/**
* Returns the best snippet that contains the query from the $text passed in
* as a parameter. For now this is a pretty dumb function, it could definitely
* be improved in the future.
*
* NOTE: This method is VERY basic and it will not get amazing results, but
* it's a start.
* NOTE: This seems to not work correctly if $query contains diacritics.
*
* ATTENTION: This doesn't current work properly if tags are in here, it just
* cuts them in half.
*
* @param {String} $text The original text from which we want to extract
* the snippet.
* @param {String} $query The query text.
* @return {String} The cut down string.
*/
function getSnippet($text, $query)
{
if ($query == '') {
$suffix = 250 < strlen($text) ? '...' : '';
return substr($text, 0, min(strlen($text), 250)) . $suffix;
}
$text = correctDiacritics($text);
$query = correctDiacritics($query);
$first_pos = stripos($text, $query);
$start = max(0, $first_pos - 70);
$end = min(strlen($text), $first_pos + 200);
$prefix = $start > 0 ? '...' : '';
$suffix = $end < strlen($text) ? '...' : '';
return $prefix . substr($text, $start, $end - $start) . $suffix;
}
function getCollegeSearch($query)
{
$ignore_words = array("str", "strada", "ale", "aleea", "din", "bld", "bulevardul", "nr", "numarul", "piata", "pta", "orasul", "comuna", "satul", "sat");
$query = getStringWithoutDiacritics($query);
$query = mysql_real_escape_string($query);
$words = explode(" ", $query);
$likes = array();
foreach ($words as $word) {
// Ignore one and two letter words, stopwords, and numbers;
if (strlen($word) > 2 && !in_array(strtolower($word), $ignore_words) && (int) $word == 0) {
$likes[] = "description LIKE '%{$word}%'";
$likes[] = "description LIKE '{$word}%'";
$likes[] = "description LIKE '%{$word}'";
}
}
if (count($likes) == 0) {
return array();
}
$where = implode(" OR ", $likes);
$s = mysql_query("\n SELECT *\n FROM electoral_colleges\n WHERE {$where}");
$result = array();
while ($r = mysql_fetch_array($s)) {
// We drop the descriptions where the match is not a full word, but instead
// only a subset of a word. So for example... if the search was "ion" and
// the description was "ionescu", we drop it.
//
// We have to do it this way because MYSQL native regexp matching does not
// handle diacritics well, whereas LIKE works perfectly.
$clean_description = strtolower(getStringWithoutDiacritics($r['description']));
if (countMatches($words, $clean_description, $ignore_words) == 0) {
continue;
}
$key = $r['name_cdep'];
$description = highlightWords(correctDiacritics($r['description']), $words);
if (!array_key_exists($key, $result)) {
$result[$key] = array();
$result[$key]['score'] = 0;
$result[$key]['description'] = array();
$result[$key]['matched_words'] = array();
$result[$key]['name_cdep'] = $r['name_cdep'];
$result[$key]['name_senat'] = $r['name_senat'];
$result[$key]['id'] = $r['id'];
}
$result[$key]['description'][] = $description;
$result[$key]['matched_words'] = setMatchedWords($result[$key]['matched_words'], $r['description'], $words);
if (startsWith($description, "Municipiul")) {
$result[$key]['score'] += 2;
}
if (startsWith($description, "Localitate componentă") || startsWith($description, "Orașul")) {
$result[$key]['score'] += 1;
}
$result[$key]['score'] += (countMatches($words, $clean_description, $ignore_words) - 1) * 2;
}
foreach ($result as $key => $value) {
foreach ($words as $word) {
$pos = strpos(strtolower($key), strtolower($word));
if ($pos === false) {
// string needle NOT found in haystack
} else {
// string needle found in haystack
$result[$key]['score'] += 2;
}
}
$result[$key]['score'] += count($result[$key]['matched_words']);
}
usort($result, "collegeResultCompare");
return array_slice($result, 0, 25);
}
