add feature that clusters similar inputs in /review table

[dmalan]

Rather than sort the HTML table at /review alphabetically, ideal if we can analyze the script column and cluster similar values (as via fuzzy matching) near each other so we can review unmatched inputs more efficiently.

CC @brianyu28