Reimplementcurrent polar plotting scripts with Makie scripts

[jellepoland]

[jellepoland]

Example:

"""
Custom Makie-based polar plotting functions for VortexStepMethod results.
This provides a temporary replacement for PyPlot-based plot_polars during the Makie migration.
"""

using GLMakie
using VortexStepMethod

"""
    generate_polar_data(solver, body_aero, angle_range; 
                        angle_type="angle_of_attack",
                        angle_of_attack=0.0,
                        side_slip=0.0,
                        v_a=10.0)

Generate polar data by sweeping through angle range and solving for each configuration.

# Returns
- `polar_data`: Dictionary with keys :alpha, :CL, :CD, :CS (side force coefficient)
- `rey`: Reynolds number
"""
function generate_polar_data(solver, body_aero, angle_range;
                            angle_type="angle_of_attack",
                            angle_of_attack=0.0,
                            side_slip=0.0,
                            v_a=10.0)
    
    CL_list = Float64[]
    CD_list = Float64[]
    CS_list = Float64[]
    alpha_list = Float64[]
    
    for angle in angle_range
        # Set angle based on type
        if angle_type == "angle_of_attack"
            alpha = angle
            beta = side_slip
        elseif angle_type == "side_slip"
            alpha = angle_of_attack
            beta = angle
        else
            error("angle_type must be 'angle_of_attack' or 'side_slip'")
        end
        
        # Set flight conditions
        VortexStepMethod.set_va!(body_aero, v_a, alpha, beta, 0.0)
        
        # Solve
        results = VortexStepMethod.solve(solver, body_aero; log=false)
        
        # Extract coefficients
        push!(alpha_list, alpha)
        push!(CL_list, results.CL)
        push!(CD_list, results.CD)
        push!(CS_list, results.CS)
    end
    
    # Calculate Reynolds number (using first wing panel as reference)
    wing = body_aero.bodies[1]
    chord = wing.chord_root  # Use root chord as reference
    rey = v_a * chord / 1.5e-5  # kinematic viscosity of air at sea level
    
    polar_data = Dict(
        :alpha => alpha_list,
        :CL => CL_list,
        :CD => CD_list,
        :CS => CS_list
    )
    
    return polar_data, rey
end

"""
    plot_polars_makie(solver_list, body_aero_list, label_list;
                      angle_range=range(0, 20, 21),
                      angle_type="angle_of_attack",
                      angle_of_attack=0.0,
                      side_slip=0.0,
                      v_a=10.0,
                      title="polar",
                      size=(1200, 800))

Plot polar curves (CL, CD, CS vs angle) using Makie.

# Arguments
- `solver_list`: Vector of VortexStepMethod.Solver objects
- `body_aero_list`: Vector of VortexStepMethod.BodyAerodynamics objects
- `label_list`: Vector of String labels for each case

# Keyword Arguments
- `angle_range`: Range of angles to sweep (default: 0° to 20° in 21 steps)
- `angle_type`: "angle_of_attack" or "side_slip"
- `angle_of_attack`: Fixed angle of attack when sweeping sideslip (default: 0.0°)
- `side_slip`: Fixed sideslip when sweeping angle of attack (default: 0.0°)
- `v_a`: Airspeed in m/s (default: 10.0)
- `title`: Plot title (default: "polar")
- `size`: Figure size in pixels (default: (1200, 800))

# Returns
- `fig`: Makie Figure object
"""
function plot_polars_makie(solver_list, body_aero_list, label_list;
                          angle_range=range(0, 20, 21),
                          angle_type="angle_of_attack",
                          angle_of_attack=0.0,
                          side_slip=0.0,
                          v_a=10.0,
                          title="polar",
                          size=(1200, 800))
    
    # Validate inputs
    if length(solver_list) != length(body_aero_list) || length(solver_list) != length(label_list)
        throw(ArgumentError("Mismatch in number of solvers ($(length(solver_list))), " *
                          "body_aero ($(length(body_aero_list))), and labels ($(length(label_list)))"))
    end
    
    # Generate polar data for all cases
    polar_data_list = []
    labels_with_re = String[]
    
    for (i, (solver, body_aero)) in enumerate(zip(solver_list, body_aero_list))
        polar_data, rey = generate_polar_data(
            solver, body_aero, angle_range;
            angle_type,
            angle_of_attack,
            side_slip,
            v_a
        )
        push!(polar_data_list, polar_data)
        # Update label with Reynolds number
        label_with_re = "$(label_list[i]) Re = $(round(Int64, rey*1e-5))e5"
        push!(labels_with_re, label_with_re)
    end
    
    # Create figure with 3 subplots
    fig = Figure(size=size)
    
    # Determine x-axis label
    x_label = angle_type == "angle_of_attack" ? "Angle of Attack [deg]" : "Sideslip Angle [deg]"
    
    # Plot CL vs angle
    ax1 = Axis(fig[1, 1], xlabel=x_label, ylabel="CL [-]", 
               title="Lift Coefficient")
    
    # Plot CD vs angle
    ax2 = Axis(fig[1, 2], xlabel=x_label, ylabel="CD [-]", 
               title="Drag Coefficient")
    
    # Plot CS vs angle
    ax3 = Axis(fig[2, 1], xlabel=x_label, ylabel="CS [-]", 
               title="Side Force Coefficient")
    
    # Plot CL vs CD (polar)
    ax4 = Axis(fig[2, 2], xlabel="CD [-]", ylabel="CL [-]", 
               title="Lift-Drag Polar")
    
    # Color palette
    colors = Makie.wong_colors()
    
    # Plot data for each case
    for (i, (polar_data, label)) in enumerate(zip(polar_data_list, labels_with_re))
        color = colors[mod1(i, length(colors))]
        
        # CL vs angle
        lines!(ax1, polar_data[:alpha], polar_data[:CL], 
               label=label, color=color, linewidth=2)
        scatter!(ax1, polar_data[:alpha], polar_data[:CL], 
                color=color, markersize=8)
        
        # CD vs angle
        lines!(ax2, polar_data[:alpha], polar_data[:CD], 
               label=label, color=color, linewidth=2)
        scatter!(ax2, polar_data[:alpha], polar_data[:CD], 
                color=color, markersize=8)
        
        # CS vs angle
        lines!(ax3, polar_data[:alpha], polar_data[:CS], 
               label=label, color=color, linewidth=2)
        scatter!(ax3, polar_data[:alpha], polar_data[:CS], 
                color=color, markersize=8)
        
        # CL vs CD polar
        lines!(ax4, polar_data[:CD], polar_data[:CL], 
               label=label, color=color, linewidth=2)
        scatter!(ax4, polar_data[:CD], polar_data[:CL], 
                color=color, markersize=8)
    end
    
    # Add legends
    axislegend(ax1, position=:lt)
    
    # Add overall title
    Label(fig[0, :], title, fontsize=20, font=:bold)
    
    return fig
end

"""
    plot_polars_makie(solver, body_aero, label; kwargs...)

Convenience method for single solver/body_aero case.
"""
function plot_polars_makie(solver, body_aero, label; kwargs...)
    return plot_polars_makie([solver], [body_aero], [label]; kwargs...)
end

# Export the main function
export plot_polars_makie, generate_polar_data

[1-Bart-1]

I don't think this is a bug?

[1-Bart-1]

But please add a PR. I wouldn't say "replace current", but "add Makie alternative", because of how extensions work you can keep both.

[ufechner7]

This is one more reason to do the 3D plotting in KiteViewers.jl and not in this package. From my point of view 2D plotting should be done using ControlPlots.jl. Then there would be no reason to reimplement this feature.

[1-Bart-1]

I would really like to have the option to 2D plot with Makie, so I don't have to use multiple plotting libraries. I want to be able to plot both in 2D and 3D in the same Julia session. And Makie is a lot easier to setup, while if you want vector graphics at some point you can always use the CairoMakie backend, or the existing ControlPlots plotting functions.

[ufechner7]

Your wrote: "I want to be able to plot both in 2D and 3D in the same Julia session. "

This is already possible. I do that in KiteControllers and in KiteSimulators. It might need some adaptations in this package, I don't understand why that currently shouldn't work. To achieve this goal, 2D plotting with Makie is not needed.

[1-Bart-1]

So you use ControlPlots and Makie at the same time?

[ufechner7]

Yes, of course. Perhaps you have to change:

using GLMakie

to

import GLMakie as ma

to avoid conflicts.

If I work in the package KiteViewers, the following commands work nicely:

julia> using ControlPlots
julia> import GLMakie as ma
julia> ma.plot(rand(3))
julia> plot(rand(3))

I get two different plots: from ControlPlots, I get a line plot, and from Makie, a scatter plot.

But this is without any package extensions; I don't know if that makes a difference.