Convert to junction graph to vertices and faces.

The goal is to be able to plot the tissue with one face per cell, to
have a simplified representation of the cells on the tissue with e.g.:
>> patch( 'Faces', F, 'Vertices', V, 'FaceColor', 'g' )

src/@deproj/deproj.m

@@ -26,6 +26,11 @@ classdef deproj
         % Exports results to a spreadsheet file.
         to_file( obj, file_name, include_header )
         
+        %% Conversion between MATLAB objects.
+        
+        % Returns the faces and vertices of a deproj collection.
+        [ V, F ] = to_vertices_and_faces( obj )
+        
         %% Generate figures.
         
         % Figure with the local plan orientation for a collection of epicells.

src/@deproj/to_vertices_and_faces.m

+function [ V, F ] = to_vertices_and_faces( obj )
+%TO_FACES Returns the faces and vertices of a deproj collection.
+% The vertices are made of the junction centroid coordinates.
+% Each face is one epicell, enclosed by the junctions around it.
+
+    % Collect vertices.
+    g = obj.junction_graph;
+    V = g.Nodes.Centroid;
+
+    % Collected junction ids.
+    ids = { obj.epicells.junction_ids }';
+    n_vertices = cellfun(@numel, ids );
+    max_n_vertices = max( n_vertices );
+    n_obj = numel( obj.epicells );
+
+    % Write into the face matrix.
+    F = NaN( n_obj, max_n_vertices );
+    for i = 1 : n_obj
+        
+        jids = ids{ i };
+        P = V( jids, : );
+        
+        % Sort the points so that we don't have intersecting polygons.
+        [ ~, I ] = deproj.sort_polygon( P );
+        F( i, 1 : n_vertices(i) ) = jids( I );
+    end
+
+end
+