Chapter Contents:
EdX Search is a Django application that provides access to search services from within edX Platform applications. Searching is accomplished by creating an index of documents, and then searching within that index for matching information.
By default, edX Search uses MockSearchEngine for testing and ElasticSearch Engine for production. You can configure edX Search to use a different search engine.
MockSearchEngine is a simple implementation using a JSON file for index storage. It has no specific requirements, but it does not scale well and should only be used for testing.
ElasticSearchEngine is a ElasticSearch back-end implementation. It uses same ElasticSearch version that is already part of edX Platform. The current version is v0.90.13, and Django Elasticsearch is 0.4.5.
EdX Search requires the following applications.
EdX Search is included in edX Platform Github requirements and is installed automatically when you install edX Platform.
For existing installations, you must install edX Search manually.
To install edX Search, make sure you are logged to your server as the
edxapp user and are located in edx-platform directory.
If you are not, run the following before continuing:
sudo su edxapp -s /bin/bash cd ~source edxapp_env
Then install edX Search using one of the three following options.
Add the Github link to edx-search to the requirements/edx/github.txt file.
-e git+https://github.com/edx/edx-search.git@ae459ead41962c656ce794619f58cdae46eb7896#egg=edx-search
Then reinstall Github requirements.
pip install -r requirements/edx/github.txt
Checkout the edx-search Github repository.
Then in the edx-search directory, run the following command.
pip install -e ./
Run pip with a Github link.
pip install -e git+https://github.com/edx/edx-search.
git@ae459ead41962c656ce794619f58cdae46eb7896
You enable course indexing by setting the ENABLE_COURSEWARE_INDEX flag.
You enable library indexing by setting the ENABLE_LIBRARY_INDEX flag.
Indexing is done from Studio as a Celery task. Every publish event triggers the reindex procedure.
You can also reindex the course manually through the Reindex button in the Course Overview page.
Which data gets indexed is determined by the module index_dictionary()
function implementation. Modules supporting this method are Sequence,
Vertical, Video and HTML Block. You can add support to any module
type.
Course metadata, including the name, description, and start and end dates are also indexed.
The following flags are supported in the CMS and LMS applications.
ENABLE_COURSEWARE_INDEX: Enables/disables courseware content and course
info indexing.
ENABLE_LIBRARY_INDEX: Enables/disables library content indexing.
SEARCH_ENGINE: Sets used search engine. There are 2 predefined values,
but more can be added:
"search.elastic.ElasticSearchEngine""search.tests.mock_search_engine.MockSearchEngine"ELASTIC_FIELD_MAPPINGS: Sets any additional field mappings that elastic
search should be aware of. For example, the following code includes the
course start date:
ELASTIC_FIELD_MAPPINGS = {
"start_date": {
"type": "date"
}
}
ENABLE_COURSEWARE_SEARCH: Enables/disables Courseware Search feature (in
course searching)ENABLE_DASHBOARD_SEARCH: Enables/disables Dashboard Search feature (in
enrolled courses searching)ENABLE_COURSE_DISCOVERY: Enables/disables Course Discovery feature (over
courses searching and facet filtering)SEARCH_ENGINE: Sets used search engine. There are 2 predefined values,
but more can be added:"search.elastic.ElasticSearchEngine""search.tests.mock_search_engine.MockSearchEngine"SEARCH_INITIALIZER: Used to set custom
SearchInitializer.SearchInitializer provides an extension to achieve
masquerade and other presearch environmental settings.SearchInitializerlms.lib.courseware_search.lms_search_initializer.LmsSearchInitializerSEARCH_RESULT_PROCESSOR: Used to set custom SearchResultProcessor.
SearchResultProcessor does post processing and data manipulation on a result
set returned by SearchEngine.SearchResultProcessorlms.lib.courseware_search.lms_result_processor.LmsSearchResultProcessorSEARCH_FILTER_GENERATOR: Used to set custom SearchFilterGenerator.
SearchFilterGenerator sets filters defined by current active user. Basic
implementation sets only course start date filter.SearchFilterGenerator"lms.lib.courseware_search.lms_filter_generator.LmsSearchFilterGenerator"