splits webcrack deobfuscated.js into seperate, manageable files

split_js.py

import re, os, sys

def split_deobfuscated(fp, od):
    print(f"reading {fp}...")
    os.makedirs(od, exist_ok=True)

    with open(fp, 'r', encoding='utf-8') as f:
        content = f.read()

    start_match = re.search(r'var \w+\s*=\s*\{', content)
    if not start_match:
        return print("could not find object start.")

    pattern = re.compile(r'(\d+|[\"\']\d+[\"\'])\s*:\s*(\([^)]*\)|function\s*\([^)]*\))\s*=>\s*\{')
    cursor, count = start_match.end(), 0
    
    while True:
        match = pattern.search(content, cursor)
        if not match: break
            
        module_id = match.group(1).strip('\"\'')
        start_pos = match.start()
        idx, balance = match.end(), 1
        in_str, escape = None, False
        
        # parse to end of module block handling strings and comments
        while idx < len(content) and balance > 0:
            char = content[idx]
            if in_str:
                if escape: escape = False
                elif char == '\\': escape = True
                elif char == in_str: in_str = None
            else:
                if char in "\"'`": in_str = char
                elif char == '{': balance += 1
                elif char == '}': balance -= 1
                elif char == '/' and idx + 1 < len(content):
                    if content[idx+1] == '/':
                        idx += 1
                        while idx < len(content) and content[idx] != '\n': idx += 1
                        continue
                    elif content[idx+1] == '*':
                        idx += 2
                        while idx + 1 < len(content) and not (content[idx] == '*' and content[idx+1] == '/'): idx += 1
                        idx += 1; continue
            idx += 1
            
        with open(os.path.join(od, f"{module_id}.js"), 'w', encoding='utf-8') as out:
            out.write(content[start_pos:idx])
            
        count += 1
        cursor = idx
        
    print(f"finished. extracted {count} modules.")

if __name__ == "__main__":
    if len(sys.argv) < 3: print("usage: python split_js.py <input_file> <output_dir>")
    else: split_deobfuscated(sys.argv[1], sys.argv[2])