create API for tstop aliasing system

[isaacsas]

Per discussions @ChrisRackauckas and I had around SciML/JumpProcesses.jl#442 it would be nice to be able to have users tell us we can alias their tstops vector in JumpProcesses so we don't need to allocate a new vector for each call to solve.

[ChrisRackauckas]

My plan is to create a keyword argument alias which then has a struct ODEAliases <: AbstractAliasSpecifier etc., that holds booleans/nothing for each thing that can be aliased. nothing for default behavior, true/false for forced aliasing, de-aliasing. This can then be rolled out for every problem type uniformly, and include booleans to cover the keyword arguments.

[ChrisRackauckas]

@jClugstor does this specification make sense?

[ChrisRackauckas]

You'll find:

alias_u0: allows the solver to alias the initial condition array that is contained in the problem struct. Defaults to false.

in DiffEq

alias_A::Bool: Whether to alias the matrix A or use a copy by default. When true, algorithms like LU-factorization can be faster by reusing the memory via lu!, but care must be taken as the original input will be modified. Default is true if the algorithm is known not to modify A, otherwise is false.
alias_b::Bool: Whether to alias the matrix b or use a copy by default. When true, algorithms can write and change b upon usage. Care must be taken as the original input will be modified. Default is true if the algorithm is known not to modify b, otherwise false.

in LinearSolve. Etc. All of that should get folded into just these more general structs, and the old way should get deprecated to just setting values of the struct.

[jClugstor]

I think I get it. alias should be a keyword of ODEProblem, LinearProblem, etc. that would be of type e.g. ODEAliases and LinearAliases respectively. And each concrete subtype of AbstractAliasSpecifier should hold boolean values that correspond to the parts of their corresponding problems that can be aliased.

[ChrisRackauckas]

Yes exactly.

[jClugstor]

Should the types ODEAliases and LinearAliases etc. be defined in SciMLBase?

[ChrisRackauckas]

yes

[jClugstor]

I'm thinking that the problem types will need a new field to hold the AbstractAliasSpecifiers, does that sound right?

[ChrisRackauckas]

I don't think so, they will just be a new common kwarg to solve.

[jClugstor]

These four PRs should add everything that's needed to use this for ODEs and Linear Problems.

Basically just like you said, I added an AbstractAliasSpecifier in SciMLBase, and then subtypes can be defined for every problem type. In DiffEqBase all I needed to do was add alias to the acceptable kwargs list.

Oh and I'm not sure I got all of the version bumps right, but I think it's just that LinearSolve and OrdinaryDiffEqCore need to be using SciMLBase@2.58 and DiffEqBase@6.160 for it all to work.

#1099
SciML/SciMLBase.jl#830
SciML/LinearSolve.jl#553
SciML/OrdinaryDiffEq.jl#2503

[ChrisRackauckas]

The interface needs a few more things. We should define the interface in AbstractAliasSpecifier to have:

• A keyword argument alias = nothing, which if true would override all other options and set all aliasing to true. If false, it would override all to false. By default it should just be nothing.
• Allow for the keyword alias in solve to also take true/false which pushes it down to alias=true/false in the AbstractAliasSpecifier.
• A contract/guarantee that if all aliases are false, then the solves are thread-safe.
• Deprecate the existing alias_* keyword arguments for the general alias one. As a non-breaking change, the concrete AbstractAliasSpecifiers should match the keyword arguments that exist by default.
• Every abstract alias specifier also needs to add alias_p and alias_f
• The docstrings should follow the standard format of SciMLStyle, so ## Keyword Arguments and a list instead of just inline text.
• IntegralProblem, OptimizationProblem, NonlinearProblem, and non-ODE DiffEq seem absent from the list right now.

Any other things you see @isaacsas @ranocha @avik-pal ?

[ranocha]

There is also a kwarg alias_du0, e.g, https://github.com/SciML/OrdinaryDiffEq.jl/pull/2503/files#diff-28a5b9e0abdb761fcac81be080bd7bb47c49f41c9d6b18df041d51881f6b1f41R73

[ranocha]

• Deprecate the existing alias_* keyword arguments for the general alias one. As a non-breaking change, the concrete AbstractAliasSpecifiers should match the keyword arguments that exist by default.

I think we have to take this one step further - if users just use the old API, everything has to work as documented. See, e.g., https://github.com/SciML/OrdinaryDiffEq.jl/pull/2503/files#r1818084222

[isaacsas]

Having an alias option around the tstops vector would also be nice, and was the original motivation for this issue. I was just going to add it to JumpProcesses, but held off to use the aliasing system.