feh-assets-json VS fehbuilder

so ik R has packages and native functions to help bypass this manual process. Eg scraping the wiki / gamepress unit list with Rvest may prove easier, furthermore you can specify web based sources when reading data. I'm not giga familiar with doing either myself, but maybe you can scrape data from the wikis or from repositories like the feh assets 1. But if youre able to set up a simple R script to read in new data and transform / clean it and save manual updates every 2 weeks

The builder works by obtaining all data from HertzDevil's dumps at https://github.com/HertzDevil/feh-assets-json and then only queries the wiki to grab images. As such is entirely possible that all new data may appear on the builder but requests will load without the correct art until the wiki is correctly updated and another rebuild cycle of 4 hours has passed.

The unit builder gets all the data from HertzDevil datamines and then downloads hero art and skills icons from the Fandom wiki.

Nope, HertzDevil is the first one to figure out this thing, but his JSON repoistory is not very easy to read for most people.

Micro-correction, it originally was like that, nowadays it uses HertzDevil's datamine dumps which he 99% always releases as soon as the update drops. https://github.com/HertzDevil/feh-assets-json

Source: https://github.com/HertzDevil/feh-assets-json

I think you should try Cargo API. Cargo is an MediaWiki extension which were used to store and query data. In this case, it's a better way than HTML parsing for retrieving data from the wiki. Unless you purge the page cache, the wiki will serve outdated pages to you. There is a hack in your code, which pretty much shows how unreliable HTML parsing is.