Scan PHP files for inline <style> and <script> tags.

inlineScan.py

import os
import re
import csv
from openpyxl import Workbook

def find_php_files(base_path):
    """
    Recursively find all .php files in the base directory.

    Args:
        base_path (str): Path to search from.

    Returns:
        list: List of PHP file paths.
    """
    php_files = []
    for root, _, files in os.walk(base_path):
        for file in files:
            if file.endswith('.php'):
                php_files.append(os.path.join(root, file))
    return php_files

def contains_tag(content, tag):
    """
    Check if the content contains a specific HTML tag.

    Args:
        content (str): File content.
        tag (str): Tag to search for, e.g. 'script', 'style'.

    Returns:
        bool: True if tag is found, False otherwise.
    """
    pattern = re.compile(rf"<{tag}\b", re.IGNORECASE)
    return bool(pattern.search(content))

def scan_php_files(php_files):
    """
    Scan each PHP file for <script> and <style> tags.

    Args:
        php_files (list): List of PHP file paths.

    Returns:
        tuple: (list of files with <script>, list of files with <style>)
    """
    script_matches = []
    style_matches = []

    for path in php_files:
        try:
            with open(path, 'r', encoding='utf-8', errors='ignore') as f:
                content = f.read()
                if contains_tag(content, 'script'):
                    script_matches.append(path)
                if contains_tag(content, 'style'):
                    style_matches.append(path)
        except Exception as e:
            print(f"Error reading {path}: {e}")

    return script_matches, style_matches

def save_to_excel(script_files, style_files, output_file='tagged_php_files.xlsx'):
    """
    Save the matched file paths into an Excel file with two tabs.

    Args:
        script_files (list): List of files with <script> tags.
        style_files (list): List of files with <style> tags.
        output_file (str): Path to save the Excel file.
    """
    wb = Workbook()
    
    # Script tag sheet
    ws_script = wb.active
    ws_script.title = "script_tags"
    ws_script.append(["Files with <script> tag"])
    for file in script_files:
        ws_script.append([file])

    # Style tag sheet
    ws_style = wb.create_sheet(title="style_tags")
    ws_style.append(["Files with <style> tag"])
    for file in style_files:
        ws_style.append([file])

    wb.save(output_file)
    print(f"\n✅ Saved results to {output_file}")

def main():
    base_path = os.getcwd()
    php_files = find_php_files(base_path)
    print(f"🔍 Scanning {len(php_files)} PHP files...")

    script_files, style_files = scan_php_files(php_files)
    
    print(f"📄 Found {len(script_files)} files with <script> tags")
    print(f"📄 Found {len(style_files)} files with <style> tags")

    save_to_excel(script_files, style_files)

if __name__ == "__main__":
    main()