Zubbi

Fixes
- **Elasticsearch:** If an index_prefix is provided, but empty, it will be ignored.

Fixes
- **Scraper:** Unfortunately, the scraper wasn't aware of the new Elasticsearch
configuration format and the SSL options.

New Features
- **Elasticsearch:** Support SSL options for the Elasticsearch connection. For a
list of available options, take a look at the settings.cfg.example file.

Deprecated
- **Configuration:** Changed the format on how to specify the Elasticsearch
connection in the config file to a single dictionary (like for CONNECTIONS).
The old format (prefixing everything with "ES_") is still supported but will
be removed in future versions. Please see the settings.cfg.example for details
to the new format.

New Features
- **Scraper:** Update repo information in Elasticsearch directly after scraping.
Previously, we updated the information for all scraped repositories in one go to
reduce the amount of requests sent to Elasticsearch. However, this had the drawback,
that none of the repo information was updated if scraping any of the repositories
failed.

Fixes
- **Scraper:** Fix a bug were the scraper was still trying to check out GitHub
repositories, although it didn't have a valid access token.
- **Scraper:** Don't fail when trying to split the owner from an invalid GitHub
repository names. Actually, the wrong name comes from a bug in the tenant scraper
which should be fixed in a future release. But for now, it's a good idea to make
this part more robust.

Backwards incompatible changes
- **Elasticsearch:** Zubbi 2.x.x is only compatible with Elasticsearch major version 7.

New Features
- **Experimental:** You can set an `ES_INDEX_PREFIX` in the `settings.cfg` file to
prefix all Elasticsearch indices with a custom value. This could be useful to avoid
name clashes if indices with the same name are already used by another part of the
system.

New Features
- **UI:** Autofocus the search field on index page. You can now visit Zubbi and
directly start typing.
- **UI:** Show "last update" timestamps for jobs and roles in search results.
- **Extensions:** Allow custom tabs/contents on details page. When extending Zubbi,
someone can now add new tabs (in addition to the already existing 'Description'
and 'Changelog').

Fixes
- **UI:** Use rendered description in search result cards to hide unparsable
Sphinx links.