Improve the way to add a service and crawl its documents

[michielbdejong]

Given a domain name, we could have a process such that:

• edit.tosdr.org calls a webhook on a Heroku-hosted instance of https://github.com/ambanum/CGUs (our new Node.js-based evolution of tosback)
• edit.tosdr.org and the crawler share the same postgres db instance
• it will find links on the home page (footer) and ask the reviewer to pick the right link, by anchor text.
• it will try to group service domains into services by looking at domain name similarity (e.g. only extension is different, like in google.*), and also if two services have the same set of documents.
• In case of doubt, we do create one service per domain, but there needs to be at least one document at the service's own domain.