text_obfuscator.py

import random
from typing import List
import unicodedata

class TextObfuscator:
    def __init__(self):
        self.zero_width_space = "‎"
        self.invisible_chars = ["⁢", "⁣", "⁤", "⁡", "⁠", ""]  # Various invisible unicode chars
        
        # Extended character mappings
        self.similar_chars = {
            '.': ['．', '｡', '。', '․', '⋅', '∙', '⸱', '⸳'],
            'a': ['а', 'ａ', '𝚊', 'ａ', 'ᴀ', 'ⓐ', '𝒂', '𝓪'],
            's': ['ѕ', 'ｓ', '𝚜', 'ѕ', 'ꜱ', 'ⓢ', '𝒔', '𝓼'],
            'i': ['і', 'ｉ', '𝚒', 'ı', 'ɪ', 'ⓘ', '𝒊', '𝓲'],
            'z': ['ｚ', '𝚣', 'ᴢ', 'ⓩ', '𝒛', '𝓏'],
            'e': ['е', 'ｅ', '𝚎', 'ᴇ', 'ⓔ', '𝒆', '𝓮'],
            'm': ['ｍ', '𝚖', 'ᴍ', 'ⓜ', '𝒎', '𝓶'],
            't': ['ｔ', '𝚝', 'ᴛ', 'ⓣ', '𝒕', '𝓽'],
            'r': ['ｒ', '𝚛', 'ʀ', 'ⓡ', '𝒓', '𝓻'],
            'f': ['ｆ', '𝚏', 'ꜰ', 'ⓕ', '𝒇', '𝓯'],
            'u': ['ｕ', '𝚞', 'ᴜ', 'ⓤ', '𝒖', '𝓾'],
            'n': ['ｎ', '𝚗', 'ɴ', 'ⓝ', '𝒏', '𝓷']
        }
        
        # Common words to replace
        self.word_replacements = {
            'size': ['ѕιzє', 'ꜱɪᴢᴇ', 'ˢⁱᶻᵉ', '𝓼𝓲𝔃𝓮', '𝕤𝕚𝕫𝕖'],
            'matters': ['ᴍᴀᴛᴛᴇʀꜱ', 'мαттєяѕ', 'ᵐᵃᵗᵗᵉʳˢ', '𝓶𝓪𝓽𝓽𝓮𝓻𝓼', '𝕞𝕒𝕥𝕥𝕖𝕣𝕤'],
            'fun': ['ғᴜɴ', 'ꜰᴜɴ', 'ᶠᵘⁿ', '𝓯𝓾𝓷', '𝕗𝕦𝕟'],
            'token': ['ᴛᴏᴋᴇɴ', 'токєη', 'ᵗᵒᵏᵉⁿ', '𝓽𝓸𝓴𝓮𝓷', '𝕥𝕠𝕜𝕖𝕟'],
            'utility': ['ᴜᴛɪʟɪᴛʏ', 'υтιℓιту', 'ᵘᵗⁱˡⁱᵗʸ', '𝓾𝓽𝓲𝓵𝓲𝓽𝔂', '𝕦𝕥𝕚𝕝𝕚𝕥𝕪']
        }

    def obfuscate_url(self, text: str) -> str:
        techniques: List[callable] = [
            self._stylize_text,
            self._mix_invisible_chars,
            self._full_width_text,
            self._mathematical_bold,
            self._bubble_text,
            self._combine_techniques
        ]
        return random.choice(techniques)(text)
    
    def _stylize_text(self, text: str) -> str:
        """Replace words with stylized versions"""
        words = text.lower().split()
        result = []
        
        for word in words:
            if word in self.word_replacements:
                result.append(random.choice(self.word_replacements[word]))
            else:
                result.append(word)
                
        return " ".join(result)
    
    def _mix_invisible_chars(self, text: str) -> str:
        """Mix invisible characters into text"""
        result = ""
        for char in text:
            result += char
            if random.random() < 0.3:  # 30% chance to add invisible char
                result += random.choice(self.invisible_chars)
        return result
    
    def _full_width_text(self, text: str) -> str:
        """Convert to full-width characters"""
        # Mapping of normal to full-width characters
        full_width_map = {
            'a': 'ａ', 'b': 'ｂ', 'c': 'ｃ', 'd': 'ｄ', 'e': 'ｅ',
            'f': 'ｆ', 'g': 'ｇ', 'h': 'ｈ', 'i': 'ｉ', 'j': 'ｊ',
            'k': 'ｋ', 'l': 'ｌ', 'm': 'ｍ', 'n': 'ｎ', 'o': 'ｏ',
            'p': 'ｐ', 'q': 'ｑ', 'r': 'ｒ', 's': 'ｓ', 't': 'ｔ',
            'u': 'ｕ', 'v': 'ｖ', 'w': 'ｗ', 'x': 'ｘ', 'y': 'ｙ',
            'z': 'ｚ', '0': '０', '1': '１', '2': '２', '3': '３',
            '4': '４', '5': '５', '6': '６', '7': '７', '8': '８',
            '9': '９', ' ': '　', '.': '．'
        }
        return ''.join(full_width_map.get(c.lower(), c) for c in text)
    
    def _mathematical_bold(self, text: str) -> str:
        """Convert to mathematical bold characters"""
        normal = 'abcdefghijklmnopqrstuvwxyz'
        math_bold = '𝐚𝐛𝐜𝐝𝐞𝐟𝐠𝐡𝐢𝐣𝐤𝐥𝐦𝐧𝐨𝐩𝐪𝐫𝐬𝐭𝐮𝐯𝐰𝐱𝐲𝐳'
        trans = str.maketrans(normal, math_bold)
        return text.translate(trans)
    
    def _bubble_text(self, text: str) -> str:
        """Convert to bubble characters"""
        normal = 'abcdefghijklmnopqrstuvwxyz'
        bubble = 'ⓐⓑⓒⓓⓔⓕⓖⓗⓘⓙⓚⓛⓜⓝⓞⓟⓠⓡⓢⓣⓤⓥⓦⓧⓨⓩ'
        trans = str.maketrans(normal, bubble)
        return text.translate(trans)
    
    def _combine_techniques(self, text: str) -> str:
        """Combine multiple techniques"""
        # Split into words
        words = text.split()
        result = []
        
        for word in words:
            if word.lower() in self.word_replacements:
                # Use stylized version for known words
                result.append(random.choice(self.word_replacements[word.lower()]))
            else:
                # Apply random technique for other words
                technique = random.choice([
                    self._mathematical_bold,
                    self._bubble_text,
                    self._full_width_text
                ])
                result.append(technique(word))
        
        # Add some invisible characters
        final_text = " ".join(result)
        return self._mix_invisible_chars(final_text)