Upload 3 files

Dockerfile

@@ -0,0 +1,29 @@
+# 1. Base Image
+FROM python:3.10-slim
+
+# 2. Set Directory
+WORKDIR /app
+
+# 3. Install System Dependencies (Needed for OpenCV)
+RUN apt-get update && apt-get install -y \
+    libgl1 \
+    libglib2.0-0 \
+    && rm -rf /var/lib/apt/lists/*
+
+# 4. User Setup (Hugging Face Requirement)
+RUN useradd -m -u 1000 user
+USER user
+ENV PATH="/home/user/.local/bin:$PATH"
+
+# 5. Copy Files
+COPY --chown=user . .
+
+# 6. Install Python Libs
+RUN pip install --no-cache-dir --upgrade pip && \
+    pip install --no-cache-dir -r requirements.txt
+
+# 7. Expose Port
+EXPOSE 7860
+
+# 8. Start Server
+CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]

main.py

@@ -0,0 +1,120 @@
+from fastapi import FastAPI, File, UploadFile, HTTPException
+from fastapi.responses import FileResponse
+from fastapi.staticfiles import StaticFiles
+from fastapi.middleware.cors import CORSMiddleware
+import tensorflow as tf
+from tensorflow.keras.models import load_model
+import numpy as np
+import cv2
+import io
+import os
+
+# 1. INITIALIZE FASTAPI
+app = FastAPI()
+
+# 2. CORS (Allow everything for safety)
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+# 3. LOAD MODEL
+MODEL_PATH = "retina_model_deploy.keras" # We use the optimized version
+model = None
+
+@app.on_event("startup")
+async def startup_event():
+    global model
+    if os.path.exists(MODEL_PATH):
+        print("🏥 Loading Medical Model...")
+        model = load_model(MODEL_PATH, compile=False)
+        print("✅ Model Ready.")
+    else:
+        print("❌ Model not found! Upload retina_model_deploy.keras")
+
+# 4. BEN GRAHAM PREPROCESSING (MUST MATCH TRAINING EXACTLY)
+def crop_image_from_gray(img, tol=7):
+    """
+    Crops the black circle borders.
+    """
+    if img.ndim == 2:
+        mask = img > tol
+        return img[np.ix_(mask.any(1),mask.any(0))]
+    elif img.ndim == 3:
+        gray_img = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
+        mask = gray_img > tol
+        
+        check_shape = img[:,:,0][np.ix_(mask.any(1),mask.any(0))].shape[0]
+        if (check_shape == 0): return img
+        else:
+            img1=img[:,:,0][np.ix_(mask.any(1),mask.any(0))]
+            img2=img[:,:,1][np.ix_(mask.any(1),mask.any(0))]
+            img3=img[:,:,2][np.ix_(mask.any(1),mask.any(0))]
+            img = np.stack([img1,img2,img3],axis=-1)
+        return img
+
+def preprocess_eye(image_bytes):
+    """
+    Full Pipeline: Decode -> RGB -> Crop -> Resize -> Gaussian Blur
+    """
+    # Decode bytes to OpenCV Image
+    nparr = np.frombuffer(image_bytes, np.uint8)
+    img = cv2.imdecode(nparr, cv2.IMREAD_COLOR)
+    
+    # Convert BGR (OpenCV) to RGB (TensorFlow)
+    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
+    
+    # 1. Crop
+    img = crop_image_from_gray(img)
+    
+    # 2. Resize (260x260 - Same as Training)
+    img = cv2.resize(img, (260, 260))
+    
+    # 3. Ben Graham Smoothing (The Secret Sauce)
+    # This highlights the veins/hemorrhages
+    img = cv2.addWeighted(img, 4, cv2.GaussianBlur(img, (0,0), 30), -4, 128)
+    
+    # Prepare for Model (Expand dims)
+    # Note: Model has Rescaling(1./255) layer inside, so we send raw pixels
+    img_array = np.expand_dims(img, axis=0)
+    
+    return img_array
+
+# 5. API ENDPOINT
+@app.post("/api/predict")
+async def predict(file: UploadFile = File(...)):
+    if not model:
+        raise HTTPException(status_code=500, detail="Model not loaded")
+    
+    try:
+        contents = await file.read()
+        processed_image = preprocess_eye(contents)
+        
+        prediction = model.predict(processed_image)
+        
+        # Binary Classification (Sigmoid output: 0 to 1)
+        # Closer to 0 = Healthy, Closer to 1 = Disease
+        score = float(prediction[0][0])
+        
+        # Logic: We set threshold at 0.5
+        label = "Disease Detected" if score >= 0.5 else "Healthy Eye"
+        confidence = score if score >= 0.5 else 1 - score
+        
+        return {
+            "diagnosis": label,
+            "confidence": round(confidence * 100, 2),
+            "raw_score": score
+        }
+    except Exception as e:
+        return {"error": str(e)}
+
+# 6. SERVE FRONTEND (The "One Place" Magic)
+# We will serve the React 'dist' folder
+if os.path.exists("dist"):
+    app.mount("/", StaticFiles(directory="dist", html=True), name="static")
+else:
+    @app.get("/")
+    def home():
+        return {"message": "API is running. Upload React 'dist' folder to serve UI."}

requirements.txt

@@ -0,0 +1,7 @@
+fastapi
+uvicorn
+python-multipart
+tensorflow-cpu>=2.16.0
+opencv-python-headless
+numpy
+pillow