This repository contains some basic, initial workings on haystack.
Haystack-test-2 contains a more in depth example with multiple types of records and different types of searching: https://github.com/cedadev/haystack-test-2
Haystack-moles-test contains a lot of moles style records and go more in depth about indexing: https://github.com/cedadev/haystack-moles-test
Haystack does not officially support ElasticSearch above 2.x I found this stated in the Haystack documentation (here: http://django-haystack.readthedocs.io/en/master/installing_search_engines.html) under the ElasticSearch section.
The unofficial fork of the Haystack repository that does support ElasticSearch 5.x (see: https://github.com/CraveFood/django-haystack-elasticsearch).
Getting this to work properly requires the regular Haystack installation as well.
The documentation for Haystack recommends using an include in our urls.py. This saves us the trouble of setting up particular views for the search related elements, however it does mean that Haystack defaults to searching for templates in the virtualenv that we are currently using rather than the Django app we have created. While Haystack does this for the templating it still relies on the standard Django method of migrating models and storing the searchIndexes in the app directory of the project.
All required components are specified in requirements.txt and can be installed with a single command pip install command (see below).
Set up the virtualenv:
virtualenv venv
source venv/bin/activate
pip install -r requirements.txt
Create the settings_local.py file:
cd testsite/
cp testsite/settings_local.py.tmpl testsite/settings_local.py
And edit the file to give the:
- Details of the ElasticSearch service and index.
- Location of the Sqlite database. This must be a file on the local (not shared) file system.
Now create the database:
export DJANGO_SETTINGS_MODULE=testsite.settings
python manage.py migrate
Now populate the database:
python create_records.py
Now index the records in ElasticSearch:
python manage.py rebuild_index
And run the server:
python manage.py runserver
Now the UI is viewable at: