Rather than converting the Mapnik XML to a JS object and back to XML for each step of filtering/modifying it (since there are a handful of steps) this changes to converting it at the beginning and having the filters and processing functions pass around and modify the object. Then it converts back to XML at the end.
This also switches the base map on the demo pages to Carto Positron, since the one they were using has disappeared.
This is primarily a simplification, though it may serve as an optimization as well. I haven't measured, and I doubt the conversions were taking huge amounts of time, but it can't be slower and it might be faster with only the one round trip.
Overview
Rather than converting the Mapnik XML to a JS object and back to XML for each step of filtering/modifying it (since there are a handful of steps) this changes to converting it at the beginning and having the filters and processing functions pass around and modify the object. Then it converts back to XML at the end.
This also switches the base map on the demo pages to Carto Positron, since the one they were using has disappeared.
Notes
Testing Instructions
Tests and demos should work the same.
Connects #129