makieorg / topoplots.jl Goto Github PK

Hi, I noticed a small problem when setting the 'colorrange' attribute for the eeg_topoplot() function. The colors inside the circles that indicate channel locations are unaffected by the new colorrange if the data is passed as Vector{Any} instead of Vector{Float32}.

This issue is used to trigger TagBot; feel free to unsubscribe.

If you haven't already, you should update your TagBot.yml to include issue comment triggers.
Please see this post on Discourse for instructions and more details.

If you'd like for me to do this for you, comment TagBot fix on this issue.
I'll open a PR within a few hours, please be patient!

I converted some MNE code to convert form x/y/z to eeg_topoplot coordinates.

Now the big question is, can/should this be included here in TopoPlots.jl, or somehwere else?

using CoordinateTransformations

	"""
		convert x/y/z electrode montage positions to spherical coordinate representation. output is a matrix
	"""
	function cart3d_to_spherical(x,y,z)
		sph = SphericalFromCartesian().(SVector.(x,y,z))
		sph = [vcat(s.r,s.θ,π/2 - s.ϕ) for s in sph] 
		sph = hcat(sph...)'
		return sph
	end

		
"""
Projects 3D electrode positions to a 2D layout.

Re-implementation of the MNE algorithm.

If you put sphere to 
"""
function pos3D_to_layout(x,y,z;sphere=[0,0,0.])
	#cart3d_to_spherical(x,y,z)
	
# translate to sphere origin
	x .-= sphere[1]
	y .-= sphere[2]
	z .-= sphere[3]
	
	# convert to spherical coordinates
	sph = cart3d_to_spherical(x,y,z)

	# get rid of of the radius for now
	pol_a = sph[:,3]
	pol_b = sph[:,2]
	
	# use only theta & phi, convert back to cartesian coordinates
	p_x = pol_a .* cos.(pol_b)
	p_y = pol_a .* sin.(pol_b)

	# scale by the radius
	p_x .*= sph[:,1] ./(π/2)
	p_y .*= sph[:,1] ./(π/2)

	# move back by the sphere coordinates
	p_x .+= sphere[1]
	p_y .+= sphere[2]

  return Point2f.(p_x,p_y)
end
		  		



	  

(for full reproducibility)

begin
using CSV
	using LinearAlgebra
	using HTTP
	using CoordinateTransformations
	using WGLMakie
using StaticArrays
using DataFrames
end
begin
pos3d = CSV.read(download("https://raw.githubusercontent.com/lrkrol/SEREEGA/b5dec01e8b98260904906cc19c38853bfbb2ff19/leadfield/hartmut/chanlocs-hartmut-NYhead_small.xyz"),DataFrame,header=false)
rename!(pos3d,:Column1=>:ix,:Column2=>:x,:Column3=>:y,:Column4=>:z,:Column5=>:label)

	# remove some whitespace in the labels
transform!(pos3d,:label=>(x->strip.(x))=>:label)

	# remove some weird channels
	subset!(pos3d,:label=>ByRow(l -> l ∉ ["Nk1","Nk2","Nk3","Nk4"]));

out =   pos3D_to_layout(pos3d.x,pos3d.y,pos3d.z)

eeg_topoplot(1:size(out,1),pos3d.label;positions=out, interpolation=NullInterpolator(),label_text=true,enlarge=0.5)
end;

checkout if we can now use
Interpolators.jl

maybe optimize the use of
ScatteredInterpolators.jl

Reading https://discourse.julialang.org/t/ann-pythoncall-and-juliacall/76778/24 and seeing the much better python package installation integration, it seems worth quite a bit to switch to PythonCall directly to call our one python dependency...

I didnt fully understand this line of code - do we need it?

It seems to be a stub to use the default values. params is not used though?

Context:

I got this message when trying to update makie to 0.19.0
It would be nice to change it somehow because in new Makie version they changed some functions.

I think we should allow
interpolation = False

This would allow to plot things like:

and

Should be straight forward to add another if interpolation == False catch where we select for Delauny/Splines Interpolations, right?

Hi,
I am plotting some EEG topographies together with @behinger. We observed strange interpolation at the top boundary. This is with the default interpolator ClaughTochter

DelaunyMesh has same issue

Any ideas? Has it to do with the fill-value?
Cheers,
Vladimir & Bene

Hello,
we noticed a weird behavior with the head position in a eeg_topoplot when the data points are asymmetrical or only in one corner.

In the image we are drawing two eeg topoplots with points in the top right and bottom left and our expected behavior was to see both heads overlap and contain all points instead of two smaller local heads.
This is the code used to draw them:

f = Figure()
axis = Axis(f[1, 1], aspect = 1)
xlims!(low = -2, high = 2)
ylims!(low = -2, high = 2)

data = [0, 0, 0]
pos1 = [Point2f(-1, -1), Point2f(-1.0, 0.0), Point2f(0, -1)]
pos2 = [Point2f(1, 1), Point2f(1.0, 0.0), Point2f(0, 1)]

eeg_topoplot!(axis, data, positions=pos1)
eeg_topoplot!(axis, data, positions=pos2)

Is there some workaround to keep the head location and size independent from the point locations ?
Or would it be possible to either add a way to set a position and size for the head or expose the draw_ear_nose! function ?

Thank you very much and kind regards,
Daniel, Sören, Niklas