function nameSimilarity($matched, $name, $type = null)
{
if ($matched == 'N/A') {
return 0;
}
//$matched_cleaned = canonical_form($matched, true);
$matched_cleaned = $matched;
if (empty($matched_cleaned)) {
return 0;
}
$score = 3;
$parts1 = explode(" ", $matched_cleaned);
$parts2 = explode(" ", $name);
if (count($parts1) === count($parts2)) {
if (levenshtein($matched_cleaned, $name) <= 3 || preg_match('/full/i', $type)) {
$score += 0.5;
}
for ($pidx = 0; $pidx < count($parts1); $pidx++) {
if ($parts1[$pidx][0] != $parts2[$pidx][0]) {
$score -= 1.5;
}
}
}
if (preg_match('/sound|look/i', $type)) {
// $score -= 0.05;
$score -= levenshtein($name, $matched_cleaned) / 20;
}
$penalty = 0;
if (count($parts1) == count($parts2)) {
$penalty = 0;
} elseif (count(array_unique($parts1)) != count(array_unique($parts2))) {
$penalty = 0.01;
}
if (count(array_unique($parts1)) > count(array_unique($parts2))) {
$penalty += 0.5;
// $penalty = $penalty / (3 - count(array_unique(array_slice($parts1, 1))));
} elseif (count(array_unique($parts1)) < count(array_unique($parts2))) {
if (count(array_unique($parts1)) < count($parts1)) {
$penalty -= 0.0;
} else {
$penalty -= 0.2;
}
// $penalty = ($penalty < 0)?0:$penalty;
}
$score -= 2 * levenshtein($parts1[0], $parts2[0]) / strlen($parts1[0]);
$sub_parts1 = array_slice($parts1, 1);
$sub_parts2 = array_slice($parts2, 1);
$total_err = 0;
foreach ($sub_parts2 as $sp2_idx => $sp2) {
// $min_err = 999.0;
// $min_errs[$sp1] = 999.0;
if (is_null(@$min_errs[$sp2])) {
$min_errs[$sp2] = 999.0;
}
foreach ($sub_parts1 as $sp1_idx => $sp1) {
if (levenshtein($sp1, $sp2) <= 3 && treat_word($sp1[0]) == treat_word($sp2[0])) {
// echo "<xmp>$sp1 $sp2 ". levenshtein($sp1, $sp2) . "</xmp>";
$tmp_err = (double) levenshtein($sp1, $sp2) / (double) strlen($sp1);
// $min_err = min($min_err, $tmp_err);
$min_errs[$sp2] = min($min_errs[$sp2], $tmp_err);
// echo "$sp1, $sp2, $min_err, $total_err, $matched<br/>";
} else {
// $min_err = 0.5;
if (count($sub_parts1) != count($sub_parts2)) {
$factor = count($sub_parts1) + count($sub_parts2) - ($sp1_idx + $sp2_idx + 1);
} else {
$factor = 1;
}
/*
echo "<xmp>";
var_dump(array($sp2, $sp1, $factor, min($min_errs[$sp2], 1 / $factor)));
echo "</xmp>";
//*/
$min_errs[$sp2] = min($min_errs[$sp2], 1 / $factor);
/*
echo "<xmp>";
var_dump(array($sub_parts1, $sub_parts2));
echo "</xmp>";
//*/
}
}
}
/*
echo "<xmp>";
var_dump($min_errs);
echo "</xmp>";
//*/
foreach (array_unique($sub_parts2) as $sp2) {
$total_err += $min_errs[$sp2];
}
// $score -= (($total_err>1.5)?1.5:$total_err);
$score -= $total_err;
/*
echo "<xmp>";
var_dump(array($matched, $penalty, $score, $score - $penalty));
echo "</xmp>";
//*/
return $score - $penalty;
}
function queryNames($name, $against, $best, $ep)
{
if (empty($ep)) {
return false;
}
$ep .= '/select?wt=json&q=*:*';
// $ep = 'http://localhost:8983/solr/taxa/select?wt=json&q=*:*';
// $ep = 'http://140.109.28.72/solr4/taxa/select?wt=json&q=*:*';
extract_results("", "", $reset = true);
// mix2; work with latin part b2, c2, and suggestions of latin part b2, c2
$mix2 = array();
$sound_mix2 = array();
$matched = array();
$info = array();
$suggestions = array();
$long_suggestions = array();
$name_cleaned = canonical_form($name, true);
$parts = explode(" ", $name_cleaned);
$lpa2 = $parts[0];
$lpb2 = @$parts[1];
$lpc2 = @$parts[2];
$spa2 = treat_word($lpa2);
$spb2 = treat_word($lpb2);
$spc2 = treat_word($lpc2);
if (!empty($parts[1])) {
$mix2[] = $parts[1];
} else {
// return null;
return array('N/A' => array('name' => $name, 'name_cleaned' => $name_cleaned, 'matched' => 'N/A', 'matched_clean' => 'N/A', 'accepted_namecode' => array(), 'namecode' => array(), 'source' => array(), 'url_id' => array(), 'a_url_id' => array(), 'kingdom' => array(), 'phylum' => array(), 'class' => array(), 'order' => array(), 'family' => array(), 'higher_than_family' => array(), 'type' => 'N/A'));
}
if (!empty($parts[2])) {
$mix2[] = $parts[2];
}
if (!empty($spb2)) {
$sound_mix2[] = $spb2;
}
if (!empty($spc2)) {
$sound_mix2[] = $spc2;
}
// Type 1
$query_url_1 = $ep . '&fq=canonical_name:"' . urlencode($name_cleaned) . '"';
extract_results($query_url_1, TYPE_1, $reset = false, $against);
// with minor spell error
$query_url_1_err_suggestion = $ep . "&rows=0&spellcheck.q=" . urlencode(implode(" ", $mix2));
$suggestion = extract_suggestion($query_url_1_err_suggestion, TYPE_1_E);
if (!empty($suggestion)) {
$query_url_1_err = $ep . '&fq=canonical_name:"' . urlencode("{$lpa2} {$suggestion}") . '"';
extract_results($query_url_1_err, TYPE_1_E, $reset = false, $against);
}
//*
$query_url_1_err_long_suggestion = $ep . "&rows=0&spellcheck.q=" . urlencode($name_cleaned);
$long_suggestion = extract_suggestion($query_url_1_err_long_suggestion, TYPE_1_E);
if (!empty($long_suggestion)) {
$query_url_1_err = $ep . '&fq=latin_part_a:' . $lpa2 . '&fq=canonical_name:"' . urlencode("{$long_suggestion}") . '"';
extract_results($query_url_1_err, TYPE_1_E, $reset = false, $against);
}
//*/
$all_matched_tmp = extract_results();
if (!empty($all_matched_tmp['']) || $best == 'no') {
// Type 2
$query_url_2 = $ep . '&fq=latin_part_a:' . urlencode($lpa2) . '&fq=latin_part_bc:(' . urlencode(implode(' OR ', $mix2)) . ")";
extract_results($query_url_2, TYPE_2, $reset = false, $against);
// with minor spell error
foreach (array_unique($mix2) as $p) {
$query_url_2_err_suggestion = $ep . "&rows=0&spellcheck.q=" . urlencode($p);
$suggestion = extract_suggestion($query_url_2_err_suggestion, TYPE_2_E);
if (!empty($suggestion)) {
$suggestions[] = $suggestion;
}
$query_url_2_err_long_suggestion = $ep . "&rows=0&spellcheck.q=" . urlencode("{$lpa2} {$p}");
$long_suggestion = extract_suggestion($query_url_2_err_long_suggestion, TYPE_2_E);
if (!empty($long_suggestion)) {
$long_suggestions[] = $long_suggestion;
}
}
if (!empty($suggestions)) {
$suggestions = array_unique(array_merge($suggestions, $mix2));
$query_url_2_err = $ep . '&fq=latin_part_a:' . urlencode($lpa2) . '&fq=latin_part_bc:(' . urlencode(implode(' OR ', $suggestions)) . ")";
extract_results($query_url_2_err, TYPE_2_E, $reset = false, $against);
}
if (!empty($long_suggestions) && count($mix2) > 1) {
foreach ($long_suggestions as $long_suggestion) {
$query_url_2_err = $ep . '&fq=canonical_name:"' . urlencode($long_suggestion) . '"';
extract_results($query_url_2_err, TYPE_2_E, $reset = false, $against);
}
}
// Genus spell error???
$query_url_2_genus_err_suggestion = $ep . "&rows=0&spellcheck.q=" . urlencode($lpa2);
$suggestion = extract_suggestion($query_url_2_genus_err_suggestion, TYPE_2_GE);
if (is_null($suggestion)) {
$query_url_2_genus_err_suggestion = $ep . "&rows=0&spellcheck.q=" . urlencode($name_cleaned);
$suggestion = array_shift(explode(" ", extract_suggestion($query_url_2_genus_err_suggestion, TYPE_2_GE)));
if (is_null($suggestion)) {
foreach ($mix2 as $mp) {
$query_url_2_genus_err_suggestion = $ep . "&rows=0&spellcheck.q=" . urlencode($lpa2 + ' ' + $mp);
$suggestion = array_shift(explode(" ", extract_suggestion($query_url_2_genus_err_suggestion, TYPE_2_GE)));
if (!is_null($suggestion)) {
break;
}
}
}
}
if (treat_word($lpa2, true) == treat_word($suggestion, true)) {
$query_url_2_genus_err = $ep . '&fq=latin_part_a:' . urlencode($suggestion) . '&fq=latin_part_bc:(' . urlencode(implode(' OR ', $mix2)) . ")";
extract_results($query_url_2_genus_err, TYPE_2_GS, $reset = false, $against);
} elseif (levenshtein($lpa2, $suggestion) == 1 && strlen($lpa2) == strlen($suggestion)) {
$len = strlen($lpa2);
for ($i = 0; $i < $len; $i++) {
if ($lpa2[$i] != $suggestion[$i]) {
if (similar_char($lpa2[$i], $suggestion[$i], @$lpa2[$i + 1], @$suggestion[$i + 1])) {
$query_url_2_genus_err = $ep . '&fq=latin_part_a:' . urlencode($suggestion) . '&fq=latin_part_bc:(' . urlencode(implode(' OR ', $mix2)) . ")";
extract_results($query_url_2_genus_err, TYPE_2_GL, $reset = false, $against);
}
}
}
} elseif (levenshtein($lpa2, $suggestion) == 1) {
$query_url_2_genus_err = $ep . '&fq=latin_part_a:' . urlencode($suggestion) . '&fq=latin_part_bc:(' . urlencode(implode(' OR ', $mix2)) . ")";
extract_results($query_url_2_genus_err, TYPE_2_GL2, $reset = false, $against);
}
$all_matched_tmp = extract_results();
}
if (!empty($all_matched_tmp['']) || $best == 'no') {
// Type 3
$sound = treat_word($name_cleaned);
$query_url_3 = $ep . '&fq=sound_name:"' . urlencode($sound) . '"';
extract_results($query_url_3, TYPE_3_S, $reset = false, $against);
// Type 3 mix
$query_url_3 = $ep . '&fq=sound_part_a:' . urlencode($spa2) . '&fq=sound_part_bc:(' . urlencode(implode(' OR ', $sound_mix2)) . ")";
extract_results($query_url_3, TYPE_3_S2, $reset = false, $against);
$sound_mix2_strip_ending = array_map("treat_word", $mix2, array_fill(0, count($mix2), true));
$query_url_3_strip_bc_ending = $ep . '&fq=sound_part_a:' . urlencode($spa2) . '&fq=sound_part_bc_strip_ending:(' . urlencode(implode(' OR ', $sound_mix2_strip_ending)) . ")";
extract_results($query_url_3_strip_bc_ending, TYPE3_S3, $reset = false, $against);
$query_url_3_strip_all_ending = $ep . '&fq=sound_part_a_strip_ending:' . urlencode(treat_word($spa2, true)) . '&fq=sound_part_bc_strip_ending:(' . urlencode(implode(' OR ', $sound_mix2_strip_ending)) . ")";
extract_results($query_url_3_strip_all_ending, TYPE_3_GUESS, $reset = false, $against);
$all_matched_tmp = extract_results();
}
foreach ($all_matched_tmp as $m) {
$all_matched[$m['matched']] = array_merge(array('name' => $name, 'name_cleaned' => $name_cleaned), $m);
}
/*
echo "<xmp>";
var_dump($all_matched);
echo "</xmp>";
//*/
//var_dump($all_matched);
return $all_matched;
}
$rec['sound_genus'] = $frags[0];
$rec['sound_part_a_strip_ending'] = treat_word($frags[0], true);
$rec['nameSpell'][] = $frags[0];
if (!empty($frags[1])) {
$rec['latin_part_bc'][] = $frags[1];
$rec['nameSpell'][] = $frags[1];
$rec['nameSpell'][] = $frags[0] . " " . $frags[1];
$rec['sound_part_bc'][] = treat_word($frags[1]);
$rec['sound_part_bc_strip_ending'][] = treat_word($frags[1], true);
} else {
continue;
}
if (!empty($frags[2])) {
$rec['latin_part_bc'][] = $frags[2];
$rec['nameSpell'][] = $frags[2];
$rec['nameSpell'][] = $frags[1] . " " . $frags[2];
$rec['nameSpell'][] = $frags[0] . " " . $frags[2];
$rec['nameSpell'][] = $frags[0] . " " . $frags[1] . " " . $frags[2];
$rec['sound_part_bc'][] = treat_word($frags[2]);
$rec['sound_part_bc_strip_ending'][] = treat_word($frags[2], true);
}
$ret[] = $rec;
if ($counter % 1000 == 999) {
submitJson($ret);
$ret = array();
}
$counter++;
}
if (!empty($ret)) {
submitJson($ret);
}
