Search at edX

Background

There are a number of search implementations within edX, all built upon Elasticsearch. This document collects together best practices learned from these different features.

Search Implementations

The following is a list of features that are currently using search.

Feature	Docs	Comments
Discussions / Forums	Discussion Forum Architecture Notes	ES 1.5 Ruby elasticsearch, elasticsearch-model gems Mongo backed Rake tasks for indexing
Course Content Search	/wiki/spaces/LEARNER/pages/158409575	ES 1.5 edx-search client in edx-platform Mongo backed Reindexing management commands and GUI
Teams	/wiki/spaces/LEARNER/pages/38863429	ES 1.5 edx-search client in edx-platform Sql backed Reindexing management commands
Student Notes	Student Notes Architecture	ES 1.5 Haystack
Course Discovery / Catalog	OpenedX Course Discovery Search Catalog Powered Search	ES 1.5 Haystack
Learner Analytics	/wiki/spaces/AN/pages/44073840	ES 1.5

Search Technologies

There are three main technologies that are used by edX:

Elasticsearch - https://github.com/elastic/elasticsearch
Haystack - http://haystacksearch.org/
- This library simplifies the process of indexing and searching Django models
- Note: it is not designed for use with Mongo (see the Haystack FAQ)
edX Search - https://github.com/edx/edx-search
- This is a home grown library on top of Elasticsearch
- It was implemented in order to support searching across the courseware Mongo database
- Notes on search implementation

Additional Proposed Uses

Bookmarks
- Search + Bookmarks Experience