restart datetimes #84
Description
Related to CIROH-UA/datastreamcli#69, CIROH-UA/ngen-datastream#344, and CIROH-UA/ngen-datastream#343
The way community t-route (and, by extension, NRDS) works is that it sets t0 to the Restart_Time attribute in the restart file, or if there is no restart file, the time in the troute config's restart_parameters[start_datetime].
Currently, NRDS is set up so that (for example) at 05:00, a short range simulation will produce forecasts for each hour in 06:00-23:00 (inclusive), and the t-route configuration restart_parameters[start_datetime] is set to 06:00. So t-route internally expects that t0=0600, otherwise it complains and won't work.
This gets weird when we bring restarts generated from the NWM's analysis/assimilation into the picture. Obviously, at 05:00, we don't have outputs from the NWM's analysis/assim for 06:00. The way I currently have restarts set up is that a restart file generated from the NWM's analysis/assim's 05:00 output is named "channel_restart_" + date + "_" + 05000000.nc", and the Restart_Date is set to date_0500. So the example commands given in CIROH-UA/datastreamcli#88 are actually incorrect because they use analysis/assim values at 202603310100 as the initial states while also forecasting outputs at 202603310100.
To make the forecasting make sense and not throw any exceptions, the mentioned PR just changes the Restart_Date to be +1 hour so that it will work with t-route. We can also change the name too, if that makes it less confusing.