deblan/deblan.io-murph
deblan/deblan.io-murph
1
0
Fork 0
Code Issues Pull requests 1 Projects Releases Wiki Activity

upgrade search engine
Some checks failed
ci/woodpecker/push/build Pipeline was successful
Details
ci/woodpecker/deployment/deploy Pipeline failed
Details

Browse source
This commit is contained in:
Simon Vieille 2023-09-24 18:59:33 +02:00
parent 45bd1b408d
commit 138b4f24ee
Signed by: deblan
GPG key ID: 579388D585F70417
1 changed files with 69 additions and 15 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

84
src/Repository/Blog/PostRepositoryQuery.php
View file

@ -63,6 +63,12 @@ class PostRepositoryQuery extends RepositoryQuery
public function search(?string $keywords, ?string $tag)
{
$keywords = explode(' ', $keywords);
$filter = fn($keyword) => trim($keyword) !== '' && preg_match('/[a-zA-Z]+/', $keyword) && mb_strlen($keyword) > 3;
$keywords = array_filter($keywords, $filter);
if ($keywords) {
$conn = $this->repository->getEm()->getConnection();
@ -70,43 +76,91 @@ class PostRepositoryQuery extends RepositoryQuery
'SELECT
post.id,
post.title,
MATCH(post.title) AGAINST(:search) AS MATCH_TITLE,
MATCH(post.content) AGAINST(:search) AS MATCH_CONTENT
post.content,
post.published_at
FROM post
WHERE
post.status = 1 AND
post.published_at < :date
ORDER BY
MATCH_TITLE DESC,
MATCH_CONTENT DESC
');
$statement = $query->execute([
':search' => $keywords,
':date' => (new \DateTime())->format('Y-m-d H:i:s'),
]);
$results = $statement->fetchAll();
$ids = [];
$matches = [];
foreach ($results as $k => $v) {
$rate = ($v['MATCH_TITLE'] * 2) + $v['MATCH_CONTENT'];
$added = false;
$initWords = explode(' ', $v['title']);
$words = [];
if ($rate >= 7) {
$ids[] = $v['id'];
foreach ($initWords as $initWord) {
$words = array_merge($words, preg_split('/[:_-]+/', $initWord));
}
}
if (0 == count($ids)) {
foreach ($results as $k => $v) {
$rate = ($v['MATCH_TITLE'] * 2) + $v['MATCH_CONTENT'];
$words = array_filter($words, $filter);
if ($rate >= 6) {
$ids[] = $v['id'];
foreach ($keywords as $keyword) {
if ($added) {
continue;
}
foreach ($words as $word) {
if ($added) {
continue;
}
if (str_contains(mb_strtolower($word), mb_strtolower($keyword))) {
$matches[] = [
'id' => $v['id'],
'published_at' => $v['published_at'],
'similarity' => 100,
];
$added = true;
} elseif(str_contains($v['content'], $keyword)) {
$matches[] = [
'id' => $v['id'],
'published_at' => $v['published_at'],
'similarity' => 99,
];
$added = true;
} else {
$lev = levenshtein($word, $keyword);
$similarity = 100 - ($lev * 100 / mb_strlen($word));
if ($similarity > 70) {
$matches[] = [
'id' => $v['id'],
'published_at' => $v['published_at'],
'similarity' => $similarity,
];
$added = true;
}
}
}
}
}
usort($matches, function($a, $b) {
if ($a['similarity'] > $b['similarity']) {
return -1;
}
if ($b['similarity'] > $a['similarity']) {
return 1;
}
return ($a['published_at'] <> $b['published_at']) * -1;
});
$ids = array_column($matches, 'id');
if (!$ids) {
$ids = [-1];
}