-
Notifications
You must be signed in to change notification settings - Fork 22
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Moi v0.10 #220
Conversation
@odow any tips on the |
Juniper supports everything so I don't know that the bridges would be used for: Juniper.jl/src/MOI_wrapper/MOI_wrapper.jl Line 358 in 3a30ee5
|
Quite a lot of tests seem to be changed and break currently like getting a variable index by name. Kind of surprised that this should be handled by the solver itself. |
For names, you could do MOI.Test.Config(
exclude = Any[
MOI.ConstraintName,
MOI.VariableName,
MOI.Name,
],
) |
@@ -15,11 +15,12 @@ end | |||
MOI.is_valid(model::Optimizer, index::MOI.Index) = MOI.is_valid(model.model_cache, index) | |||
|
|||
MOI.get(::Optimizer, ::MOI.SolverName) = "Juniper" | |||
MOI.get(::Optimizer, ::MOI.SolverVersion) = "v0.7.0" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
You can always not implement it and exclude the test.
To the contrary, because tests were opt-in, most solvers did not test against the full API of MathOptInterface, and missed implementing some critical features. The new approach is opt-out, which has exposed a lot of issues in the wrappers (many more than I expected!). |
|
Just wrap it in a cachingoptimizer. It looks like not all the tests bail early if constraint name isn't supported. |
that doesn't seem to work or am I doing something wrong? |
I'll take a look. |
Now there is a single failure:
|
Codecov Report
@@ Coverage Diff @@
## master #220 +/- ##
==========================================
+ Coverage 91.54% 91.78% +0.24%
==========================================
Files 20 20
Lines 1951 1961 +10
==========================================
+ Hits 1786 1800 +14
+ Misses 165 161 -4
Continue to review full report at Codecov.
|
@Wikunia have the feasibility pump tests always been flakey? Linux: Mac and Windows: |
Thanks a lot @odow |
Okay, fair enough. If/when I start working on nonlinear stuff I can take a longer look at some maintenance of Juniper. I've opened some issues to remind myself in the future. This is good to go by me for the near term. Things are working and updated, at least. |
Currently there are still two problems:
MOIT.runtests
(maybe without the bride optimizer?)LowerBoundAlreadySet
error in one feasibility pump problem. I'm not too familiar with the code by @blegat from removing the JuMP dependency on that. Will come back to it next Friday probably.Currently testing involves:
Not sure how to do compat only for test dependencies.
Closes #172