Natashahahahaha/darknet_market_text_classifier ? reverse-engineered prompt

Build me a Python notebook that takes raw darknet marketplace product text and turns it into a text classification pipeline.

I want it to clean messy listing text, lower case it, strip punctuation and noise, remove English stopwords, and also use a few custom stopwords for marketplace words like shipping, price, quality, and listing. Then vectorize the text with TF IDF, cluster the listings into 5 broad groups, and use those cluster based labels to train a Random Forest classifier.

Please include data loading, cleaning, feature creation, train and test split, model training, and a clear evaluation section with accuracy plus a classification report and confusion matrix. If you need to check any current library details, look up the docs online. Make the notebook easy to follow and runnable end to end, using pandas, numpy, scikit learn, and nltk.