Add argument to populate repo to allow for partial year update

[dwr-psandhu]

Description

Currently the populate scripts downloads the entire period of record available for each station.

Change

Adding an argument that allows for updating only for a partial window (resolution could be year or decade) will greatly speed up the daily update and also lower the burden for the data providers services.

[dwr-psandhu]

This just needs a flag to skip populate_repo 1 and 2 and go to populate_repo step 3, i.e. year 2000 and beyond. This will greatly increase the update speed (to minutes from hours) with the understanding that data providers rarely update past decade data. If they do, the entire repo can be rebuilt with not specifying the flag.

[water-e]

The work you delineated sounds right and the trial implementation is fine, but that was never the main concern:

• You need to think about what happens when the downloads differ because of different "heads" and "bodies". It isn't important that the agencies don't update old data often ... of course they don't. But successive downloads can still lead to different conclusions (based on download success) of what station-sublocation-variable combos even exist.
• I think the key requirement here is going to be updateing our diffing tools so that they still work when diffing is truncated.
• A number of reasonably good checking routines require statistics. If you are doing these checks on a directory that only has a couple years data, you may make very different decisions (consider 2020-2022 Sac to 2020-2023 to 1986-2024). I think that will be a non-issue in the current workflow, because we would accept the new data and then autoscreen on repo/coninuous/formatted which would have all the years.