Iterate over nodes to build up text content instead of using [node-html-parser](https://github.com/taoqf/node-html-parser)'s `structuredText` to capture lists and breaks better.