Update app.py

app.py

@@ -1,36 +1,57 @@
 import gradio as gr
-import joblib, pandas as pd
+import joblib
+import pandas as pd
+import numpy as np
 
-# Load your trained pipeline
-model = joblib.load("model.joblib")   # or "models/model.joblib"
+# Load the trained pipeline
+model = joblib.load("models/model.joblib")
 
-def predict(**kwargs):
-    # Build a single-row DataFrame from named inputs
-    df = pd.DataFrame([kwargs])
-    score = model.predict(df)[0]
-    return float(score)
+# Load team→specialty mapping from your dataset
+df = pd.read_csv("minimal_messy_task_performance.csv")
+team_specialty = df[['Team','Specialty']].drop_duplicates().set_index('Team')['Specialty'].to_dict()
+teams = list(team_specialty.keys())
 
-# Expose UI controls matching your features
-iface = gr.Interface(
-    fn=predict,
-    inputs=[
-      gr.Textbox(label="Team"),
-      gr.Textbox(label="Specialty"),
-      gr.Textbox(label="ProductType"),
-      gr.Textbox(label="TaskType"),
-      gr.Number(label="OrderQuantity"),
-      gr.Number(label="DeadlineDays"),
-      gr.Number(label="ExperienceYears"),
-      gr.Number(label="AvgTaskTime_Minutes"),
-      gr.Number(label="ErrorRate"),
-      gr.Number(label="TrainingHours"),
-      gr.Number(label="DayNumber"),
-      gr.Number(label="ThroughputRate"),
-      gr.Number(label="TimePressure"),
-      gr.Number(label="PriorityLevel"),
-    ],
-    outputs=gr.Number(label="Predicted Productivity")
-)
+# Define the static task context
+new_task_base = {
+    'ProductType':             'Mothball',
+    'TaskType':                'Packaging',
+    'OrderQuantity':           100,
+    'DeadlineDays':            3,
+    'ExperienceYears':         5,
+    'AvgTaskTime_Minutes':     25.4,
+    'ErrorRate':               0.05,
+    'TrainingHours':           20.0,
+    'DayNumber':               2,
+    'ThroughputRate':          100 / 25.4,
+    'TimePressure':            100 / (3 * 25.4),
+    'PriorityLevel':           3
+}
+
+def generate_rankings():
+    # Build one row per team, filling in Team and Specialty
+    rows = []
+    for team in teams:
+        row = new_task_base.copy()
+        row['Team'] = team
+        row['Specialty'] = team_specialty[team]
+        rows.append(row)
+    df_tasks = pd.DataFrame(rows)
+    
+    # Predict productivity for every team
+    preds = model.predict(df_tasks)
+    df_tasks['PredictedProductivity'] = np.round(preds, 2)
+    
+    # Return a sorted DataFrame
+    return df_tasks[['Team', 'PredictedProductivity']].sort_values(
+        'PredictedProductivity', ascending=False
+    ).reset_index(drop=True)
+
+# Gradio interface: single button → ranked table
+with gr.Blocks() as demo:
+    gr.Markdown("## Team Productivity Ranking for Today's Task")
+    btn = gr.Button("Generate Rankings")
+    table = gr.Dataframe(headers=["Team", "PredictedProductivity"])
+    btn.click(fn=generate_rankings, inputs=None, outputs=table)
+    
+demo.launch()
 
-if __name__ == "__main__":
-    iface.launch()