imanjani/NLP ? reverse-engineered prompt

Build me a simple NLP project that shows a few text based machine learning demos in one place.

I want one clean notebook or small app that can do fake news classification with a bidirectional LSTM, spam detection for SMS or email messages, and a text generation example from a book using an LSTM model. It should let me see the data prep, training, and a basic result for each one, with clear cells or sections so it is easy to follow.

For the text generation part, split the input text into short sequences and train the model to predict the next character. For the spam and fake news parts, make it easy to enter a message or article and get a spam or real result. Keep the output simple and readable, and add a few visualizations if they help explain what is happening.

Use current best practices and look up any needed docs online if you need to.