An Algolia backend for Haystack 2.x. Use Algolia service to index and search your django models.
First, go to https://www.algolia.com/ and create an account. Then create an index, you'll need it later.
Install haystack-algolia from github using pip:
pip install -e git+https://github.com/matagus/haystack-algolia.git#egg=dev
Then add to your settings.py:
HAYSTACK_CONNECTIONS = {
'default': {
'ENGINE': 'haystack_algolia.algolia_backend.AlgoliaEngine',
'APP_ID': '<YOUR APP ID>',
'API_KEY': '<YOUR API KEY>',
'INDEX_NAME_PREFIX': 'production-',
'TIMEOUT': 60 * 5
}
}Now you can use haystack as usual. Notice so far it only provides a few features:
- Indexing your models for first time. Each model instances get indexed in a separate Algolia's index in order to use certain Algolia optimizations
- Updating indexes
- Clearing indexes
- Removing one object from its index
- Searching models
Haystack features:
- filtering by dates, numbers or tags: https://www.algolia.com/doc/guides/python#Filtering
- narrowing results
- faceting
- more_like_this
- highlighting
- geo-search
- reindexing without data-loss using a new index & then renaming it
In general:
- tests
Notice that algolia.com does not support dinamically defined (per query) sort order, so using .order_by() in a haystack SearchQuerySet won't make any difference.
Also, filtering works in a pretty different way than in other haystack backends: if you do SearchQuerySet().models(MyModel).filter(field1="foo", field2="bar") it this backend will search for "foo bar" no matter in what fields it appears. I'll be working in providing a more accurate filtering feature using Algolia's filtering API.