Getting StackOverflowError in DFS implementation for very large input

While practising DFS problems on HackerEarth using Java, one of the test cases results in NZEC error when checking it out, it displays StackOverFlow error. I checked it many of the times but all of the test cases pass except one which results in an error, and this is not for that specific problem but for many of the problems(DFS problems) one test cases is always there out of 10 or 20 which always result in NZEC when solving with java.

I don't whether it is due to my Adjacency list implementation or something else. Here is the problem one of the problem.

Problem : https://www.hackerearth.com/practice/algorithms/graphs/depth-first-search/practice-problems/algorithm/feasible-relations/

My code:

import java.io.BufferedReader;
import java.io.IOException;
import java.io.InputStreamReader;
import java.lang.reflect.Array;
import java.util.*;
public class TestClass {
static long mod = 1000000007;
static String[] temp;
static  int[] node;
static  int[] vis;
public static final BufferedReader br = new BufferedReader(new InputStreamReader(System.in));
public static void main (String[] args) throws IOException {
int testCases= Integer.parseInt(br.readLine().split(" ")[0]);
while (testCases > 0) {
temp = br.readLine().split(" ");
int N = Integer.parseInt(temp[0]);
int K = Integer.parseInt(temp[1]);
HashMap<Integer, ArrayList<Integer>> forEqual = new HashMap<>();
ArrayList<int[]> edgeList = new ArrayList<>();
for (int i = 0; i < K; i++) {
temp = br.readLine().split(" ");
int u = Integer.parseInt(temp[0]);
int v = Integer.parseInt(temp[2]);
if (temp[1].equals("=")) {
forEqual.computeIfAbsent(u, k -> new ArrayList<>()).add(v);
forEqual.computeIfAbsent(v, k -> new ArrayList<>()).add(u);
} else {
edgeList.add(new int[]{u, v});
}
}
boolean[] vis = new boolean[1000005];
node = new int[1000005];
int c = 0;
for (int i : forEqual.keySet()) {
if (!vis[i]) {
c++;
DFS(i, forEqual, c, vis);
}
}
boolean flag = true;
for(int[] i : edgeList){
if(node[i[0]] == node[i[1]]){
flag = false;
break;
}
}
System.out.println(flag ? "YES" : "NO");
testCases--;
}
}
private static void DFS(int i, HashMap<Integer, ArrayList<Integer>> forEqual, int c, boolean[] vis) {
vis[i] = true;
node[i] = c;
for (int v : forEqual.get(i)) {
if (!vis[v]) {
DFS(v, forEqual, c, vis);
}
}
}
}

Error Log

No one answer to my question, this question is not from the live contest, it takes me a week to figure out the solution for my problem.
The problem of StackOverFlow is due to the limitation of the number of stack calls on HackerEarth IDE and similar online IDE for Java which is 100,000 calls are only allowed, to avoid such, create a thread object and pass stack size as a parameter manually, which will increase the stack size.

class Solution{
public static void main (String[] args) throws IOException {
new Thread(null, new TestClass(), "", 1 << 26).start();
}
public void run() {
// DO WHATEVER YOU WANT 
}
}