Add application file

app.py

@@ -0,0 +1,241 @@
+import os
+import torch
+import time
+import gradio as gr
+import requests
+from transformers import AutoProcessor, PaliGemmaForConditionalGeneration
+from PIL import Image
+from io import BytesIO
+import logging
+
+# Configure logging
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+# Global variables for model
+model = None
+processor = None
+device = None
+
+def load_model():
+    """Load the AI model once at startup"""
+    global model, processor, device
+    
+    logger.info("Loading AI model...")
+    
+    # Get Hugging Face token from environment
+    hf_token = os.environ.get('HF_TOKEN')
+    
+    model_id = "mychen76/paligemma-3b-mix-448-med_30k-ct-brain"
+    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+    dtype = torch.float16 if torch.cuda.is_available() else torch.float32
+    
+    logger.info(f"Using device: {device}")
+    
+    try:
+        # Load processor and model with authentication token
+        processor = AutoProcessor.from_pretrained(
+            model_id, 
+            token=hf_token
+        )
+        model = PaliGemmaForConditionalGeneration.from_pretrained(
+            model_id, 
+            torch_dtype=dtype,
+            token=hf_token
+        ).to(device).eval()
+        
+        logger.info("Model loaded successfully!")
+        return True
+        
+    except Exception as e:
+        logger.error(f"Error loading model: {e}")
+        return False
+
+def analyze_brain_scan(image, patient_name="", patient_age="", symptoms=""):
+    """Analyze brain scan image and return medical findings"""
+    try:
+        if image is None:
+            return "Please upload a brain scan image."
+        
+        # Convert to PIL Image if needed
+        if not isinstance(image, Image.Image):
+            image = Image.fromarray(image).convert("RGB")
+        
+        # Run AI inference
+        prompt = "<image> Findings:"
+        inputs = processor(
+            images=image, 
+            text=prompt, 
+            return_tensors="pt"
+        ).to(device, dtype=torch.float16 if torch.cuda.is_available() else torch.float32)
+        
+        with torch.no_grad():
+            generated_ids = model.generate(**inputs, max_new_tokens=100)
+            
+        result = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
+        
+        # Clean up the result
+        if result.startswith(prompt):
+            result = result[len(prompt):].strip()
+        
+        # Format the response
+        timestamp = time.strftime("%Y-%m-%d %H:%M:%S")
+        
+        formatted_result = f"""
+## Brain CT Analysis Results
+
+**Patient Information:**
+- Name: {patient_name or 'Not provided'}
+- Age: {patient_age or 'Not provided'}
+- Symptoms: {symptoms or 'Not provided'}
+- Analysis Date: {timestamp}
+
+**AI Findings:**
+{result}
+
+**Note:** This is an AI-generated analysis for educational purposes only. 
+Always consult with qualified medical professionals for actual diagnosis.
+"""
+        
+        return formatted_result
+        
+    except Exception as e:
+        logger.error(f"Analysis error: {e}")
+        return f"Error during analysis: {str(e)}"
+
+def create_api_response(image, patient_name="", patient_age="", symptoms=""):
+    """Create API-compatible response for integration"""
+    try:
+        if image is None:
+            return {"error": "No image provided"}
+        
+        # Convert to PIL Image if needed
+        if not isinstance(image, Image.Image):
+            image = Image.fromarray(image).convert("RGB")
+        
+        # Run AI inference
+        prompt = "<image> Findings:"
+        inputs = processor(
+            images=image, 
+            text=prompt, 
+            return_tensors="pt"
+        ).to(device, dtype=torch.float16 if torch.cuda.is_available() else torch.float32)
+        
+        with torch.no_grad():
+            generated_ids = model.generate(**inputs, max_new_tokens=100)
+            
+        result = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
+        
+        # Clean up the result
+        if result.startswith(prompt):
+            result = result[len(prompt):].strip()
+        
+        # Create API response
+        response = {
+            "prediction": result,
+            "timestamp": time.strftime("%Y-%m-%d %H:%M:%S"),
+            "patient_info": {
+                "name": patient_name,
+                "age": patient_age,
+                "symptoms": symptoms
+            },
+            "model_info": {
+                "model_id": "mychen76/paligemma-3b-mix-448-med_30k-ct-brain",
+                "device": str(device)
+            }
+        }
+        
+        return response
+        
+    except Exception as e:
+        logger.error(f"API error: {e}")
+        return {"error": f"Analysis failed: {str(e)}"}
+
+# Load model at startup
+logger.info("Initializing Brain CT Analyzer...")
+if load_model():
+    logger.info("Model loaded successfully!")
+else:
+    logger.error("Failed to load model!")
+
+# Create Gradio interface
+with gr.Blocks(title="Brain CT Analyzer", theme=gr.themes.Soft()) as demo:
+    gr.Markdown("""
+    # 🧠 Brain CT Analyzer
+    
+    Upload a brain CT scan image for AI-powered analysis. This tool uses the PaliGemma medical model 
+    to provide preliminary findings.
+    
+    **⚠️ Important:** This is for educational/research purposes only. Always consult qualified medical professionals.
+    """)
+    
+    with gr.Row():
+        with gr.Column(scale=1):
+            image_input = gr.Image(
+                label="Upload Brain CT Scan",
+                type="pil",
+                height=400
+            )
+            
+            with gr.Group():
+                patient_name = gr.Textbox(
+                    label="Patient Name (Optional)",
+                    placeholder="Enter patient name"
+                )
+                patient_age = gr.Textbox(
+                    label="Patient Age (Optional)",
+                    placeholder="Enter patient age"
+                )
+                symptoms = gr.Textbox(
+                    label="Symptoms (Optional)",
+                    placeholder="Describe symptoms",
+                    lines=3
+                )
+            
+            analyze_btn = gr.Button(
+                "🔍 Analyze Brain Scan",
+                variant="primary",
+                size="lg"
+            )
+        
+        with gr.Column(scale=1):
+            result_output = gr.Markdown(
+                label="Analysis Results",
+                value="Upload an image and click 'Analyze Brain Scan' to see results."
+            )
+    
+    # API endpoint simulation
+    with gr.Accordion("🔌 API Response (for developers)", open=False):
+        api_output = gr.JSON(label="API Response Format")
+    
+    # Event handlers
+    analyze_btn.click(
+        fn=analyze_brain_scan,
+        inputs=[image_input, patient_name, patient_age, symptoms],
+        outputs=result_output
+    )
+    
+    analyze_btn.click(
+        fn=create_api_response,
+        inputs=[image_input, patient_name, patient_age, symptoms],
+        outputs=api_output
+    )
+    
+    # Example images (if you have any)
+    gr.Markdown("""
+    ## 📋 Usage Instructions:
+    1. Upload a brain CT scan image (JPEG or PNG)
+    2. Optionally fill in patient information
+    3. Click "Analyze Brain Scan" to get AI findings
+    4. Review the results in the output panel
+    
+    ## 🔗 Integration:
+    This interface can be integrated with your medical app using the Gradio API.
+    """)
+
+if __name__ == "__main__":
+    demo.launch(
+        server_name="0.0.0.0",
+        server_port=7860,
+        share=True
+    )