Deploy from GitHub Actions

app.py

@@ -1,49 +1,121 @@
+import os
 import json
 import joblib
 import pandas as pd
-import gradio as gr
-import spaces
+import streamlit as st
 from huggingface_hub import hf_hub_download
 
+# Hugging Face Model Repository
 REPO_ID = "Swetha1929/predictive-maintenance-engine-model"
 
-model_path = hf_hub_download(repo_id=REPO_ID, filename="best_model.pkl")
-feature_path = hf_hub_download(repo_id=REPO_ID, filename="feature_names.txt")
+st.set_page_config(
+    page_title="Engine Condition Prediction",
+    layout="wide"
+)
 
-model = joblib.load(model_path)
+st.title("Engine Condition Prediction App")
+st.write("Enter the engine feature values below and click Predict.")
 
-with open(feature_path, "r", encoding="utf-8") as f:
-    feature_names = [line.strip() for line in f if line.strip()]
+# Local backup (used only if HF download fails)
+BASE_DIR = os.path.dirname(os.path.abspath(__file__))
+MODELS_DIR = os.path.join(BASE_DIR, "models")
+
+MODEL_FILE = os.path.join(MODELS_DIR, "best_model.pkl")
+FEATURE_FILE = os.path.join(MODELS_DIR, "feature_names.txt")
+INFO_FILE = os.path.join(MODELS_DIR, "model_info.json")
+
+
+@st.cache_resource
+def load_artifacts():
+    try:
+        model_path = hf_hub_download(
+            repo_id=REPO_ID,
+            filename="best_model.pkl"
+        )
+
+        feature_path = hf_hub_download(
+            repo_id=REPO_ID,
+            filename="feature_names.txt"
+        )
+
+        info_path = hf_hub_download(
+            repo_id=REPO_ID,
+            filename="model_info.json"
+        )
+
+    except Exception:
+        model_path = MODEL_FILE
+        feature_path = FEATURE_FILE
+        info_path = INFO_FILE
+
+    model = joblib.load(model_path)
+
+    with open(feature_path, "r", encoding="utf-8") as f:
+        feature_names = [
+            line.strip()
+            for line in f
+            if line.strip()
+        ]
 
-try:
-    info_path = hf_hub_download(repo_id=REPO_ID, filename="model_info.json")
-    with open(info_path, "r", encoding="utf-8") as f:
-        model_info = json.load(f)
-except Exception:
     model_info = {}
 
-@spaces.GPU
-def predict(*values):
-    input_df = pd.DataFrame([list(values)], columns=feature_names)
+    if os.path.exists(info_path):
+        with open(info_path, "r", encoding="utf-8") as f:
+            model_info = json.load(f)
+
+    return model, feature_names, model_info
+
+
+try:
+    model, feature_names, model_info = load_artifacts()
+
+except Exception as e:
+    st.error(f"Error loading model: {e}")
+    st.stop()
+
+
+if model_info:
+    st.caption(
+        f"Best Model : {model_info.get('best_model_name','Unknown')}"
+    )
+
+    st.caption(
+        f"Test F1 Score : {model_info.get('test_f1','N/A')}"
+    )
+
+
+inputs = {}
+
+cols = st.columns(2)
+
+for i, feature in enumerate(feature_names):
+
+    with cols[i % 2]:
+
+        inputs[feature] = st.number_input(
+            feature,
+            value=0.0,
+            step=1.0,
+            format="%.4f"
+        )
+
+
+if st.button("Predict"):
+
+    input_df = pd.DataFrame([inputs])
+
     prediction = model.predict(input_df)[0]
 
+    st.success(
+        f"Predicted Engine Condition : {prediction}"
+    )
+
     if hasattr(model, "predict_proba"):
-        probs = model.predict_proba(input_df)[0]
-        return str(prediction), {f"Class {i}": float(p) for i, p in enumerate(probs)}
-
-    return str(prediction), {}
-
-inputs = [gr.Number(label=feature) for feature in feature_names]
-
-demo = gr.Interface(
-    fn=predict,
-    inputs=inputs,
-    outputs=[
-        gr.Textbox(label="Predicted Engine Condition"),
-        gr.Label(label="Prediction Probabilities")
-    ],
-    title="Engine Condition Prediction",
-    description="Enter engine parameters to predict the engine condition."
-)
 
-demo.launch()
+        probabilities = model.predict_proba(input_df)
+
+        st.subheader("Prediction Probabilities")
+
+        st.dataframe(
+            pd.DataFrame(probabilities)
+        )