Open ircwaves opened 6 years ago
What's the actual purpose of the inf? Just to allow any amount of latency? If so you could base it on the start date of the asset.
@naterubin -- I think inf
was to make it so that it would never attempt to fetch, due to "not being available yet". Is that right?
That is correct, small discussion about it here: https://github.com/Applied-GeoSolutions/gips/pull/327
Right. I'm inclined to say that non-fetchables should be guarded against by their fetch methods (which landsat does). The 'latency'
may imply something about fetching, but actually it doesn't. 'SR'
assets have about the same latency as 'C1'
assets, they're just not fetchable at the moment. 'DN'
is no longer available, so that's a whole other ball of rats. So, I still vote for dropping the block in question.
Other thoughts?
https://github.com/Applied-GeoSolutions/gips/blob/60927408cf0ef41abc9147d9c0e2a68dbba909f8/gips/data/landsat/landsat.py#L648-L652
That makes for a blowout (below) on the
dhscratch
branch (which isdatahandler
withdev
merged in). At the moment, I think the simplest fix is to remove this line completely. A better fix involving supporting floating point latencies (which is reasonable) and usinginf
as a guard for things that cannot be fetched may make sense.