Get the latest tech news
Web scraping with GPT-4o: powerful but expensive
tl;dr; show me the demo and source code!
At first, I thought that the parsed Condition column was a hallucination since I did not see that in the website, however, upon inspecting the source code, I realized that those tags exist but are invisible in the table. A simple function that removes all properties except class, id, and data-testid(which are the ones I noticed the generated XPaths were using) trimmed the number of characters in the table by half. This isn’t great UX as it’d be better to ask the user to click on the table they want to extract, and to provide some sample rows so the model can understand the structure a bit better.
Or read this on Hacker News