django-oscar-pg-search 0.9.0

Pure Postgresql search backend for Django Oscar

Postgresql search handler for Django-Oscar

This creates a search handler without the need of any search backend. It is designed for the e-commerce framework Oscar.

It is implemented a little bit expensive but uses 4 annotated search vectors: * upc * title * meta_description * meta_title

This way the search can be manipulated through the meta fields. It is running productive in a heavily customized env for many months now. I think it should scale up to 5000 Products with 10 Attributes depending on how the products are loaded. We use it fully lazy with endless scrolling.

To-Do

• Provide a generic way to use filter forms
• Writing Tests

Features

• Don't need to use some additional search backend like elastic
• Creates filters (facets) for:
  • Data that is directly attached to the Product model
  • including foreign key choices
  • AttributeValues of the products
  • StockRecord entries

Installation

Install using pip:

pip install django-oscar-pg-search

# settings.py

# replace 'oscar.apps.search.apps.SearchConfig' by 'oscar_pg_search.apps.PgSearchConfig'
INSTALLED_APPS = [
	# 'oscar.apps.search.apps.SearchConfig',
    'oscar_pg_search.apps.PgSearchConfig',
]

OSCAR_PRODUCT_SEARCH_HANDLER = 'oscar_pg_search.postgres_search_handler.PostgresSearchHandler'

# To avoid exception, haystack is not used:
HAYSTACK_CONNECTIONS = {"default": {}}

Trigram search is our search algorithm. Enable it at your database by executing the following sql:

CREATE EXTENSION pg_trgm;

Settings

If you want to add some fields that are directly attached to the Product model:

# settings.py
OSCAR_ATTACHED_PRODUCT_FIELDS = ['is_public', 'deposit', 'volume', 'weight',]