夜雨聆风一、加密PDF
1.1 基本加密
import PyPDF2defencrypt_pdf(filepath, output_path, password):"""加密PDF文件"""withopen(filepath, 'rb') as file: reader = PyPDF2.PdfReader(file) writer = PyPDF2.PdfWriter()for page in reader.pages: writer.add_page(page) writer.encrypt(password)withopen(output_path, 'wb') as output: writer.write(output)print(f"PDF已加密: {output_path}")# 使用encrypt_pdf('document.pdf', 'encrypted.pdf', 'password123')1.2 不同加密算法
import PyPDF2defencrypt_with_algorithm(filepath, output_path, password, algorithm='AES-256'):""" 使用不同算法加密 算法: 'AES-256', 'RC4-128' """withopen(filepath, 'rb') as file: reader = PyPDF2.PdfReader(file) writer = PyPDF2.PdfWriter()for page in reader.pages: writer.add_page(page)# PyPDF2使用AES-256作为默认# 通过use_aes=True启用AES-256if algorithm == 'AES-256': writer.encrypt(password, use_aes=True)else: writer.encrypt(password, use_aes=False)withopen(output_path, 'wb') as output: writer.write(output)print(f"使用{algorithm}加密完成: {output_path}")# 使用不同算法encrypt_with_algorithm('document.pdf', 'aes_encrypted.pdf', 'password', 'AES-256')encrypt_with_algorithm('document.pdf', 'rc4_encrypted.pdf', 'password', 'RC4-128')1.3 设置不同权限
import PyPDF2defencrypt_with_permissions(filepath, output_path, user_password, owner_password=None):"""加密并设置权限"""withopen(filepath, 'rb') as file: reader = PyPDF2.PdfReader(file) writer = PyPDF2.PdfWriter()for page in reader.pages: writer.add_page(page)# 设置权限 permissions = {'print': False, # 禁止打印'extract': False, # 禁止提取文本'modify': False, # 禁止修改'copy': False, # 禁止复制'annotate': False, # 禁止注释'fill_forms': False, # 禁止填写表单'accessibility': False, # 禁止无障碍访问'assemble': False, # 禁止文档组装'print_high_res': False, # 禁止高分辨率打印 }# 设置权限标志if owner_password isNone: owner_password = user_password writer.encrypt( user_password=user_password, owner_password=owner_password, permissions_flag=0# 禁止所有操作 )withopen(output_path, 'wb') as output: writer.write(output)print(f"带权限加密完成: {output_path}")# 使用encrypt_with_permissions('document.pdf', 'restricted.pdf', 'user123', 'owner456')二、解密PDF
2.1 基本解密
import PyPDF2defdecrypt_pdf(filepath, output_path, password):"""解密PDF文件"""withopen(filepath, 'rb') as file: reader = PyPDF2.PdfReader(file)if reader.is_encrypted: reader.decrypt(password) writer = PyPDF2.PdfWriter()for page in reader.pages: writer.add_page(page)withopen(output_path, 'wb') as output: writer.write(output)print(f"PDF已解密: {output_path}")else:print("文件未加密")# 使用decrypt_pdf('encrypted.pdf', 'decrypted.pdf', 'password123')2.2 尝试多种密码
import PyPDF2defdecrypt_with_try(filepath, output_path, password_list):"""尝试多种密码解密"""withopen(filepath, 'rb') as file: reader = PyPDF2.PdfReader(file)ifnot reader.is_encrypted:print("文件未加密")returnfor password in password_list:try: reader.decrypt(password)# 验证是否解密成功ifnot reader.is_encrypted: writer = PyPDF2.PdfWriter()for page in reader.pages: writer.add_page(page)withopen(output_path, 'wb') as output: writer.write(output)print(f"解密成功! 密码: {password}")returnexcept Exception:continueprint("所有密码尝试失败")# 使用passwords = ['123456', 'password', 'admin', '12345', 'secret']decrypt_with_try('encrypted.pdf', 'decrypted.pdf', passwords)2.3 检查加密信息
import PyPDF2defget_encryption_info(filepath):"""获取PDF加密信息"""withopen(filepath, 'rb') as file: reader = PyPDF2.PdfReader(file) info = {'is_encrypted': reader.is_encrypted, }if reader.is_encrypted:# 获取加密信息 info['encryption_version'] = getattr(reader, 'encryption_version', 'Unknown')# 检查是否有权限限制try:# 尝试读取页面 reader.pages[0] info['has_permissions'] = Trueexcept Exception: info['has_permissions'] = Falsereturn info# 使用info = get_encryption_info('encrypted.pdf')print("加密信息:")for key, value in info.items():print(f" {key}: {value}")三、密码管理
3.1 批量加密
import PyPDF2from pathlib import Pathimport osdefbatch_encrypt(folder_path, output_dir, password):"""批量加密PDF文件""" os.makedirs(output_dir, exist_ok=True) pdf_files = list(Path(folder_path).glob('*.pdf'))for pdf_path in pdf_files:try:withopen(pdf_path, 'rb') as file: reader = PyPDF2.PdfReader(file) writer = PyPDF2.PdfWriter()for page in reader.pages: writer.add_page(page) writer.encrypt(password) output_path = os.path.join(output_dir, pdf_path.name)withopen(output_path, 'wb') as output: writer.write(output)print(f"已加密: {pdf_path.name}")except Exception as e:print(f"加密失败: {pdf_path.name} - {e}")# 使用batch_encrypt('./pdfs', './encrypted', 'batch_password')3.2 批量解密
import PyPDF2from pathlib import Pathimport osdefbatch_decrypt(folder_path, output_dir, password):"""批量解密PDF文件""" os.makedirs(output_dir, exist_ok=True) pdf_files = list(Path(folder_path).glob('*.pdf'))for pdf_path in pdf_files:try:withopen(pdf_path, 'rb') as file: reader = PyPDF2.PdfReader(file)if reader.is_encrypted: reader.decrypt(password) writer = PyPDF2.PdfWriter()for page in reader.pages: writer.add_page(page) output_path = os.path.join(output_dir, pdf_path.name)withopen(output_path, 'wb') as output: writer.write(output)print(f"已解密: {pdf_path.name}")except Exception as e:print(f"解密失败: {pdf_path.name} - {e}")# 使用batch_decrypt('./encrypted', './decrypted', 'batch_password')四、密码破解（仅用于合法用途）
4.1 暴力破解
import PyPDF2import itertoolsimport stringdefbrute_force_pdf(filepath, max_length=4):"""暴力破解PDF密码（仅用于合法测试）""" chars = string.ascii_lowercase + string.digitswithopen(filepath, 'rb') as file: reader = PyPDF2.PdfReader(file)ifnot reader.is_encrypted:print("文件未加密")returnNonefor length inrange(1, max_length + 1):print(f"尝试长度 {length}...")for password in itertools.product(chars, repeat=length): password_str = ''.join(password)try: reader = PyPDF2.PdfReader(file) reader.decrypt(password_str)# 验证是否成功ifnot reader.is_encrypted:print(f"密码找到: {password_str}")return password_strexcept Exception:continueprint("未找到密码")returnNone# 使用（仅用于测试）# brute_force_pdf('encrypted.pdf', max_length=3)4.2 字典攻击
import PyPDF2defdictionary_attack(filepath, wordlist_path):"""字典攻击破解PDF密码"""withopen(filepath, 'rb') as file: reader = PyPDF2.PdfReader(file)ifnot reader.is_encrypted:print("文件未加密")returnNonewithopen(wordlist_path, 'r', encoding='utf-8') as f:for line in f: password = line.strip()try:# 重新打开文件withopen(filepath, 'rb') as f2: reader2 = PyPDF2.PdfReader(f2) reader2.decrypt(password)ifnot reader2.is_encrypted:print(f"密码找到: {password}")return passwordexcept Exception:continueprint("未找到密码")returnNone# 使用# dictionary_attack('encrypted.pdf', 'wordlist.txt')五、实战案例
5.1 安全文档处理系统
import PyPDF2import osfrom pathlib import Pathimport hashlibfrom datetime import datetimeclassSecurePDFProcessor:"""安全PDF处理系统"""def__init__(self, master_password):self.master_password = master_passwordself.processed_files = []defencrypt_file(self, filepath, output_dir='encrypted'):"""加密文件""" os.makedirs(output_dir, exist_ok=True)# 生成基于文件内容的密码 file_hash = self._generate_file_hash(filepath) password = f"{self.master_password}_{file_hash[:8]}"withopen(filepath, 'rb') as file: reader = PyPDF2.PdfReader(file) writer = PyPDF2.PdfWriter()for page in reader.pages: writer.add_page(page) writer.encrypt(password) output_path = os.path.join(output_dir, Path(filepath).name)withopen(output_path, 'wb') as output: writer.write(output)self.processed_files.append({'file': filepath,'encrypted_file': output_path,'password': password,'timestamp': datetime.now() })print(f"已加密: {Path(filepath).name}")return output_pathdefdecrypt_file(self, filepath, output_dir='decrypted'):"""解密文件""" os.makedirs(output_dir, exist_ok=True)# 查找对应的加密信息 encrypted_info = Nonefor info inself.processed_files:if info['encrypted_file'] == filepath: encrypted_info = infobreakifnot encrypted_info:print("未找到加密信息")returnNonewithopen(filepath, 'rb') as file: reader = PyPDF2.PdfReader(file)if reader.is_encrypted: reader.decrypt(encrypted_info['password']) writer = PyPDF2.PdfWriter()for page in reader.pages: writer.add_page(page) output_path = os.path.join(output_dir, Path(filepath).name)withopen(output_path, 'wb') as output: writer.write(output)print(f"已解密: {Path(filepath).name}")return output_pathdef_generate_file_hash(self, filepath):"""生成文件哈希""" hasher = hashlib.sha256()withopen(filepath, 'rb') as f:for chunk initer(lambda: f.read(4096), b''): hasher.update(chunk)return hasher.hexdigest()defget_encrypted_files(self):"""获取已加密文件列表"""returnself.processed_filesdefgenerate_report(self, output_path='encryption_report.txt'):"""生成加密报告"""withopen(output_path, 'w', encoding='utf-8') as f: f.write("PDF加密处理报告\n") f.write("="*50 + "\n") f.write(f"生成时间: {datetime.now()}\n") f.write(f"处理文件数: {len(self.processed_files)}\n") f.write("="*50 + "\n\n")for i, info inenumerate(self.processed_files, 1): f.write(f"文件 {i}:\n") f.write(f" 原始文件: {info['file']}\n") f.write(f" 加密文件: {info['encrypted_file']}\n") f.write(f" 密码: {info['password']}\n") f.write(f" 时间: {info['timestamp']}\n") f.write("\n")# 使用processor = SecurePDFProcessor('master_key_2024')processor.encrypt_file('document.pdf')processor.encrypt_file('report.pdf')processor.generate_report()5.2 批量加密工具
import PyPDF2import osfrom pathlib import Pathimport csvimport jsonclassBatchEncryptionTool:"""批量加密工具"""def__init__(self, password_file='passwords.csv'):self.password_file = password_fileself.password_map = {}self._load_passwords()def_load_passwords(self):"""加载密码映射"""if os.path.exists(self.password_file):withopen(self.password_file, 'r', encoding='utf-8') as f: reader = csv.reader(f)for row in reader:iflen(row) >= 2:self.password_map[row[0]] = row[1]def_save_passwords(self):"""保存密码映射"""withopen(self.password_file, 'w', newline='', encoding='utf-8') as f: writer = csv.writer(f)for filename, password inself.password_map.items(): writer.writerow([filename, password])defencrypt_files(self, folder_path, output_dir='encrypted'):"""批量加密""" os.makedirs(output_dir, exist_ok=True) pdf_files = list(Path(folder_path).glob('*.pdf'))for pdf_path in pdf_files: filename = pdf_path.name# 使用已有密码或生成新密码if filename inself.password_map: password = self.password_map[filename]else: password = self._generate_password()self.password_map[filename] = passwordtry:withopen(pdf_path, 'rb') as file: reader = PyPDF2.PdfReader(file) writer = PyPDF2.PdfWriter()for page in reader.pages: writer.add_page(page) writer.encrypt(password) output_path = os.path.join(output_dir, filename)withopen(output_path, 'wb') as output: writer.write(output)print(f"已加密: {filename}")except Exception as e:print(f"加密失败: {filename} - {e}")self._save_passwords()print(f"密码已保存到: {self.password_file}")def_generate_password(self):"""生成随机密码"""import randomimport string chars = string.ascii_letters + string.digitsreturn''.join(random.choice(chars) for _ inrange(12))defget_password(self, filename):"""获取文件密码"""returnself.password_map.get(filename)# 使用tool = BatchEncryptionTool('encryption_passwords.csv')tool.encrypt_files('./pdfs')# 获取特定文件密码password = tool.get_password('document.pdf')print(f"document.pdf 密码: {password}")六、注意事项
6.1 错误处理
import PyPDF2defsafe_encrypt(filepath, output_path, password):"""安全的加密操作"""try:withopen(filepath, 'rb') as file: reader = PyPDF2.PdfReader(file)# 检查PDF是否已加密if reader.is_encrypted:print("文件已加密")return writer = PyPDF2.PdfWriter()for page in reader.pages: writer.add_page(page) writer.encrypt(password)withopen(output_path, 'wb') as output: writer.write(output)print(f"加密成功: {output_path}")except FileNotFoundError:print(f"文件不存在: {filepath}")except PyPDF2.errors.PdfReadError:print(f"无效的PDF文件: {filepath}")except Exception as e:print(f"加密失败: {e}")# 使用safe_encrypt('document.pdf', 'encrypted.pdf', 'password')6.2 密码安全
import hashlibimport base64defsecure_password(password):"""安全处理密码"""# 使用SHA-256哈希 hashed = hashlib.sha256(password.encode()).hexdigest()return hasheddefsave_password_secure(password, output_path):"""安全保存密码"""import jsonfrom datetime import datetime data = {'password': secure_password(password),'timestamp': datetime.now().isoformat(),'note': 'PDF encryption password' }withopen(output_path, 'w') as f: json.dump(data, f, indent=2)七、总结
# 快速参考# 1. 加密PDFwriter = PyPDF2.PdfWriter()writer.encrypt(password)writer.encrypt(user_password, owner_password)# 2. 使用AES-256writer.encrypt(password, use_aes=True)# 3. 解密PDFreader.decrypt(password)# 4. 检查加密状态if reader.is_encrypted:print("文件已加密")# 5. 批量处理for pdf in pdf_files: reader = PyPDF2.PdfReader(pdf)if reader.is_encrypted: reader.decrypt(password)# 6. 安全建议# - 使用强密码（大小写字母+数字+特殊字符）# - 定期更换密码# - 备份原始文件# - 记录加密信息（安全保存）PyPDF2提供了完整的PDF加密与解密功能，可以保护敏感文档的安全。加密时建议使用强密码和AES-256算法；解密时需确保有正确的密码。在批量处理时，注意密码的安全管理和备份。合理使用加密功能可以保护PDF文档的内容安全。