SMILA/Documentation/Solr Integration

org.eclipse.smila.integration.solr.SolrIndexPipelet

Description

This pipelet integrates Apache Solr's HTTP REST API (http://wiki.apache.org/solr/SearchHandler?action=show&redirect=StandardRequestHandler and http://wiki.apache.org/solr/UpdateXmlMessages ) for adding records to resp. deleting records from a Solr index. Therefore, it is now possible to build large volume search applications by basing them on SMILA and deploying Solr for coping with large indices which are not manageable on one node because of their large size which exceeds the limits of the node's file system.

Note: Currently this pipelet supports Solr 1.4. The integration of features added with the release of Solr 1.5 will be accomplished in the near future.

Configuration

Property	Constraint	Description
executionMode	Optional	The execution mode of the pipelet, default is ADD.
allowDoublets	Optional	The instruction for Solr if doublets are allowed, default is false.
commitWithin	Optional	The commit interval for Solr in miliseconds, default is 10000.
indexname	Optional	The name of the index/core on which Solr should operate. Currently we do not support more than one index/core.

Example

Pipelet configuration for SolrIndexPipelet in addpipeline.bpel

<extensionActivity>
  <proc:invokePipelet name="invokeSolrPipelet">
    <proc:pipelet class="org.eclipse.smila.integration.solr.SolrIndexPipelet" />
    <proc:variables input="request" />
    <proc:configuration>
      <rec:Val key="executionMode">ADD</rec:Val>
      <rec:Val key="allowDoublets" type="boolean">false</rec:Val>
      <rec:Val key="commitWithin" type="long">10000</rec:Val>
      <rec:Val key="indexname">test_index</rec:Val>
    </proc:configuration>
  </proc:invokePipelet>
</extensionActivity>

org.eclipse.smila.integration.solr.SolrSearchPipelet

Description

This pipelet integrates Apache Solr's HTTP REST API (http://wiki.apache.org/solr/SearchHandler?action=show&redirect=StandardRequestHandler and http://wiki.apache.org/solr/UpdateXmlMessages ) for searching in a Solr index.

Note: Currently this pipelet supports Solr 1.4. The integration of features added with the release of Solr 1.5 will be accomplished in the near future.

Configuration

From the search request, the pipelet uses the standard search parameters query, maxcount, offset, and resultAttributes. In addition to that, the following properties can be set in the pipelet configuration:

Property	Constraint	Description
highlight	Optional	The highlight flag, default is true.
highlight.params	Optional	All highlighting parameters passed directly to Solr. For more details, please see http://wiki.apache.org/solr/HighlightingParameters.

Example

Pipelet configuration for SolrSearchPipelet in searchpipeline.bpel

<extensionActivity>
  <proc:invokePipelet name="invokeSolrPipelet">
    <proc:pipelet class="org.eclipse.smila.integration.solr.SolrSearchPipelet" />
    <proc:variables input="request" />
    <proc:configuration>
      <rec:Seq key="shards">
        <rec:Val>localhost:8983/solr</rec:Val>
      </rec:Seq>
      <rec:Val key="highlight" type="boolean">true</rec:Val>
      <rec:Seq key="highlight.params">
        <rec:Val>hl.fl=Content</rec:Val>
        <rec:Val>hl.snippets=3</rec:Val>
        <rec:Val>hl.simple.pre=&lt;b&gt;</rec:Val>
        <rec:Val>hl.simple.post=&lt;/b&gt;</rec:Val>
        <rec:Val>hl.fragsize=300</rec:Val>
        <rec:Val>hl.mergeContiguous=false</rec:Val>
      </rec:Seq>
    </proc:configuration>
  </proc:invokePipelet>
</extensionActivity>