python代码 检查文件里面有没有密码之类

python代码 检查文件里面有没有密码之类

# # 遍历文件，查看文件里面有没有密码之类的要隐藏，python代码

ignore_patterns=['password = #',"password=#",'test="password',"#{password,jdbcType=VARCHAR}",
                 "item.password","#{password}","password,real_name",]
def has_pattern(line,ignore_patterns):
    for ignore_pattern in ignore_patterns:
        if ignore_pattern in line:
            return True
    return False

err_list=[]

def get_context(lines, line_num, context_length):
    start_line = max(0, line_num - context_length)
    end_line = min(len(lines), line_num + context_length)
    context = lines[start_line:end_line]
    return ''.join(context)

def get_context_chars(line, keyword, context_length=50):
    keyword_index = line.find(keyword)
    if keyword_index != -1:
        start_index = max(0, keyword_index - context_length)
        end_index = min(len(line), keyword_index + len(keyword) + context_length)
        context = line[start_index:end_index]
        # print(context)
        return context
    return None

def check_file_for_sensitive_info(file_path, sensitive_keywords):
    try:
        have_sensitive = False
        with open(file_path, 'r', encoding="utf-8") as file:
            for line_num, line in enumerate(file, start=1):
                for keyword in sensitive_keywords:
                    if keyword in line:
                        should_xzs=" xzs " in line or  "-xzs" in line or  "xzs-" in line
                        if  keyword=="xzs" and not should_xzs:
                            continue
                    
                        if has_pattern(line,ignore_patterns):
                            continue
                        print(f"{file_path}:{line_num}")
                        context_chars=get_context_chars(line, keyword, context_length=50)
                        print(context_chars.strip())
                        have_sensitive = True
        return have_sensitive
    except Exception as e:
        err_list.append({
            "file_path":file_path,
            "erorr":e
        })
      
        return False


import os


def traverse_directory(directory_path, sensitive_keywords, ignore_dirs=[]):
    for root, dirs, files in os.walk(directory_path):
    
        dirs[:] = [d for d in dirs if d not in ignore_dirs]

        for file in files:
            file_path = os.path.join(root, file)
            if check_file_for_sensitive_info(file_path, sensitive_keywords):
          
                pass

import json_util
sesentive_words_path=rf'D:\home\app\private-conf\sesentive_words.json'
conf=json_util.file_path_to_dict(sesentive_words_path)

directory_path = rf'D:\proj\python\st-util'

def is_have_idea(directory_path):
    gitignore_file_path=os.path.join(directory_path,".gitignore")
    if os.path.exists(gitignore_file_path):
        with open(gitignore_file_path, 'r',encoding="utf-8") as file:
            gitignore_data=file.read()
            if ".idea" in gitignore_data:
        
                return True
                pass

    return False


sensitive_keywords =conf.get("sesentive_words")
# 要忽略的目录列表
is_have_idea_yes=is_have_idea(directory_path=directory_path)

ignore_dirs = ['__pycache__','.git',"node_modules","target"]
if is_have_idea_yes:
    ignore_dirs.append('.idea')

traverse_directory(directory_path, sensitive_keywords, ignore_dirs)

{
  "sesentive_words": ["你的某些密码之类不适合git commit的东西"]
}