Update app.py

app.py

@@ -6,7 +6,7 @@ import lightgbm as lgb
 from sentence_transformers import SentenceTransformer
 import numpy as np
 
-# Load files from the model repo "your-username/Essay"
+# Load files from the model repo
 repo_id = "Nawal20/Essay"
 
 ridge_path = hf_hub_download(repo_id=repo_id, filename="ridge_model.pkl")
@@ -22,36 +22,36 @@ encoder = joblib.load(encoder_path)
 with open(metadata_path, "r") as f:
     metadata_columns = json.load(f)
 
-# Load SBERT model (will download at runtime)
+# Load SBERT model
 sbert = SentenceTransformer("sentence-transformers/paraphrase-mpnet-base-v2")
 
-def predict_score(essay_text, gender, race_ethnicity, assignment, prompt_name, disability, disadvantaged, ell_status):
-    # Encode essay
-    essay_embedding = sbert.encode([essay_text])
-    
-    # Prepare metadata as dict
+def predict_score(essay_text, gender, race_ethnicity, assignment, prompt_name, disadvantaged, disability, ell_status):
+    # Encode the essay
+    essay_embedding = sbert.encode([essay_text])  # shape (1, 768)
+
+    # Create metadata dict from input
     metadata_input = {
-        "gender": "female",
-        "race_ethnicity": "Asian",
-        "assignment": "Informative",  # ✅ must be here
-        "prompt_name": "Education Benefits",
-        "economically_disadvantaged": "No",
-        "student_disability_status": "None",
-        "ell_status": "No"
+        "gender": gender,
+        "race_ethnicity": race_ethnicity,
+        "assignment": assignment,
+        "prompt_name": prompt_name,
+        "economically_disadvantaged": disadvantaged,
+        "student_disability_status": disability,
+        "ell_status": ell_status
     }
 
-    # Convert to array in correct order
+    # Create input array based on column order
     metadata_values = [metadata_input[col] for col in metadata_columns]
-    metadata_array = encoder.transform([metadata_values])  # shape: (1, n)
+    metadata_array = encoder.transform([metadata_values])  # shape (1, N)
 
-    # Combine essay embedding + metadata
-    full_input = np.hstack([essay_embedding, metadata_array])
+    # Combine essay + metadata
+    full_input = np.hstack([essay_embedding.reshape(1, -1), metadata_array.toarray()])
 
-    # Predict from both models
+    # Predict scores
     ridge_score = ridge.predict(full_input)[0]
     lgb_score = lgb_model.predict(full_input)[0]
-    #final_score = round((0.5 * ridge_score + 0.5 * lgb_score), 2)
-    final_score = (ridge_pred + lgb_pred) / 2
+    final_score = round((ridge_score + lgb_score) / 2, 2)
+
     return final_score
 
 # Gradio UI
@@ -61,14 +61,14 @@ iface = gr.Interface(
         gr.Textbox(label="Essay Text", lines=10, placeholder="Paste your essay here..."),
         gr.Dropdown(["Male", "Female", "Other"], label="Gender"),
         gr.Dropdown(["Asian", "Black", "Hispanic", "White", "Other"], label="Race/Ethnicity"),
-        gr.Dropdown(["Informative", "Argumentative", "Narrative"], label="Assignment"),  # ✅ ADD
-        gr.Dropdown(["Education Benefits", "Technology Impact", "Climate Change"], label="Prompt Name"),  # ✅ ADD
+        gr.Dropdown(["Informative", "Argumentative", "Narrative"], label="Assignment"),
+        gr.Dropdown(["Education Benefits", "Technology Impact", "Climate Change"], label="Prompt Name"),
         gr.Dropdown(["Yes", "No"], label="Economically Disadvantaged"),
-        gr.Dropdown(["Yes", "No"], label="Student has Disability"),
+        gr.Dropdown(["None", "Learning", "Physical", "Other"], label="Student has Disability"),
         gr.Dropdown(["Yes", "No"], label="ELL Status"),
     ],
     outputs=gr.Number(label="Predicted Essay Score"),
-    title="Automated Essay Scoring App"
+    title="📘 Automated Essay Scoring App"
 )
 
 iface.launch()