Spaces:
Running
Running
| import gradio as gr | |
| import random | |
| import re | |
| import nltk | |
| from nltk.tokenize import sent_tokenize, word_tokenize | |
| from nltk.corpus import stopwords | |
| from nltk.tag import pos_tag | |
| import string | |
| from textstat import flesch_reading_ease, flesch_kincaid_grade | |
| import spacy | |
| from transformers import pipeline, AutoTokenizer, AutoModelForSequenceClassification | |
| import torch | |
| # Download required NLTK data | |
| try: | |
| nltk.download('punkt', quiet=True) | |
| nltk.download('averaged_perceptron_tagger', quiet=True) | |
| nltk.download('stopwords', quiet=True) | |
| nltk.download('wordnet', quiet=True) | |
| except: | |
| pass | |
| class AIContentHumanizer: | |
| def __init__(self): | |
| self.setup_models() | |
| self.setup_humanization_patterns() | |
| def setup_models(self): | |
| """Initialize AI detection and paraphrasing models""" | |
| try: | |
| # Load spaCy model | |
| self.nlp = spacy.load("en_core_web_sm") | |
| except: | |
| # Fallback if spaCy model not available | |
| self.nlp = None | |
| # Paraphrasing patterns and synonyms | |
| self.synonyms = { | |
| 'however': ['but', 'yet', 'though', 'nevertheless', 'still'], | |
| 'therefore': ['so', 'thus', 'hence', 'as a result', 'consequently'], | |
| 'furthermore': ['also', 'moreover', 'additionally', 'besides', 'plus'], | |
| 'nevertheless': ['however', 'yet', 'still', 'but', 'though'], | |
| 'consequently': ['therefore', 'so', 'thus', 'as a result', 'hence'], | |
| 'significant': ['important', 'major', 'notable', 'considerable', 'substantial'], | |
| 'utilize': ['use', 'employ', 'apply', 'make use of', 'take advantage of'], | |
| 'demonstrate': ['show', 'prove', 'display', 'exhibit', 'reveal'], | |
| 'numerous': ['many', 'several', 'various', 'countless', 'multiple'], | |
| 'substantial': ['significant', 'considerable', 'large', 'major', 'important'] | |
| } | |
| def setup_humanization_patterns(self): | |
| """Setup patterns for humanizing text""" | |
| self.ai_phrases = [ | |
| r'\b(it is important to note that|it should be noted that|it is worth noting that)\b', | |
| r'\b(in conclusion|to conclude|in summary|to summarize)\b', | |
| r'\b(furthermore|moreover|additionally|in addition)\b', | |
| r'\b(however|nevertheless|nonetheless|on the other hand)\b', | |
| r'\b(therefore|consequently|as a result|thus)\b', | |
| r'\b(various|numerous|several|multiple)\b', | |
| r'\b(significant|substantial|considerable|notable)\b', | |
| r'\b(utilize|implement|demonstrate|facilitate)\b' | |
| ] | |
| self.human_replacements = { | |
| 'it is important to note that': ['worth mentioning', 'notably', 'interestingly'], | |
| 'it should be noted that': ['keep in mind', 'remember', 'note that'], | |
| 'furthermore': ['also', 'plus', 'what\'s more', 'on top of that'], | |
| 'moreover': ['also', 'besides', 'what\'s more', 'and'], | |
| 'however': ['but', 'though', 'yet', 'still'], | |
| 'therefore': ['so', 'that\'s why', 'which means'], | |
| 'consequently': ['so', 'as a result', 'that\'s why'], | |
| 'numerous': ['many', 'lots of', 'plenty of', 'tons of'], | |
| 'significant': ['big', 'major', 'important', 'key'], | |
| 'utilize': ['use', 'make use of', 'work with'], | |
| 'demonstrate': ['show', 'prove', 'make clear'], | |
| 'facilitate': ['help', 'make easier', 'enable'] | |
| } | |
| self.contraction_map = { | |
| 'it is': 'it\'s', 'that is': 'that\'s', 'there is': 'there\'s', | |
| 'we are': 'we\'re', 'they are': 'they\'re', 'you are': 'you\'re', | |
| 'i am': 'I\'m', 'he is': 'he\'s', 'she is': 'she\'s', | |
| 'will not': 'won\'t', 'cannot': 'can\'t', 'do not': 'don\'t', | |
| 'does not': 'doesn\'t', 'did not': 'didn\'t', 'have not': 'haven\'t', | |
| 'has not': 'hasn\'t', 'had not': 'hadn\'t', 'would not': 'wouldn\'t', | |
| 'should not': 'shouldn\'t', 'could not': 'couldn\'t' | |
| } | |
| def add_human_errors(self, text): | |
| """Add subtle human-like imperfections""" | |
| sentences = sent_tokenize(text) | |
| humanized_sentences = [] | |
| for sentence in sentences: | |
| # Occasionally add filler words | |
| if random.random() < 0.3: | |
| fillers = ['actually', 'basically', 'really', 'pretty much', 'kind of', 'sort of'] | |
| filler = random.choice(fillers) | |
| words = sentence.split() | |
| if len(words) > 3: | |
| insert_pos = random.randint(1, min(3, len(words)-1)) | |
| words.insert(insert_pos, filler) | |
| sentence = ' '.join(words) | |
| # Occasionally use more casual phrasing | |
| if random.random() < 0.2: | |
| sentence = sentence.replace(' and ', ' & ') | |
| humanized_sentences.append(sentence) | |
| return ' '.join(humanized_sentences) | |
| def vary_sentence_structure(self, text): | |
| """Vary sentence structures to appear more human""" | |
| sentences = sent_tokenize(text) | |
| varied_sentences = [] | |
| for i, sentence in enumerate(sentences): | |
| # Randomly combine short sentences | |
| if i < len(sentences) - 1 and len(sentence.split()) < 8 and len(sentences[i+1].split()) < 8: | |
| if random.random() < 0.4: | |
| connectors = [' and ', ', ', ' - ', ' but '] | |
| connector = random.choice(connectors) | |
| combined = sentence.rstrip('.') + connector + sentences[i+1].lower() | |
| varied_sentences.append(combined) | |
| sentences[i+1] = "" # Skip next sentence | |
| continue | |
| if sentence: # Only add non-empty sentences | |
| varied_sentences.append(sentence) | |
| return ' '.join(varied_sentences) | |
| def replace_formal_words(self, text): | |
| """Replace formal words with casual alternatives""" | |
| for formal, casual_options in self.human_replacements.items(): | |
| if formal in text.lower(): | |
| replacement = random.choice(casual_options) | |
| text = re.sub(r'\b' + re.escape(formal) + r'\b', replacement, text, flags=re.IGNORECASE) | |
| return text | |
| def add_contractions(self, text): | |
| """Add contractions to make text more conversational""" | |
| for full_form, contraction in self.contraction_map.items(): | |
| text = re.sub(r'\b' + re.escape(full_form) + r'\b', contraction, text, flags=re.IGNORECASE) | |
| return text | |
| def adjust_punctuation(self, text): | |
| """Adjust punctuation for more natural flow""" | |
| # Replace some periods with commas for flow | |
| sentences = sent_tokenize(text) | |
| if len(sentences) > 2: | |
| # Occasionally use semicolons | |
| if random.random() < 0.3: | |
| idx = random.randint(0, len(sentences)-2) | |
| if len(sentences[idx].split()) > 6: | |
| sentences[idx] = sentences[idx].rstrip('.') + ';' | |
| sentences[idx+1] = sentences[idx+1][0].lower() + sentences[idx+1][1:] | |
| return ' '.join(sentences) | |
| def add_personal_touches(self, text): | |
| """Add personal touches and opinions""" | |
| personal_phrases = [ | |
| "I think", "In my opinion", "From what I've seen", "Personally,", | |
| "I believe", "It seems to me", "From my experience" | |
| ] | |
| sentences = sent_tokenize(text) | |
| if len(sentences) > 1 and random.random() < 0.4: | |
| # Add personal phrase to a random sentence | |
| idx = random.randint(0, len(sentences)-1) | |
| personal_phrase = random.choice(personal_phrases) | |
| sentences[idx] = personal_phrase + " " + sentences[idx].lower() | |
| return ' '.join(sentences) | |
| def paraphrase_with_synonyms(self, text): | |
| """Replace words with synonyms""" | |
| words = word_tokenize(text) | |
| new_words = [] | |
| for word in words: | |
| lower_word = word.lower() | |
| if lower_word in self.synonyms and random.random() < 0.3: | |
| synonym = random.choice(self.synonyms[lower_word]) | |
| new_words.append(synonym) | |
| else: | |
| new_words.append(word) | |
| return ' '.join(new_words) | |
| def humanize_text(self, text, intensity="medium"): | |
| """Main humanization function""" | |
| if not text or not text.strip(): | |
| return "Please provide text to humanize." | |
| # Clean input | |
| text = text.strip() | |
| # Apply humanization techniques based on intensity | |
| if intensity == "light": | |
| text = self.add_contractions(text) | |
| text = self.replace_formal_words(text) | |
| elif intensity == "medium": | |
| text = self.replace_formal_words(text) | |
| text = self.add_contractions(text) | |
| text = self.paraphrase_with_synonyms(text) | |
| text = self.vary_sentence_structure(text) | |
| else: # heavy | |
| text = self.replace_formal_words(text) | |
| text = self.add_contractions(text) | |
| text = self.paraphrase_with_synonyms(text) | |
| text = self.vary_sentence_structure(text) | |
| text = self.add_human_errors(text) | |
| text = self.adjust_punctuation(text) | |
| text = self.add_personal_touches(text) | |
| return text | |
| def get_readability_score(self, text): | |
| """Calculate readability metrics""" | |
| try: | |
| flesch_score = flesch_reading_ease(text) | |
| fk_grade = flesch_kincaid_grade(text) | |
| if flesch_score >= 90: | |
| level = "Very Easy" | |
| elif flesch_score >= 80: | |
| level = "Easy" | |
| elif flesch_score >= 70: | |
| level = "Fairly Easy" | |
| elif flesch_score >= 60: | |
| level = "Standard" | |
| elif flesch_score >= 50: | |
| level = "Fairly Difficult" | |
| elif flesch_score >= 30: | |
| level = "Difficult" | |
| else: | |
| level = "Very Difficult" | |
| return f"Flesch Score: {flesch_score:.1f} ({level})\nGrade Level: {fk_grade:.1f}" | |
| except: | |
| return "Could not calculate readability" | |
| def create_interface(): | |
| humanizer = AIContentHumanizer() | |
| def process_text(input_text, intensity): | |
| if not input_text: | |
| return "Please enter some text to humanize.", "" | |
| humanized = humanizer.humanize_text(input_text, intensity) | |
| readability = humanizer.get_readability_score(humanized) | |
| return humanized, readability | |
| # Custom CSS for better UI | |
| css = """ | |
| .gradio-container { | |
| font-family: 'Inter', sans-serif; | |
| } | |
| .main-header { | |
| text-align: center; | |
| margin-bottom: 30px; | |
| } | |
| .feature-box { | |
| border: 1px solid #e1e5e9; | |
| border-radius: 8px; | |
| padding: 15px; | |
| margin: 10px 0; | |
| background: #f8f9fa; | |
| } | |
| """ | |
| with gr.Blocks(css=css, title="AI Content Humanizer") as interface: | |
| gr.HTML(""" | |
| <div class="main-header"> | |
| <h1>🤖➡️👤 AI Content Humanizer</h1> | |
| <p>Transform AI-generated content into natural, human-like text</p> | |
| </div> | |
| """) | |
| with gr.Row(): | |
| with gr.Column(scale=2): | |
| input_text = gr.Textbox( | |
| label="📝 Enter AI-generated text", | |
| placeholder="Paste your AI-generated content here...", | |
| lines=10, | |
| max_lines=20 | |
| ) | |
| intensity = gr.Radio( | |
| choices=["light", "medium", "heavy"], | |
| value="medium", | |
| label="🎛️ Humanization Intensity", | |
| info="Light: Basic changes | Medium: Moderate restructuring | Heavy: Extensive humanization" | |
| ) | |
| humanize_btn = gr.Button("✨ Humanize Text", variant="primary", size="lg") | |
| with gr.Column(scale=2): | |
| output_text = gr.Textbox( | |
| label="✅ Humanized Text", | |
| lines=10, | |
| max_lines=20, | |
| interactive=False | |
| ) | |
| readability_info = gr.Textbox( | |
| label="📊 Readability Analysis", | |
| lines=3, | |
| interactive=False | |
| ) | |
| gr.HTML(""" | |
| <div class="feature-box"> | |
| <h3>🎯 Features:</h3> | |
| <ul> | |
| <li><strong>Smart Paraphrasing:</strong> Replaces formal AI language with natural expressions</li> | |
| <li><strong>Sentence Restructuring:</strong> Varies sentence patterns for human-like flow</li> | |
| <li><strong>Casual Contractions:</strong> Adds natural contractions and informal language</li> | |
| <li><strong>Personal Touch:</strong> Incorporates personal opinions and experiences</li> | |
| <li><strong>Readability Analysis:</strong> Provides readability scores and grade levels</li> | |
| </ul> | |
| </div> | |
| """) | |
| # Example texts | |
| examples = [ | |
| ["Artificial intelligence has numerous applications in various industries. It is important to note that AI can significantly enhance productivity and efficiency. Furthermore, machine learning algorithms can process vast amounts of data to generate valuable insights.", "medium"], | |
| ["The implementation of sustainable practices is crucial for environmental conservation. Organizations must utilize innovative technologies to facilitate eco-friendly operations. Moreover, it is essential to demonstrate commitment to reducing carbon footprint.", "heavy"], | |
| ["In conclusion, the research demonstrates that renewable energy sources are becoming increasingly viable. Therefore, governments should implement policies that promote clean energy adoption. It is worth noting that this transition requires substantial investment.", "light"] | |
| ] | |
| gr.Examples( | |
| examples=examples, | |
| inputs=[input_text, intensity], | |
| label="📚 Try these examples:" | |
| ) | |
| humanize_btn.click( | |
| fn=process_text, | |
| inputs=[input_text, intensity], | |
| outputs=[output_text, readability_info] | |
| ) | |
| return interface | |
| if __name__ == "__main__": | |
| interface = create_interface() | |
| interface.launch() |