Repository files navigation

This project offers a distance-based spatial clustering search component for Apache Solr.It addresses the problem of reducing the amount of displayed markers on a map, described asSpatial Clustering,using adistance-based clustering algorithm based onGVM.

The search component aggregates all possible search results to a maximum amount of pins and adds this information to the standard search result representation. Like faceting, it can be used to query for a paged result slice (eg. for a result list) and a geographic overview of ALL search result items (spatial clusters) at once.

Download the plugin jar fromhttp://repo1.maven.org/maven2/com/indoqa/solr/spatial-clustering/7.5.0/spatial-clustering-7.5.0-jar-with-dependencies.jar and copy it into the/lib directory of your solr core.

To enable spatial clustering, store the geo information (longitude and latitude) in your solr document:

<fieldTypename="pdouble"class="solr.DoublePointField" /><fieldname="latitude"type="pdouble"indexed="true"stored="true" /><fieldname="longitude"type="pdouble"indexed="true"stored="true" />

Note: For legacy support of old Solr 4SortableDoubleField, see branchlegacy/solr-4.3

Define the search component and map field names for id, longitude and latitude, as well as the maximum allowed number of clusters:

<searchComponentclass="com.indoqa.solr.spatial.clustering.SpatialClusteringComponent"name="spatial-clustering">  <strname="fieldId">id</str>  <strname="fieldLon">longitude</str>  <strname="fieldLat">latitude</str>  <intname="maxSize">1000000</int></searchComponent>

After that, add the spatial component to your query component chain:

<requestHandlername="/search"class="solr.SearchHandler"default="true">  <arrname="last-components">    <str>spatial-clustering</str>  </arr></requestHandler>

• spatial-clustering=true -> Enables spatial clustering
• spatial-clustering.size=20 -> Optionally sets the maximum number of clusters (=pins)
• spatial-clustering.min-result-count=100 -> Optionally sets the minimum number of documents required to do clustering

Similar to facets, the computed clusters are added to the search result after the requested documents. There are two types ofresult pins:

• single: Represents a single document, including the id of the referenced document.
• cluster: Represents an aggregated pin covering more than one document, including the cluster size.

<lstname="spatial-clustering">  <lstname="pin">    <strname="type">single</str>    <intname="size">1</int>    <doublename="longitude">16.345518</double>    <doublename="latitude">48.285202</double>    <stringname="reference">document-2313</string>  </lst>  <lstname="pin">    <strname="type">cluster</str>    <intname="size">3</int>    <doublename="longitude">16.2461115932</double>    <doublename="latitude">48.20259082573333</double>  </lst>  ...  ...</lst>

• Apache Solr 7.5.0+
• Java 8+

• Download the latest release
• runmaven clean install