core/org.eclipse.smila.solr.test/code/src/org/eclipse/smila/solr/SolrSearchPipelet_HL_Test.java

/*******************************************************************************
 * Copyright (c) 2008 empolis GmbH and brox IT Solutions GmbH. All rights reserved. This program and the accompanying
 * materials are made available under the terms of the Eclipse Public License v1.0 which accompanies this distribution,
 * and is available at http://www.eclipse.org/legal/epl-v10.html
 * 
 * Contributors: Thomas Menzel (brox IT Solution GmbH) - initial creator
 *******************************************************************************/
package org.eclipse.smila.solr;

import org.apache.commons.lang.StringUtils;
import org.eclipse.smila.datamodel.Any;
import org.eclipse.smila.datamodel.AnyMap;
import org.eclipse.smila.datamodel.AnySeq;
import org.eclipse.smila.datamodel.DataFactory;
import org.eclipse.smila.datamodel.xml.XmlSerializationUtils;
import org.eclipse.smila.search.api.SearchResultConstants;
import org.eclipse.smila.search.api.helper.ResultItemAccessor;
import org.eclipse.smila.solr.search.SolrResultAccessor;

/**
 * 
 * @author tmenzel
 */
public class SolrSearchPipelet_HL_Test extends SolrSearchPipeletTestBase {

  /**
   */
  @Override
  protected void setUp() throws Exception {
    super.setUp();
    _solrField = getName() + "_t";
  }

  /**
   * simple case to test HL, that also shows how HL works with solr.
   */
  public void test_Search_HL_Text_Simple() throws Exception {
    // index docs
    addSolrDoc("1", "text containing the word candidate that is to be highlighted once.");
    addSolrDoc("2", "text containing the word candidate that is to be highlighted twice, candidate");
    addSolrDoc("3", "text not containing the word ");
    addSolrDoc("4", "text containing the word candidate that is to be highlighted thrice. "
      + "Candidate 2nd time occurence. And the 3rd Candidate is not retuened due to maxFragsize");
    indexAndCommit();

    // setup search
    final AnyMap highlight = DataFactory.DEFAULT.createAnyMap();
    highlight.put("hl.fragsize", 84);
    _queryBuilder.addHighlightByConfig(_solrField, highlight);
    _queryBuilder.setQuery(_solrField + ":candidate");

    _pipelet.process(_blackboard, new String[] { _record.getId() });
    _log.debug("result record: " + XmlSerializationUtils.serialize2string(_record));

    // verify
    final SolrResultAccessor results = new SolrResultAccessor(WORKFLOW, _record);
    final AnySeq resultRecords = results.getResultRecords();
    assertEquals("the 3rd record should not be found", 3, resultRecords.size());

    assertHitTextContains(results, 1, "<em>candidate</em>", 1);
    assertHitTextContains(results, 2, "<em>candidate</em>", 2);
    assertHitTextContains(results, 4, "<em>candidate</em>", 2);

  }

  // /**
  // * simple case to test HL with regex. this is more an investigation test than a real test of our stuff.
  // */
  // public void test_Search_HL_RegEx() throws Exception {
  // // index docs
  // final String sentence2 = "And a flow up sentence.";
  // addSolrDoc("1", "This is a sentence that contains the search word. " + sentence2);
  // indexAndCommit();
  //
  // // setup search
  // final AnyMap highlight = DataFactory.DEFAULT.createAnyMap();
  // highlight.put("hl.fragmenter", "regex");
  // _queryBuilder.addHighlightByConfig(_solrField, highlight);
  // _queryBuilder.setQuery(_solrField + ":contains");
  //
  // _pipelet.process(_blackboard, new String[] { _record.getId() });
  // _log.debug("result record: " + XmlSerializationUtils.serialize2string(_record));
  //
  // // verify
  // final SolrResultAccessor results = new SolrResultAccessor(WORKFLOW, _record);
  // final AnySeq resultRecords = results.getResultRecords();
  // assertEquals("the 3rd record should not be found", 1, resultRecords.size());
  //
  // assertHitTextContains(results, 1, "<em>contains</em>", 1);
  // // "sentence2 should be dropped by regex fragger
  // assertHitTextContains(results, 1, sentence2, 0);
  // }

  // /**
  // * investigation/regression test to mimic the old sentence HL transformer with OOB solr classes. this just tests if
  // * there are 2 nippets returned where it is assumed that the 1st returns a merged one.
  // */
  // public void test_Search_HL_Text_SentenceStyle() throws Exception {
  // // index docs
  // addSolrDoc("1", "Text containing 1 the word. Text containing 2 the word. " //
  // + repeat("Some other content w/o the word. ", 10) //
  // + "Text containing 3 the word.");
  // indexAndCommit();
  //
  // // setup search
  // final AnyMap highlight = DataFactory.DEFAULT.createAnyMap();
  // highlight.put("hl.fragsize", 20);
  // highlight.put("hl.sippets", 5);
  // highlight.put("hl.fragmenter", "regex");
  // highlight.put("hl.mergeContiguous", true);
  // _queryBuilder.addHighlightByConfig(_solrField, highlight);
  // _queryBuilder.setQuery(_solrField + ":containing");
  //
  // _pipelet.process(_blackboard, new String[] { _record.getId() });
  // _log.debug("result record: " + XmlSerializationUtils.serialize2string(_record));
  //
  // // verify
  // final SolrResultAccessor results = new SolrResultAccessor(WORKFLOW, _record);
  // final AnySeq resultRecords = results.getResultRecords();
  // assertEquals("", 1, resultRecords.size());
  //
  // final ResultItemAccessor result = results.getResultRecord(0);
  // final Any highlightingResult =
  // result.getHighlight().getMap(_solrField).get(SearchResultConstants.HIGHLIGHT_TEXT);
  // if (!highlightingResult.isSeq()) {
  // fail("Should be sequence");
  // } else {
  // // final AnySeq solrFieldValue = result.getMetadata().getSeq(_solrField);
  // // assertEquals(2, solrFieldValue.size());
  // assertEquals(2, highlightingResult.asSeq().size());
  // }
  // }

  private void assertHitTextContains(final SolrResultAccessor results, int id, String expectedHlWordWithTags,
    int expectedCount) {
    final String id2 = getId(id + "");
    for (Any itemMap : results.getResultRecords()) {
      final ResultItemAccessor result = new ResultItemAccessor(-1, itemMap.asMap());
      if (result.getRecordId().equals(id2)) {
        final String text =
          result.getHighlight().getMap(_solrField).getStringValue(SearchResultConstants.HIGHLIGHT_TEXT);
        assertEquals(expectedCount,
          StringUtils.countMatches(text.toLowerCase(), expectedHlWordWithTags.toLowerCase()));
        return;
      }
    } // for
    fail("record for id not found: " + id2);
  }

}