...
This guide will help you get up and running with a Fedora 4 instance whose updates are automatically indexed in a Solr repository. This guide glosses over many of the details and should be considered a starting point for testing this feature. The document assumes a POSIX operating system with cURL, a text editor, Java, Git, and a download of Apache Solr 4.6.0.
The pattern for integrating Solr with Fedora 4 is to take advantage of the messages that Fedora 4 emits after every change to resources within the repository. The assumption is that you have three components running independently, in a completely decoupled fashion:
- Fedora 4
- Solr
- Indexing service
The Indexing service:
- Consumes the messages emitted by Fedora 4
- Makes a request back to Fedora 4 get a transformed representation of the Fedora 4 resource, then
- Submits the transformed resource to Solr for indexing
The most current documentation for the Indexing service can be found on GitHub.
Install and Start Fedora 4
...
Code Block | ||||||||
---|---|---|---|---|---|---|---|---|
| ||||||||
<!-- To enable dynamic schema REST APIs, use the following for <schemaFactory>: --> <schemaFactory class="ManagedIndexSchemaFactory"> <bool name="mutable">true</bool> <str name="managedSchemaResourceName">managed-schema</str> </schemaFactory> <!-- When ManagedIndexSchemaFactory is specified, Solr will load the schema from the resource named in 'managedSchemaResourceName', rather than from schema.xml. Note that the managed schema resource CANNOT be named schema.xml. If the managed schema does not exist, Solr will create it after reading schema.xml, then rename 'schema.xml' to 'schema.xml.bak'. Do NOT hand edit the managed schema - external modifications will be ignored and overwritten as a result of schema modification REST API calls. When ManagedIndexSchemaFactory is specified with mutable = true, schema modification REST API calls will be allowed; otherwise, error responses will be sent back for these requests. --> <!-- <schemaFactory class="ClassicIndexSchemaFactory"/> --> | ||||||||
The fcrepo-message-consumer SolrIndexer implementation does not commit upon updates. In order to see the changes, you must configure Solr to have a commit strategy that is appropriate for your use. |
...
...
<!-- AutoCommit
Perform a hard commit automatically under certain conditions.
Instead of enabling autoCommit, consider using "commitWithin"
when adding documents.
http://wiki.apache.org/solr/UpdateXmlMessages
maxDocs - Maximum number of documents to add since the last
commit before automatically triggering a new commit.
maxTime - Maximum amount of time in ms that is allowed to pass
since a document was added before automatically
triggering a new commit.
openSearcher - if false, the commit causes recent index changes
to be flushed to stable storage, but does not cause a new
searcher to be opened to make those changes visible.
If the updateLog is enabled, then it's highly recommended to
have some sort of hard autoCommit to limit the log size.
-->
<autoCommit>
<maxTime>${solr.autoCommit.maxTime:15000}</maxTime>
<openSearcher>false</openSearcher>
</autoCommit>
Start Solr and verify that it is running at http://localhost:8983/solr.
Code Block | ||
---|---|---|
| ||
cd solr-4.6.0/example java -jar start.jar |
Add the "uuid" field ("title" and "id" already exist).
Code Block | ||
---|---|---|
| ||
curl -X POST -H "Content-Type: application/json" --data-binary "@solr-fields.json" "http://localhost:8983/solr/schema/fields" |
Code Block | ||
---|---|---|
| ||
[{"name":"uuid","type":"text_general","stored":"true","indexed":"true"}] |
Download, Build, Configure and Start fcrepo-message-consumer
Code Block | ||
---|---|---|
| ||
git clone git@github.com:fcrepo4/fcrepo-message-consumer.git |
Edit the configuration at fcrepo-message-consumer/fcrepo-message-consumer-webapp/src/main/resources/spring/indexer-core.xml to point to your Solr installation.
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
<!-- Solr Indexer START-->
<bean id="solrIndexer" class="org.fcrepo.indexer.solr.SolrIndexer">
<constructor-arg ref="solrServer" />
</bean>
<!--External Solr Server -->
<bean id="solrServer" class="org.apache.solr.client.solrj.impl.HttpSolrServer">
<constructor-arg index="0" value="http://${fcrepo.host:localhost}:${solrIndexer.port:8983}/solr/" />
</bean>
<!-- Solr Indexer END-->
<!-- Message Driven POJO (MDP) that manages individual indexers -->
<bean id="indexerGroup" class="org.fcrepo.indexer.IndexerGroup">
<constructor-arg name="indexers">
<set>
<!--
<ref bean="jcrXmlPersist"/>
<ref bean="fileSerializer"/>
<ref bean="sparqlUpdate"/> -->
<!--To enable solr Indexer, please uncomment line below -->
<ref bean="solrIndexer"/>
</set>
</constructor-arg>
<!-- If your Fedora instance requires authentication, enter the
credentials here. Leave blank if your repo is open. -->
<constructor-arg name="fedoraUsername" value="${fcrepo.username:}" /> <!-- i.e., manager, tomcat, etc. -->
<constructor-arg name="fedoraPassword" value="${fcrepo.password:}" />
</bean> |
Start the application (in this case on port 9999).
Code Block | ||
---|---|---|
| ||
mvn clean install -DskipTests
cd fcrepo-message-consumer-webapp
mvn -Djetty.port=9999 jetty:run |
Create an Indexable resource
Code Block | ||
---|---|---|
| ||
curl -v -X PUT -H "Content-Type: text/turtle" --data-binary "@object.rdf" "http://localhost:8080/rest/indexableObject" |
Code Block | ||||
---|---|---|---|---|
| ||||
PREFIX dc: <http://purl.org/dc/elements/1.1/>
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX indexing: <http://fedora.info/definitions/v4/indexing#>
<> indexing:hasIndexingTransformation "default"; rdf:type indexing:Indexable; dc:title "This title will show up in the index."
|
Install and Start the Indexing Service
See documentationEnsure that the records are committed to Solr (either through an explicit commit or waiting until the configured commit period is up) and then verify that they show up.