collective.elasticindex 1.2.5

Index Plone content in Elastic Search

This extension index Plone content into ElasticSearch. This doesn’t replace the Plone catalog with ElasticSearch, nor interact with the Plone catalog at all, it merely index content inside ElasticSearch when it is modified or published.

In addition to this, it provides a simple search page called search.html that queries ElasticSearch using Javascript (so Plone is not involved in searching) and propose the same features than the default Plone search page. A search portlet let you redirect people to this new search page as well.

This extension have been built for Plone 4, but might work with Plone 3.

Usage

After adding this extension to your buildout (including the zcml), you can install the extension in Plone. A configuration screen is available inside site setup. It will let you configure the URLs of the ElasticSearch servers to use in order to index, and search. To proceed:

• Fill in the ElasticSearch settings,

• Click on Save,

• Click on Create Index in order to create the ElasticSearch index,

• Click on Import site content in order to index already existing content in ElasticSearch.

You can use the same ElasticSearch server (and probably index) for multiple Plone sites, creating a federated search that way.

Security disclaimer

By default is no authentication or access validation while searching or indexing content. The original purpose of this search is to be public.

If you have private content that you don’t want to be searchable or viewable by unauthorized people, please be sure to check the checkbox index only published content in the configuration screen.

In addition to this ElasticSearch is not secured by default, meaning there is no authentication to provide in order to index or look-up content. Be sure to hide it behind a firewall and use a proxy or Apache in order to restrict the requests made to it: you only need to allow access via POST to the sub-URL _search after the index name configured in the configuration screen. For instance, if the index name is plone, you shall allow only requests to http://your-public-es-url/plone/_search. After you configured your proxy, be sure to configure its public URL, like http://your-public-es-url in the configuration screen so the search page knows how to contact it.

However if you want to allow users to search though restricted and not yet published content, you can check index security and uncheck index only published content in the configuration screen. After reindexing your content, if you check proxy search requests though Plone and apply security filter, search will work on restricted and not yet published content, but will be slower as the queries will be proxied though Plone.

Changes

1.2.5 (2015-04-21)

• Less items per commit while indexing

1.2.4 (2015-04-21)

• More logging while indexing.

1.2.3 (2015-02-03)

• Fix indexation after checkin from plone.app.iterate.

1.2.2 (2015-01-29)

• Fix savepoint.

1.2.1 (25/03/2014)

• Update javascript so it longer strips double quotes away.

1.2 (28/01/2014)

• Add an option to rewrite the domain name in the URL of indexed documents.

1.1.2 (21/01/2014)

• Fix a stupid bug when a document that is unindex get reindex inside the same transaction.

1.1.1 (26/11/2013)

• Disable completely the extension when portal_factory is in the path, i.e. we have an add forms. portal_factory uses lot of hacks regarding the ZODB that are not very clean.

1.1 (05/11/2013)

• Index Plone authorization in the index and provide an option to redirect the queries through Plone in order to restrict the search results to what the current user is allowed to see. Effectively Plone will modify the search query to include a restriction on the authorization.

• Improve the public search page.

• Improve the public search javascript to ignore special Lucene characters in the query and colons except if they are before an authorized term.

1.0 (13/09/2013)

• Initial release.