think that is unlikely. Any idea how I might be able to reproduce the issue? Or do you think it might be worth digging into the rendering process itself?