Verifier Fix

Created by: jjenthought

The verifier now warns when all outputs are unassigned in a datasource. While this is likely to change in future, it's more reasonable then what we have now. Before it was warning only when a single output was missing, which was a mistake on my part when setting it up.

Created by: codecov-io

Codecov Report

Merging #176 into master will decrease coverage by 0.02%. The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #176      +/-   ##
==========================================
- Coverage   93.85%   93.83%   -0.03%     
==========================================
  Files          43       43              
  Lines        1042     1038       -4     
  Branches       83       81       -2     
==========================================
- Hits          978      974       -4     
  Misses         60       60              
  Partials        4        4

Impacted Files	Coverage Δ
force_bdss/core/verifier.py	90% <100%> (-0.39%)

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 16a69a9...eba44f0. Read the comment docs.

Created by: stefanoborini

I would change this logic to check only if all are undefined. Or as we said, the possibility that all could be undefined is not zero, although unlikely.

Created by: jjenthought

So after we spoke my thought was: if you have a datasource in the workflow just for a side-effect it’s reasonably painless to just go and name one of the output parameters, even if you don’t want to use it in the end. However it would be quite annoying to run a simulation having forgotten to name something that you did want to keep track of. But that’s all pretty subjective!

Were you thinking of having it so you only get an error if all outputs in all datasources are undefined? At the minute it gives errors if all outputs in a single datasource are undefined.

James

On 23 Jul 2018, at 11:56, Stefano Borini notifications@github.com wrote:

@stefanoborini requested changes on this pull request.

In force_bdss/core/verifier.py https://github.com/force-h2020/force-bdss/pull/176#discussion_r204359520:

@@ -178,16 +182,18 @@ def _check_data_source(data_source_model, layer_number): global_error="Missing output slot name assignment " "in layer {}".format(layer_number)))

• #: Check if any datasources have all outputs unnamed row_index_errors = [] I would change this logic to check only if all are undefined. Or as we said, the possibility that all could be undefined is not zero, although unlikely.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/force-h2020/force-bdss/pull/176#pullrequestreview-139423881, or mute the thread https://github.com/notifications/unsubscribe-auth/Al4y5arMCAGiLC830SE7vi79MgPoX3bjks5uJavygaJpZM4Vaw-W.

Created by: stefanoborini

Good point. So let's keep it as "at least one". We'll have to add a "trash" datasource later.

Created by: stefanoborini

In any case, the current logic can just check for the names, not get the errors and check that