Modifying free-text search for additional fields

Use the following file to define a new free-text field in the Guidewire Solr Extension.

/opt/gwsolr/pc/solr/policy_active/conf/schema.xml

The Guidewire Solr Extension defines the format of this file.

The schema configuration file contains <field> elements, one for each full-text field of the index documents in the Guidewire Solr Extension.

The following example defines a full-text field that stores postal codes.

<field name="postalCode" type="gw_unanalyzed" indexed="true" 
        stored="true" required="false"
        multiValued="false"/>

The definition directs the Guidewire Solr Extension to index the values of postal code fields, so search criteria can include postal codes. The definition directs the Guidewire Solr Extension to store the values of postal code fields, so items returned in search results include them.

The type attribute in the preceding definition identifies the analyzer with which Guidewire Solr Extension processes postal code fields. The analyzer type indicates the category of matches that are possible on applicable search fields. In this case, a value for the type attribute of gw_unanalyzed indicates that postal codes are raw text fields and that Guidewire Solr Extension does not analyze them. Given this value, only exact matches are possible for searches on the values of postal code fields.

Use the following file to define a new free-text field in PolicyCenter.

PolicyCenter/modules/configuration/config/search/policy-search-config.xml

Access the file in the Project window in Studio by navigating to configuration > config > search policy-search-config.xml. PolicyCenter defines the format of this file.

Free-text search configuration files have these main elements:

• <Indexer> – Contains <IndexField> elements to define field names and their locations within object graphs of the root object and in the index documents sent to the Guidewire Solr Extension.
• <Query> – Contains the following types of elements:
  • <QueryTerm> elements define how specific fields are matched and how a match contributes to the overall score. A query term is one of two types: term and subquery. A term type searches a single index for a single field. A subquery type searches multiple indices for multiple fields simultaneously and scores the most appropriate match.
    Important: Do not confuse the term type query term with the subquery type query term. A subquery type can search multiple indexes for multiple fields. A term type can only search one index for a single field. An XSD only validates the structure of a query. The XSD does not validate the inputs for a term type. Thus, if you attempt to enter multiple terms in a term type, the XSD will not catch this error. Instead, the platform will throw a run-time exception.
  • <FilterTerm> elements restrict Guidewire Solr Extension from returning a result if a field in that result matches given search criteria. The elements do not affect the score of any results. Instead, they help users to limit the results that a search returns, such as by a date range. The elements can also facilitate security by controlling the results that users receive.
• <QueryResult> – Contains <ResultProperty> elements to configure whether and how specific fields are returned in query results from the Guidewire Solr Extension.

To add a free-text field to policy-search-config.xml, first add an <IndexField> element to the <Indexer> element. The following example defines a free-text field for postal codes.

<IndexField field="postalCode">
  <DataProperty path="root.PolicyAddress.PostalCode"/>
</IndexField>

The definition specifies that values of postalCode fields in the index documents sent to the full-text engine come from addresses on policy periods, the root object.

Next, add <FilterTerm> elements for the new free-text field to the <Query> element. The following example specifies how the Guidewire Solr Extension matches PostalCodeCriteria values in search criteria with postalCode values in the Guidewire Solr Extension.

<FilterTerm>
  <DataProperty path="root.PostalCodeCriteria"/>
  <QueryField field="postalCode"/>
</FilterTerm>

The definition directs the Guidewire Solr Extension to accept postal codes in search criteria and where to find them in the XML structure of the search criteria.

Finally, add a <ResultProperty> element to the <QueryResult> element. The following example defines how the Guidewire Solr Extension returns postalCode values in query results.

<ResultProperty name="PostalCode">
  <ResultField name="postalCode"/>
</ResultProperty>

The definition assigns the postalCode value in the result to the PostalCode on the result object.

To load data from an existing PolicyCenter database, the new free-text field must be listed in the batchload-config-database brand.xml files and the data-config.xml file. You decide whether to include or exclude the field from the digest that prevents duplicate index entries.

SELECT DISTINCT
  ...
  addr.postalCode
  ...
  FROM pc_policyperiod AS pp
    ...
    INNER JOIN pc_policyaddress paddr
      ON paddr.BranchID = pp.ID
    ...
    INNER JOIN pc_address addr
      ON addr.ID = paddr.Address
    ...
  WHERE 
    ...
    AND (paddr.EffectiveDate  IS NULL OR paddr.EffectiveDate  <= pp2.EditEffectiveDate )
    AND (paddr.ExpirationDate IS NULL OR paddr.ExpirationDate >  pp2.EditEffectiveDate )
    ...

<CONTAINER_ELEM>
  <POLICY>
    <!-- data for one policy -->
  </POLICY>
</CONTAINER_ELEM>

<field column="postalCode" xpath="/CONTAINER_ELEM/POLICY/POSTALCODE"/>

sb.append("postalCode", period.PolicyAddress.PostalCode)

<transformer
  name="digestTransformer"
  class="com.guidewire.solr.batchload.xform.PCDigestTransformer"
  ignoreElems="urn, periodID, policyPublicID, sliceDate, periodStart, periodEnd, policyStart,
          policyEnd, periodIdWithSliceDate, jobType"
  algorithm="SHA"
/>

static function initXformer() : com.guidewire.solr.batchload.xform.DigestTransformer {
  try {
    var xml =
      "<transformer name=\"digestTransformer\"
         class=\"com.guidewire.solr.batchload.xform.PCDigestTransformer\""
                + " algorithm=\"SHA\""
                + " ignoreElems=\"urn, periodID, policyPublicID, sliceDate, periodStart, periodEnd, ""
                + " policyStart, policyEnd, periodIdWithSliceDate, jobType\"/>";
    var xf = new com.guidewire.solr.batchload.xform.PCDigestTransformer(false)
    var doc = com.guidewire.solr.batchload.Utils.parseXml(xml)
    xf.configure(doc.getDocumentElement())
  }
}