import numpy as np import gradio as gr import pickle from sklearn.feature_extraction.text import TfidfVectorizer # Load the trained model (same as before) import joblib # Load the trained model (if saved using joblib) def load_model(): model = joblib.load("random_forest_model.pkl") return model # Prediction function def predict_project(fname, lname, email, role, experience, skills): # Load the model and vectorizer model = load_model() vectorizer=TfidfVectorizer() # Prepare the new employee data new_employee = { 'fname': fname, 'lname': lname, 'email': email, 'role': role, # Role can be 'Senior' or 'Intern' 'experience': experience, 'skills': skills, } # Encode the 'role' to numerical format role_mapping = {'Intern': 0, 'Senior': 1} # Map 'Intern' -> 0, 'Senior' -> 1 new_employee['role'] = role_mapping.get(new_employee['role'], -1) # Default to -1 if unknown role # Transform the skills using the same vectorizer used during training skills_vector = vectorizer.transform([new_employee['skills']]) # Combine the features (experience, role, and skills) X_new_employee = np.hstack(( np.array([[new_employee['experience']]]), # Reshape experience to 2D np.array([[new_employee['role']]]), # Reshape role to 2D skills_vector.toarray() # Convert skills vector to array )) # Predict the project ID for the new employee predicted_project = model.predict(X_new_employee)[0] return f"The predicted project ID for {new_employee['fname']} {new_employee['lname']} is {predicted_project}" # Create the Gradio interface interface = gr.Interface( fn=predict_project, inputs=[ gr.Textbox(label="First Name"), gr.Textbox(label="Last Name"), gr.Textbox(label="Email"), gr.Dropdown(choices=["Intern", "Senior"], label="Role"), gr.Slider(minimum=0, maximum=50, step=1, label="Experience (in years)"), gr.Textbox(label="Skills (comma separated)"), ], outputs="text", ) # Launch the Gradio app interface.launch()