The core templates are specified in the URL used for creating shards, as shown below:
|rerank||This template is an enhanced core configuration for Alfresco Content Services. To use rerank, you need to reindex using this template, when creating a new core. It has more appropriate settings for sharding and supports indexes containing approximately 50-80M documents per shard.|
|noRerank||This template matches how the alfresco and archive cores were defined in Alfresco Content Services 5.0. In addition, it supports auto-phrasing and query re-ranking.|
The core templates include schema.xml and solrconfig.xml. The main purpose is to create multiple cores on multiple machines with the same configuration.
Comparison between the rerank and noRerank templates
|No.||Rerank template||noRerank template|
|1||The rerank template causes less duplication of the index, and therefore the index is more compact.||The noRerank template causes more duplication of the index, and therefore the index is large.|
|2||In the rerank template, stop words are included and indexed as
common grams. By default, majority of the 100 most frequently used words in English
language text are now treated as stop words.
For more information, see <SOLR_HOME>/templates/rerank/conf/lang/stopwords_en.txt.
|In the noRerank template, stop words are removed from the
words that are tokenised in the English language.
For more information, see <SOLR_HOME>/templates/norerank/conf/lang/stopwords_en.txt.
|3||The rerank template supports real rerank with automatic phrasing*
(or auto-phrasing). Queries are run in two stages:
* When a user provides individual search terms in a query, the automatic phrasing feature groups those individual terms into a search phrase and returns the query results for the phrase.
|The noRerank core performs auto-phrasing without re-ranking but the auto-phrase is added to the query.|