How does indexing work

We use haystack, this is an API that allows one to index Django models using various indexing engines.

Indexing is controlled by SearchIndex instances, we use single index named: ContentIndex.

Indexing in dev environment

In dev environment we use Whoosh index, it is completely inside docker container. Whoosh is a python only library.

To index your data you’ll need to call: ./manage.py rebuild_index.

Indexing in prod environment

First you’ll need to set up elastic search index service. Indexing should be done outside of request context, so you’ll need to call ./manage.py update_index --age HOURS, it will update index for ContentItems updated less than HOURS ago.

How to add new fields

How to add field searchable by full text search

Haystack uses document indexing, so you’ll need to add this field to contentitem_text.txt template. This template will be rendered and indexed.

How to add field searchable as an extra field

If you want to add filtering capabilities, like filter by license, you’ll need to add additional field to ContentIndex.

How to add ranked field

Ranks are a way of saying: “this field is more important”. To add a ranked field you’ll need to just add a “rank” attribute to field.

If this field is also a part of “full text search” (or is indexed inside document) you’ll need to add some magic inside search form, like here: http://django-haystack.readthedocs.io/en/v2.4.1/boost.html#field-boost