KalyanM45/Spam-Email-Detection ? reverse-engineered prompt

Build me a spam email classification app in Python that can tell Spam from Ham with a clean web interface.

I want a Streamlit app where I can paste in one email and तुरंत get a prediction with a confidence score, and also upload an mbox file to check a bunch of emails at once and download the results as a CSV. Please make it use a proper training pipeline too, so I can retrain it on a dataset in the data folder and save the trained model and vectorizer into outputs. It should compare a few common machine learning models, like SVM, Logistic Regression, Decision Tree, and Random Forest, then pick the best one for inference.

Please include basic logging and show useful metrics like precision, recall, and F1 score during training. Keep the project organized so the training and prediction code are separated cleanly, and if you need to look up current docs online, feel free to do that.