twhiteaker / pynwm Goto Github PK

example data for tests are in the data folder

apps.hydroshare.org is temporarily hs-apps.hydroshare.org. Check that the endpoints pynwm uses don't rely on that.

Probably good to have a module that mimics the HydroShare API, and then have other modules that call that module for streamlining common tasks such as downloading streamflow time series.

NWM v1.2 came out in March, 2018. It includes a few more variables in streamflow. Make sure our example for getting the latest forecast still works.

noaa_list.list_sims returns a set of objects with links to download NWM result files. The links are FTP links, e.g.,
ftp://ftpprd.ncep.noaa.gov/pub/data/nccf/com/nwm/prod/nwm.20170808/medium_range/nwm.t06z.medium_range.channel_rt.f003.conus.nc

In Python 2.7, using urllib.urlretrieve to get a file from an FTP link works the first time. However, subsequent calls (at least to NOAA) fail with
File "...\lib\ftplib.py", line 830, in parse227
raise error_reply, resp
IOError: [Errno ftp error] 200 Switching to Binary mode.

Suggest using NOAA's HTTP server instead of FTP.

nwmv2 includes time variable. Are we reading that or the global attribute?

HydroShare is moving toward token based authentication. That will need to be incorporated into pynwm.

See
https://hs-apps.hydroshare.org/apps/nwm-forecasts/api-page/

local tests run fine, but Travis miniconda fails in netCDF4.date2num.

hs_list.py -> def list_dates()

Get a 404 Not Found