Update app.py

app.py

@@ -1,31 +1,60 @@
+
 import gradio as gr
 from PIL import Image
+import requests
 import hopsworks
+import joblib
+import pandas as pd
 
 project = hopsworks.login()
 fs = project.get_feature_store()
 
-dataset_api = project.get_dataset_api()
 
-dataset_api.download("Resources/images/latest_wine.png")
-dataset_api.download("Resources/images/actual_wine.png")
-dataset_api.download("Resources/images/df_recent.png")
-dataset_api.download("Resources/images/confusion_matrix.png")
+mr = project.get_model_registry()
+model_red = mr.get_model("wine_red_model", version=2)
+model_dir_red = model_red.download()
+model_red = joblib.load(model_dir + "/wine_red_model.pkl")
+print("Red Model downloaded")
+
+model_white = mr.get_model("wine_white_model", version=2)
+model_dir_white = model_white.download()
+model_white = joblib.load(model_dir_white + "/wine_white_model.pkl")
+print("White Model downloaded")
 
-with gr.Blocks() as demo:
-    with gr.Row():
-      with gr.Column():
-          gr.Label("Today's Predicted Image")
-          input_img = gr.Image("latest_wine.png", elem_id="predicted-img")
-      with gr.Column():          
-          gr.Label("Today's Actual Image")
-          input_img = gr.Image("actual_wine.png", elem_id="actual-img")        
-    with gr.Row():
-      with gr.Column():
-          gr.Label("Recent Prediction History")
-          input_img = gr.Image("df_recent.png", elem_id="recent-predictions")
-      with gr.Column():          
-          gr.Label("Confusion Maxtrix with Historical Prediction Performance")
-          input_img = gr.Image("confusion_matrix.png", elem_id="confusion-matrix")        
+def wine(alcohol, chlorides, density, type, volatil_acidity):
+    print("Calling function")
+    df = pd.DataFrame([[type, alcohol, chlorides, density, volatil_acidity]], 
+                      columns=['type','alcohol','chlorides','density','volatil_acidity','fixed_acidity','citric_acid','total_sulfur_dioxide'])
+    print("Predicting")
+    print(df)
+    
+    if type == "red":
+         res = model_red.predict(df) 
+         print(res)
+         wine_url = "https://raw.githubusercontent.com/Anniyuku/wine_quality/main/" + res[0] + ".png"
+         img = Image.open(requests.get(wine_url, stream=True).raw)
+    else :
+         res = model_white.predict(df) 
+         print(res)
+         wine_url = "https://raw.githubusercontent.com/Anniyuku/wine_quality/main/" + res[0] + ".png"
+         img = Image.open(requests.get(wine_url, stream=True).raw)            
+    return img
+        
+demo = gr.Interface(
+    fn=wine,
+    title="Wine Predictive Analytics",
+    description="Experiment with type, alcohol, chlorides, density, volatil_acidity, fixed_acidity, citric_acid, total_sulfur_dioxide to predict which flower it is.",
+    allow_flagging="never",
+    inputs=[
+        gr.inputs.Radio(choices=['red', 'white'], label='Type'),
+        gr.inputs.Number(default=9.00, label="alcohol"),
+        gr.inputs.Number(default=0.60, label="chlorides"),
+        gr.inputs.Number(default=1.00, label="density"),
+        gr.inputs.Number(default=1.00, label="volatil_acidity"),
+        gr.inputs.Number(default=1.00, label="fixed_acidity"),
+        gr.inputs.Number(default=1.00, label="citric_acid"),
+        gr.inputs.Number(default=1.00, label="total_sulfur_dioxide"),
+        ],
+    outputs=gr.Image(type="pil"))
 
-demo.launch()
+demo.launch(debug=True)