Terrafirma debugging - Githubissues

valeriupredoi / bgcval2

Package for BGCVal v2.0

3 stars 0 forks source link

Terrafirma debugging #87

Closed ledm closed 1 year ago

ledm commented 1 year ago

Here's a few minor changes for Terrafirma. Including:

Removing old code from analysis_compare.py
Fix for weird time_counter, time_centered units
Using "live" volume for calculating global volume means.
Masking values that are exactly zero.

Closes #82 Addresses but does not close #72

ledm commented 1 year ago

you have a buggy bug, have a look at the failed test buds

It's here: https://github.com/valeriupredoi/bgcval2/pull/87/commits/60cf6b0fcc3ae39eb441c0bfd45dc24f0322b6dd

The tests are passing netcdf filenames as a string, but the function requires an openned netcdf as a bgcval.dataset object.

Ergo: Test is wrong, not code.

valeriupredoi commented 1 year ago

test is never wrong - it means you have changed the handling in a function - hence the test failing is a good thing (catches that and is waiting for a fix or test change); if you have tests passing on main and tests failing here it means once you "fix" the test here the behaviour in main will make it fail. I'll have a look

valeriupredoi commented 1 year ago

you were missing a check on the existence of kwargs in globalVolumeMean() standard function - always good to check if stuff's there before starting to call it :grin: