Delete app.py

app.py

@@ -1,224 +0,0 @@
-"""
-HuggingFace Space App for PII Detection
-This app uses a BERT model to identify Personal Identifiable Information in text.
-"""
-
-import gradio as gr
-from transformers import AutoTokenizer, AutoModelForTokenClassification
-import torch
-
-# Load the model and tokenizer
-MODEL_PATH = "./Bert_NER_PII_Multilingual"
-tokenizer = AutoTokenizer.from_pretrained(MODEL_PATH)
-model = AutoModelForTokenClassification.from_pretrained(MODEL_PATH)
-
-# Entity label colors for visualization
-ENTITY_COLORS = {
-    "NAME": "#FF6B6B",
-    "EMAIL": "#4ECDC4",
-    "CREDITCARDNUMBER": "#FFE66D",
-    "IP": "#95E1D3",
-    "PASSWORD": "#F38181",
-    "STREET": "#AA96DA",
-    "ACCOUNTNAME": "#FCBAD3",
-    "USERNAME": "#A8E6CF",
-    "ZIPCODE": "#FFD3B6",
-    "IBAN": "#FFAAA5",
-    "URL": "#FF8B94",
-    "JOB": "#C7CEEA",
-    "GENDER": "#FFDAC1",
-    "ADDRESS": "#B5EAD7",
-    "MAC": "#C9CBA3",
-    "GEO": "#FFE2E2",
-    "NEARBYGPSCOORDINATE": "#F7D9C4",
-    "COINADDRESS": "#FAACA8",
-    "CREDITCARDISSUER": "#DCD6F7",
-    "CURRENCY": "#A6D9F7",
-    "NUM": "#D4F1F4",
-    "BIC": "#FFB6B9",
-    "ORDINALDIRECTION": "#F6EAC2",
-    "PHONENUMBER": "#FFB3BA",
-    "SSN": "#FF677D",
-    "DATE": "#BAE1FF",
-    "TIME": "#FFFFB5",
-    "AGE": "#FFDFBA",
-    "ORG": "#BAFFC9",
-    "VEHICLEVIN": "#D4A5A5",
-    "VEHICLEVRM": "#9B9B9B",
-    "PHONEIMEI": "#E0BBE4",
-    "PREFIX": "#FFDFD3",
-    "HEIGHT": "#C7CEEA",
-    "WEIGHTS": "#F0E68C",
-    "BLOODTYPE": "#FFB6C1",
-    "COLOR": "#E6E6FA",
-    "MISC": "#D3D3D3",
-}
-
-
-def detect_pii(text):
-    """
-    Detect PII entities in the input text.
-    
-    Args:
-        text (str): Input text to analyze
-        
-    Returns:
-        list: Highlighted entities for Gradio display
-        str: Summary of detected entities
-    """
-    if not text.strip():
-        return None, "Please enter some text to analyze."
-    
-    # Tokenize input
-    inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=512)
-    
-    # Get predictions
-    with torch.no_grad():
-        outputs = model(**inputs)
-        predictions = torch.argmax(outputs.logits, dim=2)
-    
-    # Convert tokens back to words and align with predictions
-    tokens = tokenizer.convert_ids_to_tokens(inputs["input_ids"][0])
-    predicted_labels = [model.config.id2label[pred.item()] for pred in predictions[0]]
-    
-    # Reconstruct words and their labels
-    highlighted_entities = []
-    current_word = ""
-    current_label = None
-    
-    for token, label in zip(tokens, predicted_labels):
-        # Skip special tokens
-        if token in ["[CLS]", "[SEP]", "[PAD]"]:
-            continue
-            
-        # Handle subword tokens (starting with ##)
-        if token.startswith("##"):
-            current_word += token[2:]
-        else:
-            # Save previous word if it exists
-            if current_word:
-                if current_label and current_label != "O":
-                    highlighted_entities.append((current_word, current_label))
-                else:
-                    highlighted_entities.append((current_word, None))
-                current_word = " "  # Add space between words
-            
-            current_word += token
-            current_label = label
-    
-    # Add the last word
-    if current_word.strip():
-        if current_label and current_label != "O":
-            highlighted_entities.append((current_word, current_label))
-        else:
-            highlighted_entities.append((current_word, None))
-    
-    # Create summary
-    detected_entities = {}
-    for word, label in highlighted_entities:
-        if label and label != "O":
-            if label not in detected_entities:
-                detected_entities[label] = []
-            detected_entities[label].append(word.strip())
-    
-    if detected_entities:
-        summary = "**Detected PII:**\n\n"
-        for entity_type, words in detected_entities.items():
-            summary += f"- **{entity_type}**: {', '.join(words)}\n"
-    else:
-        summary = "No PII detected in the text."
-    
-    return highlighted_entities, summary
-
-
-# Example texts for users to try (multilingual)
-examples = [
-    ["My name is John Smith and my email is john.smith@example.com. I was born on January 15, 1985."],
-    ["Please send the payment to IBAN GB29 NWBK 6016 1331 9268 19 or call me at +1-555-123-4567."],
-    ["Mi nombre es María García y vivo en Calle Mayor 123, Madrid. Mi teléfono es +34-91-123-4567."],
-    ["Je m'appelle Pierre Dubois, mon email est pierre.dubois@email.fr et j'habite à Paris."],
-    ["My SSN is 123-45-6789 and my credit card number is 4532-1234-5678-9010. My blood type is O+."],
-    ["车辆识别号: 1HGBH41JXMN109186, 联系电话: +86-138-0013-8000"],
-]
-
-# Create Gradio interface
-with gr.Blocks(title="PII Detection with BERT", theme=gr.themes.Soft()) as demo:
-    gr.Markdown(
-        """
-        # 🌍 Multilingual PII Detector
-        
-        This tool uses a fine-tuned **multilingual BERT model** to automatically detect and highlight personal information in text.
-        It can identify **39 different types** of PII including names, emails, phone numbers, SSN, dates, and more.
-        
-        **Supports multiple languages!** 🌏
-        
-        ### How to use:
-        1. Enter or paste text in the box below (in any supported language)
-        2. Click "Detect PII" to analyze
-        3. View highlighted entities and summary
-        """
-    )
-    
-    with gr.Row():
-        with gr.Column():
-            input_text = gr.Textbox(
-                label="Input Text",
-                placeholder="Enter text to analyze for PII...",
-                lines=6,
-            )
-            detect_btn = gr.Button("🔍 Detect PII", variant="primary")
-        
-        with gr.Column():
-            output_highlighted = gr.HighlightedText(
-                label="Highlighted PII Entities",
-                combine_adjacent=True,
-                color_map=ENTITY_COLORS,
-            )
-            output_summary = gr.Markdown(label="Summary")
-    
-    gr.Markdown("### 📝 Try these examples:")
-    gr.Examples(
-        examples=examples,
-        inputs=input_text,
-    )
-    
-    gr.Markdown(
-        """
-        ### 🏷️ Detectable Entity Types (39 types):
-        
-        **Identity**: NAME, USERNAME, PREFIX, GENDER, AGE, JOB, BLOODTYPE  
-        **Contact**: EMAIL, PHONENUMBER, PHONEIMEI, STREET, ADDRESS, ZIPCODE, GEO, NEARBYGPSCOORDINATE  
-        **Financial**: CREDITCARDNUMBER, CREDITCARDISSUER, IBAN, BIC, ACCOUNTNAME, CURRENCY, COINADDRESS  
-        **Government IDs**: SSN (Social Security Number)  
-        **Vehicle**: VEHICLEVIN, VEHICLEVRM  
-        **Technical**: IP, MAC, URL, PASSWORD  
-        **Organization**: ORG  
-        **Temporal**: DATE, TIME  
-        **Physical**: HEIGHT, WEIGHTS, COLOR  
-        **Other**: NUM, ORDINALDIRECTION, MISC
-        
-        ---
-        **Model**: Multilingual BERT-base fine-tuned for PII detection  
-        **Base Model**: [google-bert/bert-base-multilingual-uncased](https://huggingface.co/google-bert/bert-base-multilingual-uncased)  
-        **Languages**: Supports 100+ languages including English, Spanish, French, German, Chinese, Arabic, and more!
-        """
-    )
-    
-    # Connect the button to the function
-    detect_btn.click(
-        fn=detect_pii,
-        inputs=input_text,
-        outputs=[output_highlighted, output_summary]
-    )
-    
-    # Also trigger on Enter key
-    input_text.submit(
-        fn=detect_pii,
-        inputs=input_text,
-        outputs=[output_highlighted, output_summary]
-    )
-
-# Launch the app
-if __name__ == "__main__":
-    demo.launch()
-