Lemmatization Comparison Tool for spaCy and NLTK

Reducing words to their dictionary base forms, mapping ran and running to run, is a core normalization step in search, text analysis, and machine learning pipelines, but different NLP libraries produce meaningfully different results. Learners who treat lemmatization as a single well-defined operation get surprised when switching libraries silently changes their features. The intern builds a Python command-line application that lemmatizes input text using both major approaches side by side: NLTK's WordNet lemmatizer, which depends on part-of-speech hints supplied for each word, and spaCy's pipeline, which infers part of speech from sentence context automatically. The tool accepts text from arguments or files, processes it through both libraries, and prints an aligned comparison showing each original token beside its two lemmas, flagging the cases where the libraries disagree. Those disagreements become the teaching moments, revealing how context changes what the correct base form even is. Through the project the intern learns what lemmatization actually involves beneath the API call, why part-of-speech tagging matters to normalization quality, and how to evaluate competing NLP libraries empirically. They finish with a tidy command-line utility and the habit of verifying tool behavior rather than assuming it.