Dijkstra description.txt

for each pair of rectangles(you can consider index of rectangle while taking pair) for eg: 01, 02,......011,...1011:
	create a graph with all the vertices apart from source and destination, centre of source, centre of destination
	create edges by connecting all the rectangles corners with another rectangle corner ignoring the corners of source and destination
  	Add edges connecting source centr to destination centre, and add edges connecting source centr and destination centre to all other rectangle vertices
  	Now, we get all possible edges. Then, we need to select valid candidate edges.
	for each edges:
    		check (if this edge crosses any other rectangle edges(without source and destination rectangle edges):
    		if cross:
	 		discard.
		if does not cross: 
      			add the edge to graph edge list. 
      			compute distance which will be the weight of the edge
  
run dijaktra across these graph which will give distance from i to j



class Vertex里面存有坐标(x,y),到source的distance，visit flag，和neighbor edge vector。
class Edge里面存有起点和终点的坐标，以及weight。

申请了2个vector<Vertex>,一个是存放确定最短的顶点determined_vex，一个是待定candidate_vex。

初始化，把Vertex source的visit flag置为true，顶点push到determined_vex容器里。
1. 把Vertex source里的neighbor edge对应的vertex push到candidate_vex里面。并把visit flag置为true
2. 根据source的distance（此时=0），和edge的weight,计算candidate_vex里顶点到source的distance。并赋值给这些candidate_vex里的顶点。
3. 遍历candidate_vex容器，找到distance最短的顶点i，将顶点i加入到determined_vex，并且从candidate_vex删除掉顶点i。
开始循环：
4. 遍历Vertex i里的neighbor edge对应的vertex。如果vertex j没有visit，push到candidate_vex。push之后把visit flag置为true；如果vertex j有visit，比较vertex j.distance
   和vertex i.distance + edge weight的大小，前者大的话更新，小的话不变。
5. 遍历candidate_vex容器，找到distance最短的顶点k，将顶点k加入到determined_vex，并且从candidate_vex删除掉顶点k。
重复4~5直到找到destination.