Issue 334 - local inputs connected to promoted variables

[ScottDelbecq]

This PR fixes #334 (partly solved by #332) where some variables are tagged as inputs instead of outputs when using VariableList.from_problem().

To fix this, VariableList.from_problem() was modified to handle the effect of using connect() with local inputs and promoted outputs. This behaviour can be observed when using list_variables() or the variable_viewer().

However, generate_inputs() works fine and gives the correct input list because it uses _utils.get_unconnected_input_names(). This brings up the fact that VariableList.from_problem() is too complicated for what we really need it to do and should deserve a rework in a later PR.

[christophe-david]

That works Ok, thanks.

A remark on the order of commits: it is better to implement the test, then the fix. It makes easier for the reviewer to checkout the commit with the test, see that it fails, then checkout the head of the branch and see that the test passes.
Even if you did not work in this order (though it is generally advised to do so), an interactive rebase allows to put your commit history in the right order.

I just ask for a change in a comment and it will be Ok.

[ScottDelbecq]

Thanks for you comments. Can we integrate this PR in #340?

[codecov]

Codecov Report

Merging #341 (fbe8bbb) into master (682f212) will increase coverage by 0.02%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #341      +/-   ##
==========================================
+ Coverage   91.09%   91.11%   +0.02%     
==========================================
  Files         191      191              
  Lines        7648     7655       +7     
  Branches      726      730       +4     
==========================================
+ Hits         6967     6975       +8     
+ Misses        546      542       -4     
- Partials      135      138       +3     

Impacted Files	Coverage Δ
src/fastoad/openmdao/variables.py	91.16% <100.00%> (+0.58%)	⬆️

---

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 682f212...fbe8bbb. Read the comment docs.

[christophe-david]

Thanks for the changes.
Sorry, in the previous review, I missed you commented out some code (that, indeed, should legitimately be removed).

[christophe-david]

The change has been integrated in #340.
Looking forward your review :)