Closed jleinonen closed 1 month ago
Yeah, I put in the fixed time out to have a cap on the download. Figured 100s would be long enough for people but we can increase to say 10 minutes. The idea was to push people away from very large requests into memory in one go. But rather split it up with multiple calls. But if you are timing out even for just 73 channels, then we can bump it up.
Of course in the mean time you can change it via:
ds = ARCO()
ds.async_timeout = 1200
If this continues to be a bigger problem, I'll change this to use an environment variable instead similar to the model packages, but I expect 10 minutes to be sufficiently enough time for the models we have.
Yes, I got a timeout for one timestep and 73 channels. Thanks for the change, that should fix it for me given that I managed to download about 50% of the data before it timed out.
Version
Latest from Github
On which installation method(s) does this occur?
Source
Describe the issue
When trying to inference the example workflow from https://github.com/NVIDIA/earth2studio/issues/91#issuecomment-2229512188, I get the following error while the script is downloading data:
It looks like the ARCO data source hardcodes https://github.com/NVIDIA/earth2studio/blob/cb1c2306467f013601fc606596a2a5da1de4fa5d/earth2studio/data/arco.py#L85 so maybe that is the issue? If to, making the timeout longer and/or user configurable would probably solve the problem.