This issue is a minor bugfix update that fixes an issue with websites that have multiple ld+json metadata blobs.

This change adds support for scraping:

* json+ld schema.org metadata
* DCMI metadata
* the unofficial metadata used by some journals

This hugely improves scraping accuracy on most modern news articles, worldcat books and academic journal articles.

Other changes include:

* switch to the generic {{citation}} template to better support journals and books as well as websites
* add support for new fields: issn, doi, volume, issue, pmid, publication-place and others
* add automated scrape tests for more common websites
* persist the removal of deleted keys from the template
* add in-browser tests for parsing behaviour

* Add support for directly citing an archive.org page
* Strip out non-breaking spaces and other weird unicode whitespace
* Add new scraping capabilities for newspapers on Papers Past
* Add new scraping capabilities for books on WorldCat
* Fix issue causing extension to crash on non-HTML pages (such as PDFs)