User will be able to provide a url for an article/page that contains a recipe, and the system should be able to return back a recipe entry, or a recipe form filled out for the user to submit/correct.
Will rely on AI to structure the data correctly and return in a json format.
However, we need to provide OpenAI with the text in the html document, as might not me possible for AI to navigate to a url.
I have built a basic scrapper using with cherrio: https://www.npmjs.com/package/cheerio
that returns the main text in the page. However the code might only work for the scenario of Guardian recipes as other pages might have another html structure.
User will be able to provide a url for an article/page that contains a recipe, and the system should be able to return back a recipe entry, or a recipe form filled out for the user to submit/correct.
Will rely on AI to structure the data correctly and return in a json format.
However, we need to provide OpenAI with the text in the html document, as might not me possible for AI to navigate to a url.
I have built a basic scrapper using with cherrio: https://www.npmjs.com/package/cheerio that returns the main text in the page. However the code might only work for the scenario of Guardian recipes as other pages might have another html structure.