Closed zackarno closed 1 year ago
@zackarno is there an objects directory inside _targets
?
@zackarno is there an objects directory inside
_targets
?
yes they get made when the targets pipeline is run with targets::tar_make()
. Some of them take quite a while to make so I recommend running it over night when you do not need your computer (or in the background). Also a few use GEE, so you would need to have {rgee}
set up and able to authenticate.
If you want to run without GEE components you can do that by commenting out those components in the _targets.R
script
Also if you just want to inspect and explore some of the targets I've copied the latest version of them here: Data/private/processed/yem/targets_20230526
. There is a way to run {targets}
so it outputs them directly into cloud buckets like AWS, but haven't figure that out yet.
Also FYI - with aee555d I copied the chirps-gefs zonal stats to the gdrive as a csv
so that you can use it as input for python script for mae/bias analysis. It might need to just a bit of formatting to fit your code, but should be straight forward
Comparing the two chirps-gefs files, there is an offset of 1 day of lead time. So, for example, for data for published on the 28th of Apr 2002, the rolling sum of 4.74 is allotted to LT=8 in the file you have prepared vs the window ending in LT=7 in the one Monica prepared. Are we calling "day 0" LT=1?
Comparing the two chirps-gefs files, there is an offset of 1 day of lead time. So, for example, for data for published on the 28th of Apr 2002, the rolling sum of 4.74 is allotted to LT=8 in the file you have prepared vs the window ending in LT=7 in the one Monica prepared. Are we calling "day 0" LT=1?
sorry for the late reply, I think we cleared this up in a separately , but adding hear for clarity/transparency on the PR. Yes what was originally called 0 makes sense to call 1 as it is the first 24 hours.
thought it would be good to have a target for chirps-gefs processing so most of the analysis is in one clear place.