Add user-facing Model predicates and restrict predicates to Model only

BREAKING.md

@@ -4,7 +4,7 @@
 
 This document describes breaking changes in CTModels releases and how to migrate your code.
 
-## [0.9.15] - 2026-04-18
+## [0.9.15-beta] - 2026-04-18
 
 ### Breaking Changes: Dual Dimension Function Renaming
 
@@ -32,39 +32,44 @@ The old function names were misleading because they returned the dimension of du
 
 The functions `dim_*_constraints_box(ocp::Model)` (for Model, not Solution) remain unchanged and still refer to constraint dimension in the model.
 
-### Non-Breaking Changes
+### Breaking Changes: PreModel Predicate Removal
 
-This release also introduces consistent variable and control checking functions without breaking existing functionality:
+The following predicate methods have been removed for `PreModel` and are now exclusive to `Model`:
 
-#### New Functions (Non-Breaking)
+- `is_autonomous(ocp::PreModel)` - removed
+- `is_variable(ocp::PreModel)` - removed
+- `is_control_free(ocp::PreModel)` - removed
 
-- **New functions**: Added `is_variable()` and `is_control_free()` for checking problem properties
-- **Dual methods**: Both functions have methods for `PreModel` and `Model` types
-- **Consistent API**: These functions follow the same pattern as `is_autonomous()`
-- **Runtime dimension checks**: Unlike time dependence (type-parameterized), variable and control use runtime dimension checks
+These methods remain available for `Model` instances.
 
-#### What Changed
+#### Migration Guide
 
 ```julia
-# New functions available (non-breaking)
-is_variable(ocp)        # Returns true if variable_dimension > 0
-is_control_free(ocp)   # Returns true if control_dimension == 0
-
-# Works for both PreModel and Model
-ocp = PreModel()
-state!(ocp, 2)
-control!(ocp, 1)
-variable!(ocp, 2)
-
-is_variable(ocp)        # Returns true
-is_control_free(ocp)    # Returns false
+# Before (PreModel access)
+pre = PreModel()
+state!(pre, 2)
+control!(pre, 1)
+variable!(pre, 2)
+time_dependence!(pre; autonomous=true)
+
+# These no longer work:
+is_autonomous(pre)    # MethodError
+is_variable(pre)      # MethodError
+is_control_free(pre)  # MethodError
+
+# After (use direct field access or internal predicates)
+pre.autonomous                      # true/false
+!CTModels.OCP.__is_variable_empty(pre)  # true/false
+CTModels.OCP.__is_control_empty(pre)    # true/false
 ```
 
-#### Migration
+#### Rationale
 
-- **No action required**: Existing code continues to work unchanged
-- **Optional enhancement**: Can use new functions for more readable code instead of inline dimension comparisons
-- **Same API**: No changes to existing user-facing API; behavior is fully backward compatible
+Predicate methods are now exclusive to immutable `Model` types to enforce a clear separation between mutable construction (`PreModel`) and immutable problem definition (`Model`). Internal predicates (`__is_*_empty`) are used for construction-time checks.
+
+#### Note
+
+The predicate methods `is_autonomous(model)`, `is_variable(model)`, and `is_control_free(model)` for `Model` remain unchanged and continue to work as before.
 
 ## [0.9.14] - 2026-04-12
 

CHANGELOG.md

@@ -7,7 +7,7 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
-## [0.9.15] - 2026-04-18
+## [0.9.15-beta] - 2026-04-18
 
 ### 🚀 Enhancements
 
@@ -60,37 +60,43 @@ expr = expression(pre.definition)  # Returns the Expr
 model = build(pre)  # Works even without definition
 ```
 
-#### Consistent Variable and Control Checking Functions
+#### User-Facing Model Predicates
 
-- **New functions**: Added `is_variable()` and `is_control_free()` for checking problem properties
-- **Dual methods**: Both functions have methods for `PreModel` and `Model` types
-- **Consistent API**: These functions follow the same pattern as `is_autonomous()`
-- **Runtime dimension checks**: Unlike time dependence (type-parameterized), variable and control use runtime dimension checks
-- **Display integration**: Updated display code to use the new functions instead of inline comparisons
+- **New predicates**: Added user-friendly predicate methods for `Model` instances
+- **Exclusive to Model**: Predicates are only available for immutable `Model`, not `PreModel`
+- **Consistent naming**: Follows pattern `has_*` for presence checks, `is_*` for property checks
+- **New exports**: `has_variable`, `has_control`, `has_abstract_definition`, `is_abstractly_defined`, `is_nonautonomous`, `is_nonvariable`
 
 #### API Enhancements
 
 ```julia
 # Check if problem has optimisation variables
-is_variable(ocp::PreModel)   # Returns true if variable_dimension > 0
-is_variable(ocp::Model)     # Returns true if variable_dimension > 0
+has_variable(model)       # Alias for is_variable(model)
+is_nonvariable(model)    # Opposite of is_variable(model)
 
-# Check if problem is control-free (no control input)
-is_control_free(ocp::PreModel)   # Returns true if control_dimension == 0
-is_control_free(ocp::Model)     # Returns true if control_dimension == 0
+# Check if problem has control input
+has_control(model)        # Opposite of is_control_free(model)
+
+# Check if problem has abstract definition
+has_abstract_definition(model)      # Checks if definition is non-empty
+is_abstractly_defined(model)        # Alias for has_abstract_definition
+
+# Check time dependence
+is_nonautonomous(model)             # Opposite of is_autonomous(model)
 ```
 
 ### 📊 API Changes
 
-- **New exports**: `is_variable` and `is_control_free` are now exported from CTModels
-- **Display code**: Internal display functions now use the new checking functions instead of inline dimension comparisons
+- **Breaking**: `is_variable(ocp::PreModel)`, `is_control_free(ocp::PreModel)`, `is_autonomous(ocp::PreModel)` removed
+- **New exports**: `has_variable`, `has_control`, `has_abstract_definition`, `is_abstractly_defined`, `is_nonautonomous`, `is_nonvariable`
+- **Display code**: Internal display functions use `__is_*_empty` predicates for PreModel, public predicates for Model
 
 ### 🔧 Internal Changes
 
-- **New methods**: Added `is_variable()` and `is_control_free()` methods in `time_dependence.jl` and `model.jl`
-- **Display refactoring**: Replaced inline `v_dim > 0` with `is_variable(ocp)` in `print.jl`
-- **Display refactoring**: Replaced inline `u_dim > 0` with `!is_control_free(ocp)` in `print.jl`
-- **New tests**: Added comprehensive test suite in `test_variable_control_checks.jl` with 20 tests
+- **Predicate refactoring**: Removed `__is_*_set` methods for `Model` (only `__is_*_empty` remains)
+- **PreModel access**: Display code uses direct field access (`ocp.autonomous`) and internal predicates (`__is_variable_empty`, `__is_control_empty`)
+- **Model access**: Public predicates (`is_variable`, `is_control_free`, `is_autonomous`) work for Model only
+- **Test updates**: Migrated tests to use internal predicates for PreModel, public predicates for Model
 
 ## [0.9.14] - 2026-04-12
 

Project.toml

@@ -1,6 +1,6 @@
 name = "CTModels"
 uuid = "34c4fa32-2049-4079-8329-de33c2a22e2d"
-version = "0.9.15"
+version = "0.10.0"
 authors = ["Olivier Cots <olivier.cots@toulouse-inp.fr>"]
 
 [deps]

src/CTModels.jl

@@ -91,4 +91,4 @@ using .Serialization
 include(joinpath(@__DIR__, "Init", "Init.jl"))
 using .Init
 
-end
+end

src/Display/Display.jl

@@ -39,7 +39,8 @@ import ..OCP: Model, PreModel, Solution, AbstractSolution
 import ..OCP: AbstractDefinition, Definition, EmptyDefinition
 
 # Import internal helpers from OCP for display
-import ..OCP: __is_empty, __is_definition_set, definition, __is_consistent
+import ..OCP: __is_empty, definition, __is_consistent
+import ..OCP: __is_variable_empty, __is_control_empty
 import ..OCP: state_dimension, control_dimension, variable_dimension
 import ..OCP: time_name, initial_time_name, final_time_name
 import ..OCP: dimension, name, state_name, control_name, variable_name

src/Display/pre_model.jl

@@ -38,9 +38,9 @@ function Base.show(io::IO, ::MIME"text/plain", ocp::PreModel)
         vi_names = components(ocp.variable)
 
         # dependencies
-        is_variable_dependent = is_variable(ocp)
-        is_time_dependent = !is_autonomous(ocp)
-        is_control_free_ocp = is_control_free(ocp)
+        is_variable_dependent = v_dim > 0
+        is_time_dependent = !ocp.autonomous
+        is_control_free_ocp = u_dim == 0
 
         # cost
         has_a_lagrange_cost = has_lagrange_cost(ocp.objective)

src/OCP/Building/model.jl

@@ -493,7 +493,7 @@ function build(pre_ocp::PreModel; build_examodel=nothing)::Model
     objective = pre_ocp.objective
     constraints = build(pre_ocp.constraints)
     definition = pre_ocp.definition
-    TD = is_autonomous(pre_ocp) ? Autonomous : NonAutonomous
+    TD = pre_ocp.autonomous ? Autonomous : NonAutonomous
 
     # create the model
     model = Model{TD}(
@@ -619,6 +619,114 @@ function is_control_free(ocp::Model)::Bool
     return control_dimension(ocp) == 0
 end
 
+"""
+$(TYPEDSIGNATURES)
+
+Check whether the problem has optimisation variables.
+
+# Arguments
+- `ocp::Model`: The optimal control problem model.
+
+# Returns
+- `Bool`: `true` if the problem has optimisation variables (variable dimension > 0), `false` otherwise.
+
+# Example
+```julia-repl
+julia> has_variable(model)  # returns true if variables are present
+```
+"""
+has_variable(ocp::Model)::Bool = is_variable(ocp)
+
+"""
+$(TYPEDSIGNATURES)
+
+Check whether the problem has control input.
+
+# Arguments
+- `ocp::Model`: The optimal control problem model.
+
+# Returns
+- `Bool`: `true` if the problem has control input (control dimension > 0), `false` otherwise.
+
+# Example
+```julia-repl
+julia> has_control(model)  # returns true if control is present
+```
+"""
+has_control(ocp::Model)::Bool = !is_control_free(ocp)
+
+"""
+$(TYPEDSIGNATURES)
+
+Check whether the problem has an abstract definition.
+
+# Arguments
+- `ocp::Model`: The optimal control problem model.
+
+# Returns
+- `Bool`: `true` if the model has a non-empty abstract definition, `false` otherwise.
+
+# Example
+```julia-repl
+julia> has_abstract_definition(model)  # returns true if definition was attached
+```
+"""
+has_abstract_definition(ocp::Model)::Bool = !__is_definition_empty(definition(ocp))
+
+"""
+$(TYPEDSIGNATURES)
+
+Check whether the problem is abstractly defined.
+
+# Arguments
+- `ocp::Model`: The optimal control problem model.
+
+# Returns
+- `Bool`: `true` if the model has a non-empty abstract definition, `false` otherwise.
+
+# Example
+```julia-repl
+julia> is_abstractly_defined(model)  # returns true if definition was attached
+```
+"""
+is_abstractly_defined(ocp::Model)::Bool = has_abstract_definition(ocp)
+
+"""
+$(TYPEDSIGNATURES)
+
+Check whether the problem is non-autonomous (time-dependent).
+
+# Arguments
+- `ocp::Model`: The optimal control problem model.
+
+# Returns
+- `Bool`: `true` if the system is non-autonomous (time-dependent), `false` otherwise.
+
+# Example
+```julia-repl
+julia> is_nonautonomous(model)  # returns true if time-dependent
+```
+"""
+is_nonautonomous(ocp::Model)::Bool = !is_autonomous(ocp)
+
+"""
+$(TYPEDSIGNATURES)
+
+Check whether the problem has no optimisation variables.
+
+# Arguments
+- `ocp::Model`: The optimal control problem model.
+
+# Returns
+- `Bool`: `true` if the problem has no optimisation variables (variable dimension == 0), `false` otherwise.
+
+# Example
+```julia-repl
+julia> is_nonvariable(model)  # returns true if no variables
+```
+"""
+is_nonvariable(ocp::Model)::Bool = !is_variable(ocp)
+
 # State
 """
 $(TYPEDSIGNATURES)

src/OCP/Components/constraints.jl

@@ -319,7 +319,7 @@ function constraint!(
 
     # checks: control must be set for :control constraint type
     if type == :control
-        @ensure __is_control_set(ocp) Exceptions.PreconditionError(
+        @ensure !__is_control_empty(ocp) Exceptions.PreconditionError(
             "Control must be set for type=:control constraints",
             reason="control has not been defined yet but constraint type requires it",
             suggestion="Call control!(ocp, dimension) before adding :control constraints, or use a different constraint type",
@@ -328,7 +328,7 @@ function constraint!(
     end
 
     # checks: variable must be set if using type=:variable
-    @ensure (type != :variable || __is_variable_set(ocp)) Exceptions.PreconditionError(
+    @ensure (type != :variable || !__is_variable_empty(ocp)) Exceptions.PreconditionError(
         "Variable must be set for type=:variable constraints",
         reason="OCP has no variable defined but constraint type requires it",
         suggestion="Call variable!(ocp, dimension) before adding variable constraints, or use a different constraint type",

src/OCP/Components/control.jl

@@ -59,7 +59,7 @@ function control!(
 )::Nothing where {T1<:Union{String,Symbol},T2<:Union{String,Symbol}}
 
     # checks using @ensure
-    @ensure !__is_control_set(ocp) Exceptions.PreconditionError(
+    @ensure __is_control_empty(ocp) Exceptions.PreconditionError(
         "Control already set",
         reason="control has already been defined for this OCP",
         suggestion="Create a new OCP instance or use the existing control definition",

src/OCP/Components/times.jl

@@ -56,14 +56,14 @@ function time!(
         context="time! function - duplicate definition check",
     )
 
-    @ensure __is_variable_set(ocp) || (isnothing(ind0) && isnothing(indf)) Exceptions.PreconditionError(
+    @ensure !__is_variable_empty(ocp) || (isnothing(ind0) && isnothing(indf)) Exceptions.PreconditionError(
         "Variable must be set for free time",
         reason="variable is required when t0 or tf is free (ind0/indf provided)",
         suggestion="Call variable!(ocp, dimension) before time! with free time parameters, or use fixed times (t0, tf)",
         context="time! function - free time validation",
     )
 
-    if __is_variable_set(ocp)
+    if !__is_variable_empty(ocp)
         q = dimension(ocp.variable)
 
         @ensure isnothing(ind0) || (1 ≤ ind0 ≤ q) Exceptions.IncorrectArgument(

src/OCP/Components/variable.jl

@@ -39,7 +39,7 @@ function variable!(
     name::T1=__variable_name(q),
     components_names::Vector{T2}=__variable_components(q, string(name)),
 )::Nothing where {T1<:Union{String,Symbol},T2<:Union{String,Symbol}}
-    @ensure !__is_variable_set(ocp) Exceptions.PreconditionError(
+    @ensure __is_variable_empty(ocp) Exceptions.PreconditionError(
         "Variable already set",
         reason="variable has already been defined for this OCP",
         suggestion="Create a new OCP instance or use the existing variable definition",