Open dabhicusp opened 1 year ago
I definitely see the value in this language feature, I think it would reduce a lot of boilerplate. Can we explore a few options for what the keyword or syntax to express this? How about you propose a few options, and we'll decide together on the best approach.
I have a suggestion for an enhancement to the
weather-dl
tool. Currently, when downloading data for amonth range
that spans multiple years, the process requires manual splitting of themonth range
into separate configurations and executing multiple jobs for each part. This process can be tedious and time-consuming.For instance, if we wish to download data from
April 1994
toAugust 2012
, and thetarget path
follows this format:(gs://xyz/{year}/{year}{month:02d}_hres_var.grb2)
, we need to split the dates into three parts as follows:1st part :
year=1994
month=04/to/12
day=all
For the 2nd part:
year=1995/to/2011
month=01/to/12
day=all
And for the last part:
year=2012
month=01/to/08
day=all
I propose an enhancement in the
weather-dl parser
by introducing a new keyword, such asmonth_year
, in the configuration. The value of this keyword would look like08-2012
or01-2022
. Additionally, we can add a script to theparser
ofweather-dl
that can correctly interpret this keyword.P.S.: A similar use-case can be found in the Arco-era5 dataset. Please refer to any file containing
sfc
in the filename for more context.