/// <summary>
/// http://alrightchiu.github.io/SecondRound/graph-li-yong-dfsxun-zhao-dagde-topological-sorttuo-pu-pai-xu.html
/// </summary>
public static void TopologicalSort(int start, Graph <int> graph)
{
int num_vertex = graph.VerticesCount;
// 第一次DFS(), 目的是取得finish[]
DFS_Algo dfs = new DFS_Algo();
dfs.DFS(start, graph);
// 顯示 第一次DFS()後的finish[]
Console.WriteLine("First DFS() on G, finish time:");
dfs.finish.Print("finish");
// 矩陣 finishLargetoSmall[] 用來儲存 finish[] 由大至小的vertex順序
int[] finishLargetoSmall = new int[num_vertex];
for (int i = 0; i < num_vertex; i++)
{
finishLargetoSmall[i] = i;
}
// QuickSort()會更新 finishLargetoSmall[] 成 finish[] 由大至小的vertex順序
Sort_QuickSort_Graph.QuickSort(dfs.finish, 0, num_vertex - 1, finishLargetoSmall);
// 列印出 finish[] 由大至小的vertex順序
Console.WriteLine("Topological Sort:");
Console.WriteLine("finish time Large to Small:");
finishLargetoSmall.Print();
}
public static void main()
{
DFS_Algo dfs = new DFS_Algo();
var vertices = new[] { 0, 1, 2, 3, 4, 5, 6, 7 };
var edges = new[] { Tuple.Create(0, 1), Tuple.Create(0, 2), Tuple.Create(1, 3), Tuple.Create(2, 1), Tuple.Create(2, 5), Tuple.Create(3, 4), Tuple.Create(3, 5), Tuple.Create(5, 1), Tuple.Create(6, 4), Tuple.Create(6, 7), Tuple.Create(7, 6) };
var graph = new Graph <int>(vertices, edges, GraphOption.Diirected);
dfs.DFS(0, graph);
dfs.PrintInfo();
//想了很久總算可能想出為什麼 跑的結果 和 文章不一樣
//應該是因為 BFS的時候 例子是 無向圖 undirected graph
// graph的Add Edge方法也是針對無向圖設計的
// 所以要再增加能接受有向圖的機制
// 已加 推測沒錯 加了就成功了
}
/// <summary>
/// http://alrightchiu.github.io/SecondRound/graph-li-yong-dfsxun-zhao-strongly-connected-componentscc.html#ref
/// </summary>
public static void PrintSCCs(int start, Graph <int> graph)
{
int num_vertex = graph.VerticesCount;
// 第一次DFS(), 目的是取得finish[]
DFS_Algo dfs = new DFS_Algo();
dfs.DFS(start, graph);
// 顯示 第一次DFS()後的finish[]
Console.WriteLine("First DFS() on G, finish time:");
dfs.finish.Print("finish");
// 矩陣 finishLargetoSmall[] 用來儲存 finish[] 由大至小的vertex順序
int[] finishLargetoSmall = new int[num_vertex];
for (int i = 0; i < num_vertex; i++)
{
finishLargetoSmall[i] = i;
}
// QuickSort()會更新 finishLargetoSmall[] 成 finish[] 由大至小的vertex順序
Sort_QuickSort_Graph.QuickSort(dfs.finish, 0, num_vertex - 1, finishLargetoSmall);
// 列印出 finish[] 由大至小的vertex順序
Console.WriteLine("finish time Large to Small:");
finishLargetoSmall.Print();
// gT代表Transpose of Graph
var gT = graph.GetTranspose();
// 第二次DFS(), 執行在gT
DFS_Algo dfsGT = new DFS_Algo();
dfsGT.DFS_WithOrder(start, gT, finishLargetoSmall);
// 顯示 第二次DFS()後的finish[]
Console.WriteLine("Second DFS() on gT, finish time:\n");
dfsGT.finish.Print("finish");
// 顯示 第二次DFS()後的predecessor[]
Console.WriteLine("predecessor[] before SetCollapsing:\n");
dfsGT.predecessor.Print("predecessor");
dfsGT.SetCollapsing();
//這個有點沒品 他竟然沒有附 SetCollapsing 的程式碼
//http://alrightchiu.github.io/SecondRound/graph-li-yong-dfshe-bfsxun-zhao-connected-component.html
//原文是有圖說解釋得很詳細 簡單來說 把一個圖的所有點 改變成直接連結到Root
//看來是出作業自己寫
//不過找到這篇有類似的程式碼 http://alrightchiu.github.io/SecondRound/setyi-arraybiao-shi.html
//來參考看看
//實際想想看，這邊使用的機制 就是 Adj List 所以就要用AdjList的概念做法來做到
//那就主要是 刪除邊 增加邊 還有把predecessor設定為root
//predecessor隔天起來就忘記 他也就是上一層的父節點 就是DFS 一點走到下個點 一點就是下點的pred
//所以就是一個點對一個pred 也就是dfs類中的那個array啦
// 顯示 SetCollapsing後的predecessor[]
Console.WriteLine("predecessor after SetCollapsing:\n");
dfsGT.predecessor.Print("predecessor");
// 如同在undirected graph中尋找connected component
int num_cc = 0;
for (int i = 0; i < num_vertex; i++)
{
if (dfsGT.GetPredecessor(i) < 0)
{
Console.WriteLine($"SCC#{++num_cc}: {i} ");
for (int j = 0; j < num_vertex; j++)
{
if (dfsGT.GetPredecessor(j) == i)
{
Console.WriteLine(j + " ");
}
}
}
}
}
