RuleFit: an open model rewrites messy product titles, and plain code decides if they pass
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle What I Built RuleFit cleans up messy product titles so they follow a store platform's rules. You upload a CSV, pick the title column, and download a CSV with a rewritten title, a status (PASS, FIXED or FLAGGED) and the reasons for any failure. The first rule pack is Google Shopping titles. Google's title docs…
RuleFit is an open-source AI model that simplifies messy e-commerce product titles to align with store platform guidelines. Users upload a CSV file containing titles, then receive a new CSV with rewritten titles, a status (PASS, FIXED or FLAGGED), and reasons for any failures. The first rule pack focuses on Google Shopping titles, which must adhere to a 150-character limit and avoid promotional text or all-caps titles.
The open model generates rewrite suggestions, which are then checked using plain Python code. If a title fails the validation, it undergoes one retry using the specific failure reasons. If it still fails, it is flagged, and the original title is kept. RuleFit uses a dual-engine architecture, with an open language model hosted by Tinker generating rewrite suggestions and deterministic Python validators enforcing compliance.
The model's code is released under the MIT license and includes 67 automated tests. The author built the application using Flask and Antigravity (Gemini), and it is not currently deployed on any live platform.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
