Conversation
| return MOI.set(model, attr, bridge.constraint, mapped_func) | ||
| end | ||
|
|
||
| """ |
| return MOI.get(model, ForwardObjectiveSensitivity()) | ||
| end | ||
|
|
||
| """ |
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #345 +/- ##
==========================================
+ Coverage 90.73% 91.16% +0.42%
==========================================
Files 16 16
Lines 2300 2457 +157
==========================================
+ Hits 2087 2240 +153
- Misses 213 217 +4 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
src/jump_wrapper.jl
Outdated
| JuMP.backend(model), | ||
| ForwardConstraintFunction(), | ||
| JuMP.index(con_ref), | ||
| JuMP.moi_function(value), |
There was a problem hiding this comment.
This won't work if value is not a Vector. We actually always need to look at the shame of con_ref and do a reshaping, this will be needed for example for PSD constraints
There was a problem hiding this comment.
I see JuMP.vectorize is used below, maybe it should be used here as well ?
src/bridges.jl
Outdated
| ) where {T} | ||
| indices = _square_to_triangle_indices(bridge) | ||
| tri_func = MOI.Utilities.eachscalar(func)[indices] | ||
| return MOI.set(model, attr, bridge.triangle, tri_func) |
There was a problem hiding this comment.
ForwardConstraintFunction is set like ConstraintPrimalStart is set so you should imitate
https://github.com/jump-dev/MathOptInterface.jl/blob/b5778d01ec690e5514630ff2a2fe51ae0bd480f0/src/Bridges/Constraint/bridges/SquareBridge.jl#L328-L352
So you should also set functions to sym.
There was a problem hiding this comment.
Now I followed very closely the MOI code
|
|
||
| function MOI.get( | ||
| model::MOI.ModelLike, | ||
| attr::DiffOpt.ReverseConstraintFunction, |
There was a problem hiding this comment.
This should imitate MOI.get for MOI.ConstraintDual
https://github.com/jump-dev/MathOptInterface.jl/blob/b5778d01ec690e5514630ff2a2fe51ae0bd480f0/src/Bridges/Constraint/bridges/SquareBridge.jl#L354-L416
Again, bridge.sym is ignored
There was a problem hiding this comment.
Now I followed very closely the MOI code
|
Open question: |
2 participants
Most important to agree on are the functions on jump_wrapper.jl
Later on, we will need to deprecate the MOI functions that are not compliant with #288
Claude helped me with the bridging stuff and created first version of tests.
close #103
close #283
close #288
close #305