fix: copy initials to u0 if u0 not provided to remake
#3572
Conversation
AayushSabharwal
force-pushed
the
as/remake-copy-initials
branch
from
7aa8c0c to
8a19a49
Compare
| @@ -658,6 +659,18 @@ function SciMLBase.late_binding_update_u0_p( | |||
| prob, sys::AbstractSystem, u0, p, t0, newu0, newp) | |||
| supports_initialization(sys) || return newu0, newp | |||
| u0 === missing && return newu0, (p === missing ? copy(newp) : newp) | |||
| # If the user passes `p` to `remake` but not `u0` and `u0` isn't empty, | |||
| # and if the system supports initialization (so it has initial parameters), | |||
| # and if the initialization solves for `u0`, | |||
There was a problem hiding this comment.
We do. Initialization solves for unknowns if initializeprobmap !== nothing
There was a problem hiding this comment.
You don't know what the initialization will be until solve is called.
There was a problem hiding this comment.
We have all the information about the initialization in the ODEProblem?
There was a problem hiding this comment.
We don't know which initialization method will be used until later. CheckInit and NoInit don't solve for u0.
AayushSabharwal
force-pushed
the
as/remake-copy-initials
branch
from
ddc22ca to
c385c11
Compare
|
Waiting for SciML/DiffEqBase.jl#1143 |
AayushSabharwal
force-pushed
the
as/remake-copy-initials
branch
from
f788eeb to
0aa0d8b
Compare
| getter = if meta isa InitializationMetadata | ||
| meta.get_initial_unknowns | ||
| else | ||
| getu(sys, Initial.(unknowns(sys))) |
There was a problem hiding this comment.
I assume this will get the guesses too?
There was a problem hiding this comment.
Why would the guesses come into the picture here?
There was a problem hiding this comment.
It's the same reason as https://github.com/SciML/SciMLSensitivity.jl/pull/1168/files#diff-76127b1004a68143584d305e1c9ee36126238fd5a05ddeff5985f5b5450ed2c2R92, the u0 vector on DAEs is canonically interpreted as either the initial value or the guess for u0, depending on the choice of initialization and which one exists.
There was a problem hiding this comment.
I see. That's a little more annoying, but not difficult to do. I'll fix it so this propagates guesses.
AayushSabharwal
force-pushed
the
as/remake-copy-initials
branch
from
0aa0d8b to
1450a5f
Compare
|
Cancelled CI because I forgot to add a test |
AayushSabharwal
force-pushed
the
as/remake-copy-initials
branch
from
a611042 to
997da8d
Compare
|
SciMLBase/Downstream failure is SciML/SciMLBase.jl#1006. |
Close #3570