The Solr plugin provides and interface between Newscoop and the Apache Solr full-text search server. A requirement is have a running Solr server which is configured correctly. By default the plugin provides two configuration possibilities. The first is an indexing only configuration and the second an indexing and storage configuration.
The advantage of using Solr, compared to the default Newscoop search, is that Solr is specifically designed for very fast full-text searches. Solr provides and extensive technical interface for searching and filtering results based on the request parameters. E.g. its possible to search in multiple data fields and give more weight to one or more fields, so if a keyword is matches in an article title, it has more weight then when it's mached in the article body. For both configurations the filter possibilities remain exactly the same.
The original Solr interface implementation for Newscoop was an indexing only one. That means that data is only being stored in Solr for indexing purposes and when performing queries only article numbers are being returned.
For TagesWoche the indexing and storage implementation was developed. This means that not only data is being stored for indexing purposes, but also for presentation purposes. E.g. the omniticker, search and topic pages all heavily rely on this. Per object stored in Solr, we also store the most important data attributes (think of name, id, dates, etc.) When performing a search query not only identifiers are returned, but also the object data. This data is then used in the presentation layer. By using this mechanism we skip additional requests to the internal Newscoop storage which has a positive effect on performance.
For a more technical and detailed description of setting up Solr and implementing it in templates, i would refer to the wiki page.
There has been error in communication with Booktype server. Not sure right now where is the problem.
You should refresh this page.