The Code Template

Here is the standard recursive DFS template you will use constantly:

visited = [false] * (n + 1)
adj = adjacency list of the graph

function dfs(u):
    visited[u] = true
    // process u here
    for each v in adj[u]:
        if not visited[v]:
            dfs(v)

// in main:
// build adj from input
// call dfs(start_vertex)

Declare a global visited array. Write a function dfs that takes a vertex $u$. Inside: mark $u$ visited, process $u$, loop through neighbors of $u$, and recurse on unvisited neighbors.

This template works for nearly every DFS problem. You will modify what happens when you process a vertex, but the structure stays the same.