TF-IDF

Suppose we search a movie dataset for cyborg bear and score the following documents.

Document 1: "The Terminator – A cyborg from the future"

• cyborg: TF = 1, IDF = 3.9 → TF-IDF = 1 × 3.9 = 3.9
• bear: TF = 0, IDF = 0.05 → TF-IDF = 0 × 0.05 = 0
• Total score: 3.9

Document 2: "Ted – A talking bear who loves honey and bear friends"

• cyborg: TF = 0, IDF = 3.9 → TF-IDF = 0 × 3.9 = 0
• bear: TF = 2, IDF = 0.05 → TF-IDF = 2 × 0.05 = 0.1
• Total score: 0.1

Document 3: "Cyborg Bear – A robotic bear saves the city"

• cyborg: TF = 1, IDF = 3.9 → TF-IDF = 1 × 3.9 = 3.9
• bear: TF = 2, IDF = 0.05 → TF-IDF = 2 × 0.05 = 0.1
• Total score: 4.0

Final Rankings:

1. Document 3 ("Cyborg Bear") - 4.0 - has both terms.
2. Document 1 ("The Terminator") - 3.9 - has the rare cyborg term.
3. Document 2 ("Ted") - 0.1 - has only the common bear term.

The movie that's actually about a cyborg bear ranks highest, and cyborg matters more than bear in a dataset where bear is common.

1. Add a tfidf command to your keyword_search_cli.py entrypoint script:
   1. Take a document ID and term as arguments.
   2. Load the index and docmap from disk using the .load() method you implemented previously.
   3. Calculate the TF-IDF score for the given document and term.
   4. Print the TF-IDF value.
2. Format the floating-point value to 2 decimal places when printing:

   print(f"TF-IDF score of '{args.term}' in document '{args.doc_id}': {tf_idf:.2f}")
   

Run and submit the CLI tests.