Anderson Banihirwe

> But since all the other wrapped functions just rewrite the arguments, I think it would be good to make the use of bounds an explicit opt-in behaviour. This would...

@TomAugspurger and I are looking into this issue. We changed `get_data_chunk` to an awaited call without any success. With Tom's help, I will keep looking into this for the next...

While looking at the dashboard for the server, I am noticing that only worker is active. It is likely that #12 didn't fully address the asynchronous fetching: ![Screen Shot 2020-03-03...

Not sure if this is a sign that we still need an async client on the server side

> I did have one thing that's confusing me. Why do we ever get a Dask array back in get_data_chunk? Shouldn't that always be a chunk for a single key,...

> @rabernat, following on [#361 (comment)](https://github.com/pangeo-forge/pangeo-forge-recipes/issues/361#issuecomment-1260096498), is the only thing left to do here pinning `kerchunk>=0.0.7`? @rabernat the gentlest of bumps on this :)

@Surojit88, thank you for reporting this. This is likely a Tensorflow version issue. Unfortunately, I don't have free cycles to look into this issue any time soon. I will update...

@oussa055, Thank you for opening this issue! Unfortunately, I am very swamped with work right now. How about I get back to you some time over the weekend?

@oussa055, 1) To make sure that your input is not the problem, can you try to use a different video input and see if you encounter the same error? 2)...

It appears that at one point when the algorithm tries to compute the road horizon gap, your `projMgr.curGradient` is None, as a result, it is unable to perform `self.projMgr.curGradient -...