Currently, the MediaBiasFactCheck data we are getting from another scraper (https://github.com/drmikecrowe/mbfcext) distinguishes sites into "left", "left-center", "center", "right", and "right-center". However, this scraper is missing MediaBiasFactCheck's extreme-left and extreme-right ratings.

We need to build a new scraper that pulls the extreme-left and extreme-right sources so that we can distinguish them from the rest. This can be done by analyzing the URL of each source's scale image (which includes a rating and number, such as left2 or extremeright3).

This new scraper will also make our script more efficient as we won't be pulling down extra data we don't need.