Document is not found for users searching in another language

Issue

• When searching for a text document (e.g., PDF) by keywords inside the text using the default portal language, the document is shown in the search results.
• When the language is different, the document cannot found.

Environment

• Liferay DXP 7.0, 7.1, 7.2, 7.3

Resolution

• The text of the document is indexed under the Elasticsearch element content_$locale, where $locale is the default language of the portal (e.g., content_en_US). So searching in another language won't return any results. See Searching for Localized Content for more information.
• As a workaround, create a DLFileEntryIndexerPostProcessor module using the BLADE sample available in this DXP 7.0 guide on IndexerPostProcessor.
• Starting sample code for DXP 7.0.

  /**
   * Copyright 2000-present Liferay, Inc.
   */
  
  package com.liferay.document.library.indexerpostprocessor;
  
  import com.liferay.document.library.kernel.model.DLFileEntry;
  import com.liferay.petra.string.StringPool;
  import com.liferay.portal.kernel.search.BooleanQuery;
  import com.liferay.portal.kernel.search.Document;
  import com.liferay.portal.kernel.search.Field;
  import com.liferay.portal.kernel.search.IndexerPostProcessor;
  import com.liferay.portal.kernel.search.SearchContext;
  import com.liferay.portal.kernel.search.Summary;
  import com.liferay.portal.kernel.search.filter.BooleanFilter;
  import com.liferay.portal.kernel.util.LocalizationUtil;
  import com.liferay.portal.kernel.util.PortalUtil;
  import com.liferay.portal.kernel.util.Validator;
  
  import java.util.Locale;
  
  import org.osgi.service.component.annotations.Component;
  
  @Component(
      immediate = true,
      property = {
          "indexer.class.name=com.liferay.document.library.kernel.model.DLFileEntry"
      },
      service = IndexerPostProcessor.class
  )
  public class DLFileEntryIndexerPostProcessor implements IndexerPostProcessor {
  
      @Override
      public void postProcessContextBooleanFilter(
          BooleanFilter booleanFilter, SearchContext searchContext)
      throws Exception {
  
      }
  
      @Override
      public void postProcessContextQuery(
          BooleanQuery contextQuery, SearchContext searchContext)
      throws Exception {
  
      }
  
      @Override
      public void postProcessDocument(Document document, Object obj)
      throws Exception {
  
          DLFileEntry dlFileEntry = (DLFileEntry) obj;
  
          Locale defaultLocale = PortalUtil.getSiteDefaultLocale(dlFileEntry.getGroupId());
  
          String localizedField = LocalizationUtil.getLocalizedName(Field.CONTENT, defaultLocale.toString());
  
          Field localizedContentField = document.getField(localizedField);
  
          document.addText(Field.CONTENT, localizedContentField.getValue());
      }
  
      @Override
      public void postProcessFullQuery(
          BooleanQuery fullQuery, SearchContext searchContext)
      throws Exception {
  
      }
  
      @Override
      public void postProcessSearchQuery(
          BooleanQuery searchQuery, BooleanFilter booleanFilter,
          SearchContext searchContext)
      throws Exception {
  
      }
  
      @Override
      public void postProcessSearchQuery(
          BooleanQuery searchQuery, SearchContext searchContext)
      throws Exception {
  
      }
  
      @Override
      public void postProcessSummary(
          Summary summary, Document document, Locale locale, String snippet) {
  
          if (Validator.isBlank(summary.getContent())) {
              String prefix = Field.SNIPPET + StringPool.UNDERLINE;
  
              String content = document.get(prefix + Field.CONTENT, Field.CONTENT);
  
              summary.setContent(content);
          }
      }
  
  }

  • For DXP 7.1, this sample may be a good point to start. It might need some adjustments and following this DXP 7.1 guide on IndexerPostProcessor. 
  • For DXP 7.2 and DXP 7.3, the methods postProcessContextQuery(BooleanQuery contextQuery, SearchContext searchContext) and postProcessSearchQuery(BooleanQuery searchQuery, SearchContext searchContext) should be skipped.
• Note: It is necessary to reindex the Documents and Media entities after deploying this module for the changes to take effect.

Additional Information

• Searching for Localized Content