lilbool/vuln-code-analysis

1import os
2import re
3import requests
4from bs4 import BeautifulSoup
5
6def sanitize_filename(filename):
7    """Remove ou substitui caracteres inválidos em nomes de arquivos."""
8    return re.sub(r'[<>:"/\\|?*]', '_', filename)
9
10def fetch_task_links(category_url):
11    """Fetch all task links from Rosetta Code's category page."""
12    response = requests.get(category_url)
13    if response.status_code != 200:
14        print(f"[ERROR] Failed to fetch {category_url}. Status code: {response.status_code}")
15        return []
16
17    soup = BeautifulSoup(response.text, 'html.parser')
18    links = soup.select('.mw-category-group ul li a')
19    return [("https://rosettacode.org" + link['href'], link.text) for link in links]
20
21def fetch_code_from_task(task_url):
22    """Fetch code snippets from a specific task on Rosetta Code."""
23    response = requests.get(task_url)
24    if response.status_code != 200:
25        print(f"[ERROR] Failed to fetch {task_url}. Status code: {response.status_code}")
26        return []
27
28    soup = BeautifulSoup(response.text, 'html.parser')
29    code_blocks = soup.find_all('pre')
30    return [code.text for code in code_blocks]
31
32def save_safe_codes(task_name, codes, save_dir):
33    """Save the safe codes as text files."""
34    os.makedirs(save_dir, exist_ok=True)
35    task_name = sanitize_filename(task_name)  # Sanitizar o nome da tarefa
36    for i, code in enumerate(codes):
37        filename = f"{task_name}_{i+1}.txt"
38        filepath = os.path.join(save_dir, filename)
39        try:
40            with open(filepath, 'w', encoding='utf-8') as f:
41                f.write(code)
42            print(f"[SUCCESS] Saved: {filepath}")
43        except Exception as e:
44            print(f"[ERROR] Could not save file {filepath}: {e}")
45
46if __name__ == "__main__":
47    category_url = "https://rosettacode.org/wiki/Category:Programming_Tasks"
48    save_directory = "safe-code-analyzer/safe_codes"
49
50    # Fetch tasks
51    tasks = fetch_task_links(category_url)
52    print(f"[INFO] Found {len(tasks)} tasks on Rosetta Code.")
53
54    # Fetch and save codes
55    for task_url, task_name in tasks[:10]:  # Ajuste o número de tarefas a serem processadas
56        print(f"[INFO] Fetching codes for task: {task_name}")
57        codes = fetch_code_from_task(task_url)
58        save_safe_codes(task_name, codes, save_directory)
59