import gradio as gr import pytesseract from PIL import Image from transformers import pipeline import re from langdetect import detect from deep_translator import GoogleTranslator # Translator instance translator = GoogleTranslator(source="auto", target="es") # 1. Load separate keywords for SMiShing and Other Scam (assumed in English) with open("smishing_keywords.txt", "r", encoding="utf-8") as f: SMISHING_KEYWORDS = [line.strip().lower() for line in f if line.strip()] with open("other_scam_keywords.txt", "r", encoding="utf-8") as f: OTHER_SCAM_KEYWORDS = [line.strip().lower() for line in f if line.strip()] # 2. Zero-Shot Classification Pipeline model_name = "joeddav/xlm-roberta-large-xnli" classifier = pipeline("zero-shot-classification", model=model_name) CANDIDATE_LABELS = ["SMiShing", "Other Scam", "Legitimate"] def get_keywords_by_language(text: str): """ Detect language using `langdetect` and translate keywords if needed. """ snippet = text[:200] # Use a snippet for detection try: detected_lang = detect(snippet) except Exception: detected_lang = "en" # Default to English if detection fails if detected_lang == "es": smishing_in_spanish = [ translator.translate(kw).lower() for kw in SMISHING_KEYWORDS ] other_scam_in_spanish = [ translator.translate(kw).lower() for kw in OTHER_SCAM_KEYWORDS ] return smishing_in_spanish, other_scam_in_spanish, "es" else: return SMISHING_KEYWORDS, OTHER_SCAM_KEYWORDS, "en" def boost_probabilities(probabilities: dict, text: str): """ Boost probabilities based on keyword matches and presence of URLs. """ lower_text = text.lower() smishing_keywords, other_scam_keywords, detected_lang = get_keywords_by_language(text) smishing_count = sum(1 for kw in smishing_keywords if kw in lower_text) other_scam_count = sum(1 for kw in other_scam_keywords if kw in lower_text) # Example: 30% per found keyword smishing_boost = 0.30 * smishing_count other_scam_boost = 0.30 * other_scam_count found_urls = re.findall(r"(https?://[^\s]+)", lower_text) if found_urls: # 35% boost for Smishing if there's a URL smishing_boost += 0.35 p_smishing = probabilities.get("SMiShing", 0.0) p_other_scam = probabilities.get("Other Scam", 0.0) p_legit = probabilities.get("Legitimate", 1.0) p_smishing += smishing_boost p_other_scam += other_scam_boost p_legit -= (smishing_boost + other_scam_boost) # Clamp to 0 p_smishing = max(p_smishing, 0.0) p_other_scam = max(p_other_scam, 0.0) p_legit = max(p_legit, 0.0) # Re-normalize total = p_smishing + p_other_scam + p_legit if total > 0: p_smishing /= total p_other_scam /= total p_legit /= total else: p_smishing, p_other_scam, p_legit = 0.0, 0.0, 1.0 return { "SMiShing": p_smishing, "Other Scam": p_other_scam, "Legitimate": p_legit, "detected_lang": detected_lang } def smishing_detector(input_type, text, image): """ Main detection function: - If input_type == "Text": use `text` as the message - If input_type == "Screenshot": use OCR on `image` to get text """ if input_type == "Text": # Use the pasted text combined_text = text.strip() if text else "" else: # input_type == "Screenshot" if image is not None: ocr_text = pytesseract.image_to_string(image, lang="spa+eng") combined_text = ocr_text.strip() else: combined_text = "" if not combined_text: return { "text_used_for_classification": "(none)", "label": "No text provided", "confidence": 0.0, "keywords_found": [], "urls_found": [] } # Zero-shot classification result = classifier( sequences=combined_text, candidate_labels=CANDIDATE_LABELS, hypothesis_template="This message is {}." ) original_probs = {k: float(v) for k, v in zip(result["labels"], result["scores"])} # Boost logic boosted = boost_probabilities(original_probs, combined_text) # Convert to float boosted = {k: float(v) for k, v in boosted.items() if isinstance(v, (int, float))} detected_lang = boosted.pop("detected_lang", "en") # Final classification final_label = max(boosted, key=boosted.get) final_confidence = round(boosted[final_label], 3) # For display lower_text = combined_text.lower() smishing_keys, scam_keys, _ = get_keywords_by_language(combined_text) found_urls = re.findall(r"(https?://[^\s]+)", lower_text) found_smishing = [kw for kw in smishing_keys if kw in lower_text] found_other_scam = [kw for kw in scam_keys if kw in lower_text] return { "detected_language": detected_lang, "text_used_for_classification": combined_text, "original_probabilities": {k: round(v, 3) for k, v in original_probs.items()}, "boosted_probabilities": {k: round(v, 3) for k, v in boosted.items()}, "label": final_label, "confidence": final_confidence, "smishing_keywords_found": found_smishing, "other_scam_keywords_found": found_other_scam, "urls_found": found_urls, } # Create a Radio for user choice + text input + image input demo = gr.Interface( fn=smishing_detector, inputs=[ gr.Radio( choices=["Text", "Screenshot"], label="Choose input type", value="Text", # default info="Select 'Text' to paste a message, or 'Screenshot' to upload an image." ), gr.Textbox( lines=3, label="Paste Suspicious SMS Text", placeholder="Type or paste the message here..." ), gr.Image( type="pil", label="Upload a Screenshot", ) ], outputs="json", title="SMiShing & Scam Detector", description=""" Select "Text" or "Screenshot" above. - If "Text", only use the textbox. - If "Screenshot", only upload an image. The app will classify the message as SMiShing, Other Scam, or Legitimate. """, allow_flagging="never" ) if __name__ == "__main__": demo.launch()