search

SolarArbiter / solarforecastarbiter-api

HTTP API and database schema for the Solar Forecast Arbiter
https://api.solarforecastarbiter.org
MIT License
10 stars 6 forks source link

Limit variables allowed at weather stations #271

Closed lboeman closed 4 years ago

lboeman commented 4 years ago

closes #241 This adds a MYSQL procedure to check if a site has modeling parameters by checking for a non-null tracking_type column. The procedure is called from python before the subsequent call to store_obs/fx/cdf and raises a BadAPIRequestError with an appropriate message.

@alorenzo175 what do you think of this approach?

alorenzo175 commented 4 years ago

Hmm, I guess the approach I was thinking about would be to have a function like what you've written, and call that function from within the store_forecast/obs/cdf procedures and abort if appropriate. Otherwise, we could just use read_site and check in python instead of adding a new procedure.

Now that I think about it again, I'm not sure if checking in sql or checking in python is best. Let's go with python for now. We might also want to check more than just tracking type (like capacity). Also need to make sure forecast and aggregate variables match if we don't already

lboeman commented 4 years ago

Hmm, I guess the approach I was thinking about would be to have a function like what you've written, and call that function from within the store_forecast/obs/cdf procedures and abort if appropriate. Otherwise, we could just use read_site and check in python instead of adding a new procedure.

Now that I think about it again, I'm not sure if checking in sql or checking in python is best. Let's go with python for now. We might also want to check more than just tracking type (like capacity). Also need to make sure forecast and aggregate variables match if we don't already

That's a good point, this procedure is unnecessary. I had originally started working on something in line with what you intended, but I think that keeping the validation in python and the RBAC based control stuff in SQL is a good separation of concerns.

I'll remove the procedure, and check for matching aggregate variable if it doesn't exist.

alorenzo175 commented 4 years ago

I think the test failures are due to a recent pymysql update. Try restricting to 0.9.3 for datastore/test-requirements.txt and requirements.txt