Create app.py

app.py

@@ -0,0 +1,80 @@
+import os
+import numpy as np
+import tensorflow as tf
+from tensorflow.keras.models import load_model # type: ignore
+from tensorflow.keras.preprocessing import image # type: ignore
+from tensorflow.keras.applications.densenet import preprocess_input # type: ignore
+
+from fastapi import FastAPI, File, UploadFile, Request
+from fastapi.templating import Jinja2Templates
+from fastapi.responses import HTMLResponse
+import uvicorn
+
+# Initialize FastAPI app
+app = FastAPI()
+
+# Set up template rendering (similar to Flask’s render_template)
+templates = Jinja2Templates(directory="templates")
+
+# Define paths
+BASE_DIR = os.path.dirname(__file__)
+MODEL_PATH = os.path.join(BASE_DIR, 'uploads', "densenet_ship.h5")
+
+# Load the model
+model = load_model(MODEL_PATH)
+
+# Define ship categories
+val_dict = {
+    0: 'Aircraft Carrier',
+    1: 'Bulkers',
+    2: 'Car Carrier',
+    3: 'Container Ship',
+    4: 'Cruise',
+    5: 'DDG',
+    6: 'Recreational',
+    7: 'Sailboat',
+    8: 'Submarine',
+    9: 'Tug'
+}
+
+# Define Routes
+
+@app.get("/", response_class=HTMLResponse)
+async def index(request: Request):
+    return templates.TemplateResponse("index.html", {"request": request})
+
+@app.get("/about", response_class=HTMLResponse)
+async def about(request: Request):
+    return templates.TemplateResponse("about.html", {"request": request})
+
+@app.get("/service", response_class=HTMLResponse)
+async def service(request: Request):
+    return templates.TemplateResponse("service.html", {"request": request})
+
+@app.post("/predict/")
+async def predict(image_file: UploadFile = File(...)):
+    # Save uploaded file
+    file_path = os.path.join(BASE_DIR, 'uploads', image_file.filename)
+    
+    with open(file_path, "wb") as buffer:
+        buffer.write(await image_file.read())
+
+    # Load and preprocess image
+    img = image.load_img(file_path, target_size=(224, 224))
+    img = image.img_to_array(img)
+    img = img.reshape((1, img.shape[0], img.shape[1], img.shape[2]))
+    img = preprocess_input(img)
+
+    # Make prediction
+    pred = model.predict(img)
+    pred = pred.flatten()
+    
+    # Get predicted category
+    predicted_class = val_dict[np.argmax(pred)]
+    
+    # Return result
+    return {"category": predicted_class, "message": f"The Ship Category is {predicted_class}"}
+
+# Run the FastAPI app with uvicorn (needed when not using Docker Spaces)
+if __name__ == "__main__":
+    uvicorn.run(app, host="0.0.0.0", port=7860)