Deploy KidneyDL CT Scan Classifier

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+artifacts/training/model.keras filter=lfs diff=lfs merge=lfs -text

app.py

@@ -1,4 +1,5 @@
 import os
+import traceback
 from flask import Flask, request, jsonify, render_template
 from flask_cors import CORS
 from cnnClassifier.pipeline.prediction import PredictionPipeline
@@ -9,12 +10,24 @@ CORS(app)
 UPLOAD_FOLDER = "uploads"
 os.makedirs(UPLOAD_FOLDER, exist_ok=True)
 
+MODEL_PATH = os.path.join("artifacts", "training", "model.h5")
+
 
 @app.route("/", methods=["GET"])
 def home():
     return render_template("index.html")
 
 
+@app.route("/health", methods=["GET"])
+def health():
+    model_exists = os.path.isfile(MODEL_PATH)
+    return jsonify({
+        "status": "ok" if model_exists else "degraded",
+        "model_found": model_exists,
+        "model_path": os.path.abspath(MODEL_PATH),
+    }), 200 if model_exists else 503
+
+
 @app.route("/train", methods=["GET", "POST"])
 def train():
     os.system("python main.py")
@@ -31,12 +44,17 @@ def predict():
         return jsonify({"error": "No file selected"}), 400
 
     filepath = os.path.join(UPLOAD_FOLDER, file.filename)
-    file.save(filepath)
-
-    pipeline = PredictionPipeline(filepath)
-    result = pipeline.predict()
-
-    return jsonify(result)
+    try:
+        file.save(filepath)
+        pipeline = PredictionPipeline(filepath)
+        result = pipeline.predict()
+        return jsonify(result)
+    except Exception as e:
+        traceback.print_exc()
+        return jsonify({"error": str(e)}), 500
+    finally:
+        if os.path.exists(filepath):
+            os.remove(filepath)
 
 
 if __name__ == "__main__":

artifacts/training/model.keras

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e92b2657c65573f928f9f5f20f63db4001cc93d8ae1ed58f5f396918b0a2ebda
+size 59141294

requirements.txt

@@ -1,5 +1,5 @@
-tensorflow
-keras
+tensorflow==2.20.0
+keras==3.13.2
 dvc
 numpy
 pandas

resave_model.py

@@ -0,0 +1,21 @@
+"""
+One-time script: resave model.h5 as model.keras (Keras 3 native format).
+This ensures the deployed model loads correctly regardless of minor Keras version
+differences between training and serving environments.
+
+Run once:
+    python resave_model.py
+"""
+
+from tensorflow.keras.models import load_model
+
+h5_path    = "artifacts/training/model.h5"
+keras_path = "artifacts/training/model.keras"
+
+print(f"Loading  {h5_path} ...")
+model = load_model(h5_path)
+
+print(f"Saving   {keras_path} ...")
+model.save(keras_path)
+
+print("Done. Deploy with:  python deploy_to_hf.py")

src/cnnClassifier/pipeline/prediction.py

@@ -9,7 +9,10 @@ class PredictionPipeline:
         self.filename = filename
 
     def predict(self):
-        model = load_model(os.path.join("artifacts", "training", "model.h5"))
+        keras_path = os.path.join("artifacts", "training", "model.keras")
+        h5_path    = os.path.join("artifacts", "training", "model.h5")
+        model_path = keras_path if os.path.isfile(keras_path) else h5_path
+        model = load_model(model_path)
 
         img = image.load_img(self.filename, target_size=(224, 224))
         img_array = image.img_to_array(img)

templates/index.html

@@ -670,6 +670,7 @@
       try {
         const res  = await fetch('/predict', { method: 'POST', body: fd });
         const data = await res.json();
+        if (!res.ok) throw new Error(data.error || `Server error ${res.status}`);
         const pred = data[0]?.image || 'Unknown';
 
         const resultEl = document.getElementById('result');
@@ -695,8 +696,8 @@
         document.getElementById('confPct').textContent  = conf + '%';
         resultEl.style.display = 'block';
 
-      } catch {
-        alert('Something went wrong during analysis. Please try again.');
+      } catch (err) {
+        alert('Something went wrong during analysis.\n\n' + (err?.message || err));
       } finally {
         document.getElementById('loading').style.display = 'none';
         document.getElementById('predictBtn').disabled   = false;