Closed dshemetov closed 2 months ago
I'm not sure that actually exists yet, as latency "doesn't exist" at the moment. I'm adding one in step_adjust_latency
, but that's not on dev yet
Take a look at that comment, there's some text in the vignettes that comments on such a warning, maybe I misunderstood, but that warning is not showing up for the same code anymore 🤷
I removed that warning, because it was annoying. All it did was always warn if the epi_df
had as_of < today()
.
Oh that makes sense. I figured it checked if as_of < today() - 90
or something. Is a warning with a better threshold something we want in principle or should we just scrap this?
I think scrap it completely. The print methods will show all this information.
_Originally posted by @dajmcdon in https://github.com/cmu-delphi/epipredict/pull/319#discussion_r1567551994_
There is supposed to be a warning when you try to train and forecast on data with a severe latency, but it has stopped showing up. Investigate and bring back?