Aditicodes07/CineSense-5-Star ? reverse-engineered prompt

Build me a small Streamlit app called CineSense 5 Star that lets someone paste in a movie review and get back a predicted rating from 1 to 5 stars.

I want it to feel simple and polished, with a clean input box, a big result area, and a confidence score so the user can see how sure the model is. Please include the full NLP pipeline too, with basic text cleaning, stopword removal, lemmatization, TF IDF vectorization, and a Logistic Regression model for the prediction. Add a probability chart for all five star levels so the output feels more interactive and easy to understand.

Also make sure the project can be trained from the included dataset, saves the model and vectorizer, and then loads them in the app. If anything needs current library behavior or setup details, look up the latest docs online and keep it working with the versions in the requirements.