[General] Control logic

[ocots]

Since the control is piecewise constant, use this and step for plot.

[ocots]

Add also in the build solution function some assert to check if the size of the state, control and costate are N or N-1, where N is the size of the grid. For the state, it must be N I think.

[ocots]

Control Logic - Validation & Visualization

Issue: #207 - Control logic
Date: 2025-12-17
Status: Planning Complete ✅

TL;DR

Add defensive validation in build_solution to ensure state/control/costate dimensions match the time grid (either N or N-1), and enforce steppre visualization for controls.

---

1. Analysis

Current Behavior

• build_solution automatically slices T to match the data size (T[1:size(data,1)]).
• Risk: It implicitly accepts arbitrary sizes (e.g., data covering only half the time grid), which is likely a bug in user code.

Requirement

1. Validation: Enforce that size(X,1), size(U,1), size(P,1) are either length(T) or length(T)-1.
2. Visualization: Always use steppre for controls in ext/plot.jl.

---

2. Technical Design

Solution Building (src/ocp/solution.jl)

Insert checks before interpolation:

dim_t = length(T)
N = size(X, 1)
M = size(U, 1)
L = size(P, 1)

@ensure N == dim_t || N == dim_t - 1 "State dimension mismatch"
@ensure M == dim_t || M == dim_t - 1 "Control dimension mismatch"
@ensure L == dim_t || L == dim_t - 1 "Costate dimension mismatch"

Plotting (ext/plot.jl)

Update __plot recipe:

# For controls
seriestype := :steppre

---

3. Tasks

Task	Description
T1.1	Add dimension validation in build_solution.
T1.2	Update ext/plot.jl to use :steppre for controls.

---

4. Test Commands

julia --project=. -e 'using Pkg; Pkg.test("CTModels"; test_args=["solution"]);'