request: use XPath (or class/id) for text version of page instead of raw HTML

The current behavior is that an XPath job returns + compares the raw HTML.
I would very much prefer that the XPath (or CSS class / ID) returns a text version of that section of the page.
Similar to changedetection.io.

E.g. I want to know when wikipedia.org adds a new project:
```
//*[contains(@class, 'other-projects')]
```
or
```
.other-projects
```