<?php
// $Id: search.test,v 1.11 2008-04-19 16:57:44 boombatower Exp $

define('SEARCH_TYPE', '_test_');

class SearchMatchTestCase extends DrupalWebTestCase {
  /**
   * Implementation of getInfo().
   */
  function getInfo() {
    return array(
      'name' => t('Search engine queries'),
      'description' => t('Indexes content and queries it.'),
      'group' => t('Search'),
    );
  }

  /**
   * Implementation setUp().
   */
  function setUp() {
    parent::setUp('search');
  }

  /**
   * Test search indexing.
   */
  function testMatching() {
    $this->_setup();
    $this->_testQueries();
  }

  /**
   * Set up a small index of items to test against.
   */
  function _setup() {
    variable_set('minimum_word_size', 3);

    for ($i = 1; $i <= 7; ++$i) {
      search_index($i, SEARCH_TYPE, $this->getText($i));
    }
    search_update_totals();
  }

  /**
   * Helper method for generating snippets of content.
   *
   * Generated items to test against:
   *   1  ipsum
   *   2  dolore sit
   *   3  sit am ut
   *   4  am ut enim am
   *   5  ut enim am minim veniam
   *   6  enim am minim veniam es cillum
   *   7  am minim veniam es cillum dolore eu
   */
  function getText($n) {
    $words = explode(' ', "Ipsum dolore sit am. Ut enim am minim veniam. Es cillum dolore eu.");
    return implode(' ', array_slice($words, $n - 1, $n));
  }

  /**
   * Run predefine queries looking for indexed terms.
   */
  function _testQueries() {
    /*
      Note: OR queries that include short words in OR groups are only accepted
      if the ORed terms are ANDed with at least one long word in the rest of the query.

      e.g. enim dolore OR ut = enim (dolore OR ut) = (enim dolor) OR (enim ut) -> good
      e.g. dolore OR ut = (dolore) OR (ut) -> bad

      This is a design limitation to avoid full table scans.
    */
    $queries = array(
      // Simple AND queries.
      'ipsum' => array(1),
      'enim' => array(4, 5, 6),
      'xxxxx' => array(),
      'enim minim' => array(5, 6),
      'enim xxxxx' => array(),
      'dolore eu' => array(7),
      'dolore xx' => array(),
      'ut minim' => array(5),
      'xx minim' => array(),
      'enim veniam am minim ut' => array(5),
      // Simple OR queries.
      'dolore OR ipsum' => array(1, 2, 7),
      'dolore OR xxxxx' => array(2, 7),
      'dolore OR ipsum OR enim' => array(1, 2, 4, 5, 6, 7),
      'ipsum OR dolore sit OR cillum' => array(2, 7),
      'minim dolore OR ipsum' => array(7),
      'dolore OR ipsum veniam' => array(7),
      'minim dolore OR ipsum OR enim' => array(5, 6, 7),
      'dolore xx OR yy' => array(),
      'xxxxx dolore OR ipsum' => array(),
      // Negative queries.
      'dolore -sit' => array(7),
      'dolore -eu' => array(2),
      'dolore -xxxxx' => array(2, 7),
      'dolore -xx' => array(2, 7),
      // Phrase queries.
      '"dolore sit"' => array(2),
      '"sit dolore"' => array(),
      '"am minim veniam es"' => array(6, 7),
      '"minim am veniam es"' => array(),
      // Mixed queries.
      '"am minim veniam es" OR dolore' => array(2, 6, 7),
      '"minim am veniam es" OR "dolore sit"' => array(2),
      '"minim am veniam es" OR "sit dolore"' => array(),
      '"am minim veniam es" -eu' => array(6),
      '"am minim veniam" -"cillum dolore"' => array(5, 6),
      '"am minim veniam" -"dolore cillum"' => array(5, 6, 7),
      'xxxxx "minim am veniam es" OR dolore' => array(),
      'xx "minim am veniam es" OR dolore' => array()
    );
    foreach ($queries as $query => $results) {
      $set = do_search($query, SEARCH_TYPE);
      $this->_testQueryMatching($query, $set, $results);
      $this->_testQueryScores($query, $set, $results);
    }
  }

  /**
   * Test the matching abilities of the engine.
   *
   * Verify if a query produces the correct results.
   */
  function _testQueryMatching($query, $set, $results) {
    // Get result IDs.
    $found = array();
    foreach ($set as $item) {
      $found[] = $item->sid;
    }

    // Compare $results and $found.
    sort($found);
    sort($results);
    $this->assertEqual($found, $results, "Query matching '$query'");
  }

  /**
   * Test the scoring abilities of the engine.
   *
   * Verify if a query produces normalized, monotonous scores.
   */
  function _testQueryScores($query, $set, $results) {
    // Get result scores.
    $scores = array();
    foreach ($set as $item) {
      $scores[] = $item->score;
    }

    // Check order.
    $sorted = $scores;
    sort($sorted);
    $this->assertEqual($scores, array_reverse($sorted), "Query order '$query'");

    // Check range.
    $this->assertEqual(!count($scores) || (min($scores) > 0.0 && max($scores) <= 1.0001), TRUE, "Query scoring '$query'");
  }
}