Plotting a graph - Different line styles for different edges - matlab

Question

I have a graph of nodes and edges that I am plotting against (x, y, z) coordinates for each node.
I would like to use a dotted line for the "inner" edges, and a straight line for the "outer" edges.
LineStyle may help me, but I think that will make all of my edges dotted lines. I would like to specify which edges are dotted.
Here is the code I am using to plot the graph: (G is a graph)
plot(G, 'XData', A(1,:), 'YData', A(2,:), 'ZData', A(3,:))
Any help is much appreciated!
I have a function to create the graph, it gives this output:
G =
graph with properties:
Edges: [11×1 table]
Nodes: [6×0 table]
And a coordinate system similar to this:
A =
0 1 0 -1 0 0
0 1 3 2 1 2
1 1 1 1 1 1
Using the plot function above, this provides this output:

Answer 1

Ok you didn't really put enough effort in making your question easy to answer for everybody but I was curious about a method and it seem to work so I'll post it anyway.
The classic trick for these cases where a sub-group of points/lines have to be plotted differently than the whole group is to:
Plot the whole group
Plot the sub-group on top with different line properties
In your case, it means you have to find the points and edges which are part of the boundary. I'll explain one method below. Be aware that this method will only work for convex boundaries (see the difference with a concave set where the method wouldn't work: concave vs. convex hull ).
First I need to get to the state of your question, so let's consider:
%% Initial data
A =[ 0 1 0 -1 0 0
0 1 3 2 1 2
1 1 1 1 1 1 ];
G = graph ;
G = addedge(G,[1 1 1 1 1 2 2 2 3 4 4],[2 3 4 5 6 3 5 6 4 5 6]) ;
plot(G, 'XData', A(1,:), 'YData', A(2,:) , 'LineStyle','--')
grid on ; hold on
np = size(A,2) ; % Number of points/nodes
This will produce a figure exactly like yours, with only one linestyle, which was set to dashed lines for the full graph. The exercise is now to plot a contour with a solid line.
To find the contour (at least in 2D cases like yours), the method used is:
find one point of the contour (initial point), the lowest x
possible.
calculate the angle between unit vector ux=[1,0,0] and
all the other vectors (formed by the initial point and all the other
points of the matrix).
Assign the next point of the contour (it is the one which vector has
the minimum angle) . You now have a current vector vec (between
point 1 and 2), and a current point pt2.
Calculate the angle between vec and all the other vectors (formed
by the current point pt2 and all the other points of the matrix)
the next point of the contour is the one which vector has the
minimum angle.
Repeat (3) and (4) until the next point of the contour is the
starting point.
Translated in code:
%% find a point on the edge
% (I chose to start with the lowest point on Y axis)
[~,idxNext] = min(A(2,:)) ;
%% initialise counters
isOnEdge = false(np,1) ; % this will hold a logical register of contour points
idxEdge = idxNext ; % this will hold the points of the contour, in order
%% setup start conditions
isOnEdge(idxNext) = true ; % initial point
p = A(:,idxNext) ; % initial point
v = [1;0;0] ; % initial vector (start with unit vector oriented Ox)
%% search for contour
isRunning = true ;
iter = 0 ;
while isRunning
iter = iter + 1 ; % make sure we're not stuck in infinite loop
angs = find_angles(v,p,A) ; % find angles between initial vector and all other points
angs(idxNext) = Inf ; % Exclude current point
if numel(idxEdge) > 1 % Exclude last point (if not at first iteration)
angs(idxEdge(end-1)) = Inf ;
end
[~,idxNext] = min(angs) ; % find the index of the minimum angle
if isOnEdge(idxNext)
% we've completed the close profile, bail out
isRunning = false ;
else
% move on to next point/vector
idxEdge = [idxEdge idxNext] ; %#ok<AGROW>
isOnEdge(idxNext) = true ;
p = A(:,idxNext) ;
v = A(:,idxNext) - A(:,idxEdge(end-1)) ;
end
if iter > np
break % make sure we're not stuck in infinite loop
end
end
%% present results
if isRunning
fprintf('Could''t find a closed profile\n')
else
fprintf('Found points defining a closed profile:\n')
disp(idxEdge)
end
%% Plot on top of graph
% create a matrix conataining only the contour points, close the contour by
% replicating the first point in last position
C = [A(:,idxEdge) , A(:,idxEdge(1))] ;
% plot
plot3( C(1,:) , C(2,:) , C(3,:) ,'b', 'LineWidth',2)
view(2)
Which will yield:
In the script above I am using the function find_angles.m. The code for it:
function angs = find_angles(Uin,p,M)
% find angle between Uin vector and the vectors formed between p and
% all the points in M
np = size(M,2) ;
angs = zeros(np,1) ;
for iv=1:np
vec = M(:,iv) - p ;
c = cross(Uin,vec) ;
d = dot(Uin,vec) ;
angs(iv) = rad2deg( atan2( norm(c) , d ) ) * sign( c(3) ) ;
% take care of an edge case
if c(3)==0 && d==-1 ; angs(iv) = 180 ; end
% give [0 360] degree results
if angs(iv) < 0 ; angs(iv) = 180 + (180+angs(iv) ) ; end
end