DirectoryScanner.java

import java.io.File;
import java.util.Arrays;
import java.util.HashSet;
import java.util.List;
import java.util.Set;

public class DirectoryScanner {

    // Basic filter for system files/folders to ignore
    private static final Set<String> IGNORED_NAMES = new HashSet<>(Arrays.asList(
            "Windows", "Program Files", "Program Files (x86)", ".git", ".idea", "node_modules", "$Recycle.Bin",
            "System Volume Information"));

    private PatternMatcher patternMatcher;
    private InfectionReport report;

    public DirectoryScanner(PatternMatcher patternMatcher, InfectionReport report) {
        this.patternMatcher = patternMatcher;
        this.report = report;
    }

    public void scanDirectory(File directory) {
        if (directory == null || !directory.exists()) {
            System.out.println("[-] Invalid directory.");
            return;
        }

        if (IGNORED_NAMES.contains(directory.getName())) {
            // System.out.println("[*] Skipping system/ignored directory: " +
            // directory.getName());
            return;
        }

        File[] files = directory.listFiles();
        if (files != null) {
            for (File file : files) {
                if (file.isDirectory()) {
                    scanDirectory(file); // Recursive call
                } else {
                    scanFile(file);
                }
            }
        }
    }

    private void scanFile(File file) {
        // Skip binary files or massive files if needed (simplified check)
        if (file.length() > 10 * 1024 * 1024) { // Skip > 10MB
            return;
        }

        // Basic check to try and only scan text-based files
        // (A real tool would check magic numbers, but this is simple)
        if (!isLikelyTextFile(file)) {
            // System.out.println("Skipping non-text file: " + file.getName());
            return;
        }

        System.out.print("."); // Progress indicator
        List<String> threats = patternMatcher.scanFile(file);
        if (!threats.isEmpty()) {
            report.addInfection(file, threats);
        }
    }

    private boolean isLikelyTextFile(File file) {
        String name = file.getName().toLowerCase();
        return name.endsWith(".txt") || name.endsWith(".java") || name.endsWith(".py") ||
                name.endsWith(".js") || name.endsWith(".html") || name.endsWith(".css") ||
                name.endsWith(".json") || name.endsWith(".xml") || name.endsWith(".md") ||
                name.endsWith(".log") || name.endsWith(".csv") || name.endsWith(".bat") ||
                name.endsWith(".sh");
    }
}