v9.65.0

ModelingToolkit v9.65.0

Diff since v9.64.3

Merged pull requests:

• Replace Expronicon with Moshi (#3354) (@visr)
• fix(nonlinearsystem): Fix codegen issue for vector parameters (#3382) (@SebastianM-C)
• feat: run trivial initialization in problem constructor (#3404) (@AayushSabharwal)
• docs: add doc page for FMU import capability (#3405) (@AayushSabharwal)
• fix: allow passing guesses to linearize, improve linearization docs (#3412) (@AayushSabharwal)
• feat: add map_variables_to_equations (#3417) (@AayushSabharwal)
• feat: add isInitial (#3433) (@vyudu)
• CompatHelper: add new compat entry for FMI at version 0.14 for package docs, (keep existing compat) (#3434) (@github-actions[bot])
• CompatHelper: add new compat entry for FMIZoo at version 1 for package docs, (keep existing compat) (#3435) (@github-actions[bot])
• fix: separate Initial parameters into initials portion (#3439) (@AayushSabharwal)
• feat: allow NonlinearSystem(::ODESystem) and NonlinearProblem(::ODESystem) (#3440) (@AayushSabharwal)

Closed issues:

• Trivial initialization problem isn't solved when creating ODEProblem (#3290)
• remake failures (#3396)
• Add ability to provide additional guesses to linearize and family (#3400)
• remake issues (#3410)
• Cases of creation of NonlinearSystems have been broken (#3411)
• SciMLBase - equations bug fixed, but introduces new problems (#3421)
• isInitial function (#3431)
• Faulty non-boolean error in ifelse() (#3441)
• Broken amount of parameters: much more parameters in resulting ODEProblem than symbolic parameters in MTK system (#3444)

v9.64.3

ModelingToolkit v9.64.3

Diff since v9.64.2

Merged pull requests:

• fix: fix SCCNonlinearProblem reordering parameters (#3425) (@AayushSabharwal)
• build: bump patch version (#3426) (@AayushSabharwal)

Closed issues:

• Conditional branches inside @component statements do not support multiple components (#3424)

v9.64.2

ModelingToolkit v9.64.2

Diff since v9.64.1

Merged pull requests:

• add utilities and tests for disturbance modeling (#3314) (@baggepinnen)
• BVProblem with constraints (#3323) (@vyudu)
• fix: validate u0map and pmap (#3350) (@vyudu)
• Renaming discrete variables during structural simplification (plus refactoring tearing_reassemble) (#3379) (@vyudu)
• fix: fix array varables split across SCCs in SCCNonlinearProblem (#3384) (@AayushSabharwal)
• refactor: rewrite HomotopyContinuationProblem to target NonlinearSolveHomotopyContinuation.jl (#3385) (@AayushSabharwal)
• refactor: format (#3413) (@AayushSabharwal)
• fix: make the diffvar type check less restrictive (#3415) (@vyudu)
• feat: mark getproperty(::AbstractSystem, ::Symbol) as non-differentiable (#3422) (@AayushSabharwal)
• Replace vcat(eqs...) with reduce(vcat, eqs) (#3423) (@devmotion)

Closed issues:

• Nonsensical stuff is allowed in input paraemter maps (#3325)
• Calling reinit! makes the solver unstable (#3375)
• Solution somehow corrupted with IfLifting (#3383)
• Diffvar type check introduced in v9.63 causes error with complex diffvars (#3406)
• Re-creating MTK problems causes side effects (#3407)
• Can no longer simulate coupled ReactionSystem ODESystem (#3409)
• equations not defined??? (#3414)

v9.64.1

ModelingToolkit v9.64.1

Diff since v9.64.0

Merged pull requests:

• fix: fix several bugs (#3390) (@AayushSabharwal)
• build: bump patch version (#3402) (@AayushSabharwal)

Closed issues:

• Parameter splitting breaks runtime generated jacobians (#3294)
• Cases where remake seems to be mutating the original input in some way (#3395)
• breaking change in behavior of complete in non-breaking release (#3397)

v9.64.0

ModelingToolkit v9.64.0

Diff since v9.63.0

Merged pull requests:

• CompatHelper: add new compat entry for PreallocationTools at version 0.4 for package docs, (keep existing compat) (#3388) (@github-actions[bot])
• fix: allow vector of parameters for split system of pure tunables (#3389) (@AayushSabharwal)

v9.63.0

ModelingToolkit v9.63.0

Diff since v9.62.0

Merged pull requests:

• fix: Add metadata getters (#3334) (@vyudu)
• Validate types of state variables in ODESystem/SDESystem construction (#3340) (@vyudu)
• feat: always build initialization problem (#3347) (@AayushSabharwal)
• fix: handle scalarized array parameters in initialization (#3358) (@AayushSabharwal)
• refactor: centralize all code generation (#3360) (@AayushSabharwal)
• refactor: format (#3361) (@AayushSabharwal)
• Add @description to @mtkmodel (#3363) (@hersle)
• feat: add assertions functionality (#3364) (@AayushSabharwal)
• latexify equation (#3366) (@owiecc)
• refactor: default create_bindings to false in build_function_wrapper (#3367) (@AayushSabharwal)
• Fix parameter array handling in ImperativeAffect (#3368) (@BenChung)
• Fix @mtkbuild macro keyword parsing (#3371) (@AayushSabharwal)
• Support specialization in ODEFunctionExpr (#3373) (@devmotion)
• test: update reference tests (#3377) (@AayushSabharwal)
• add documentation page for DiscreteSystem (#3380) (@vyudu)
• build: bump minor version (#3387) (@AayushSabharwal)

Closed issues:

• It is possible to designate non-default types for variables, which are then ignored. (#2589)
• Missing metadata getters (#2643)
• Initialization System fails to remove an equation (#3312)
• remake ignores u0 updates when the initialization is trivial (#3318)
• Operating point passed to linearization_function closure has no effect (#3319)
• MTK does not allow building initialization systems for un-simplified systems (#3330)
• Initialization warning when creating certain SDEProblems from Catalyst (but not corresponding ODEProblems (#3331)
• Remake does not handle dual type promotion with dummy derivatives (#3336)
• Cannot solve ODE with Duals if it has initialization and callback (#3342)
• Determined complete system raises an overdetermined warn when creating ODEProblem (#3349)
• Initialization and Linearization (#3352)
• ODESystem not dealing with dynamic parameters (#3365)
• The IfLifting pass doesn't seem to be working quite right (#3369)

v9.62.0

ModelingToolkit v9.62.0

Diff since v9.61.0

Merged pull requests:

• feat: add FMUComponent (#3282) (@AayushSabharwal)
• test: added tests for SDEFunctionExpr (#3313) (@sivasathyaseeelan)
• Handle events in isequal (#3317) (@vyudu)
• fix: fix add_fallbacks! (#3332) (@AayushSabharwal)
• Start using Moshi compatible syntax for singleton ADT variants (#3353) (@visr)

Closed issues:

• Systems are considered equal, even if they have different events (#2638)
• remake failing for several cases with vector variables (#2660)
• remake not working for Symbol + array variable (u0 only, p is fine) (#3281)
• ERROR: ArgumentError: ... is neither an observed nor an unknown variable (#3355)

v9.61.0

ModelingToolkit v9.61.0

Diff since v9.60.0

Merged pull requests:

• fix: @mtkmodel no longer makes @variables callable (#3217) (@AayushSabharwal)
• fix: properly handle rational functions in HomotopyContinuation (#3265) (@AayushSabharwal)
• feat: better inbounds handling and propagation for generated functions (#3277) (@AayushSabharwal)
• feat: add analysis points (#3285) (@AayushSabharwal)
• fix: minor bug fix in PolynomialTransformation (#3293) (@AayushSabharwal)
• Generic mechanism for debugging/logging functions (#3296) (@hersle)
• feat: allow remade initialization systems to be incomplete (#3301) (@AayushSabharwal)
• feat: add support for causal connections of variables (#3304) (@AayushSabharwal)
• DifferentialEquations -> OrdinaryDiffEqDefaults in readme (#3305) (@TorkelE)
• Add better SDESystem error when it hasn't been structural_simplify (#3316) (@vyudu)
• feat: support caching of different types of subexpressions in SCCNonlinearProblem (#3324) (@AayushSabharwal)
• fix: fix initialization of DiscreteProblem(::JumpSystem) (#3329) (@AayushSabharwal)
• test: added test for dae_order_lowering (#3335) (@sivasathyaseeelan)
• Fix dual type promotion in remake with dummy derivatives (#3337) (@hersle)
• fix: respect checkbounds in build_explicit_observed_function (#3339) (@AayushSabharwal)
• fix: only solve parameter initialization for NonlinearSystem (#3345) (@AayushSabharwal)

Closed issues:

• SDESystem's not displayed properly with latexify (#2069)
• Wonky interface for Dicrete/Continious callbacks when using ModelingToolkit (#2303)
• Events not working with automatic solver selection (#2741)
• Latexify display does not work for SDESystems (#2782)
• SDEs with brownian motions requires structural_simplify (#2831)
• Regression: Problem with indexing of JumpProblem (#2838)
• support connections between (causal) variables connect(var1, var2) (#3287)
• HomotopyContinuation extension of MTK breaks creation of ODESystems in PEtab.jl (#3291)
• Cannot remake problems if not providing full u0 and the first version did not require an initialization problem (#3295)
• Parameter initial conditions broken for DiscreteProblems (#3297)
• deepcopy leads to a failure in the initialization (#3307)
• Circular dependencies on Julia 1.10 (#3308)
• Dropped Variable (#3311)
• Some new problems with remake for Catalyst update (#3326)

v9.60.0

ModelingToolkit v9.60.0

Diff since v9.59.0

Merged pull requests:

• feat: create initialization systems for all problem types (#3253) (@AayushSabharwal)
• fix: fix timeseries_parameter_index for array symbolics (#3284) (@AayushSabharwal)
• build: bump OrdinaryDiffEqDefault compat (#3286) (@AayushSabharwal)
• build: bump minor version (#3289) (@AayushSabharwal)

v9.59.0

ModelingToolkit v9.59.0

Diff since v9.58.0

Merged pull requests:

• Add a mechanism to add additional passes to the end of structural simplification before completion. (#3201) (@BenChung)
• feat: enable structural_simplify(::SDESystem) (#3266) (@AayushSabharwal)
• CompatHelper: add new compat entry for Setfield at version 1 for package docs, (keep existing compat) (#3270) (@github-actions[bot])
• feat: add optional IfLifting simplification pass (#3272) (@AayushSabharwal)
• add option to include disturbance args in generate_control_function (#3273) (@baggepinnen)
• fix: retain system metadata when calling flatten (#3276) (@AayushSabharwal)

Closed issues:

• structural_simplify fails on SDESystem (#3261)
• Type instability with parameters that are abstract at system-level, but concrete at problem-level (#3262)
• Output of generate_control_function can no longer be traced symbolically (#3269)
• Unexpected behaviour of @register_symbolic (#3271)
• System metadata disapearing when calling complete on model with subsystems (#3275)