Fix the possible-duplicates finder; share it as SimilarTicketService (#113)

check_duplicates.php's logic moves to services/SimilarTicketService.php
(used next by the MCP find_similar_tickets tool). Fixes found while
testing it:

- Non-admins never got matches: the query had no `t` alias but the
  visibility filter's SQL uses t.*, so it failed and the error was
  swallowed into "no duplicates".
- The word-overlap scoring could never fire, since candidates were only
  whole-title substring or SOUNDEX matches. Tickets sharing a significant
  word (4+ letters) are now candidates too; overlap needs 2+ shared words
  so one common word (e.g. every automated ticket's [problem] tag) isn't
  a match.
- SOUNDEX compared PHP's 4-char code (effectively the first word) with
  MariaDB's full-length code: the SQL side almost never matched, and the
  scorer marked any two titles sharing a first word as "sounds alike".
  Both sides now use SQL SOUNDEX on the full title.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MGDKHiU5RJdo3dqQUDow3X
This commit is contained in:
2026-09-26 17:08:04 -04:00
co-authored by Claude Opus 5.5
parent 72feafb359
commit a9bc72fdcc
2 changed files with 153 additions and 98 deletions
+2 -98
View File
@@ -8,7 +8,7 @@
require_once __DIR__ . '/bootstrap.php';
require_once dirname(__DIR__) . '/helpers/ResponseHelper.php';
require_once dirname(__DIR__) . '/models/TicketModel.php';
require_once dirname(__DIR__) . '/services/SimilarTicketService.php';
// Only accept GET requests
if ($_SERVER['REQUEST_METHOD'] !== 'GET') {
@@ -18,100 +18,4 @@ if ($_SERVER['REQUEST_METHOD'] !== 'GET') {
// Get title parameter
$title = isset($_GET['title']) ? trim($_GET['title']) : '';
if (strlen($title) < 5) {
ResponseHelper::success(['duplicates' => []]);
}
// Search for similar titles
// Use both LIKE for substring matching and SOUNDEX for phonetic matching
$duplicates = [];
// Prepare search term for LIKE
$searchTerm = '%' . $title . '%';
// Get SOUNDEX of title
$soundexTitle = soundex($title);
// Build visibility filter so users only see titles they have access to
$ticketModel = new TicketModel($conn);
$visFilter = $ticketModel->getVisibilityFilter($currentUser);
// First, search for exact substring matches (case-insensitive)
$sql = "SELECT ticket_id, title, status, priority, created_at
FROM tickets
WHERE (
title LIKE ?
OR SOUNDEX(title) = ?
)
AND status != 'Closed'
AND ({$visFilter['sql']})
ORDER BY created_at DESC
LIMIT 10";
$types = "ss" . $visFilter['types'];
$params = array_merge([$searchTerm, $soundexTitle], $visFilter['params']);
// Duplicate detection is advisory (it must not block ticket creation), so on any
// DB error degrade gracefully to "no duplicates" rather than fataling the request.
// mysqli may throw (default exception mode) or return false depending on config.
try {
$stmt = $conn->prepare($sql);
if (!$stmt) {
throw new RuntimeException('prepare failed: ' . $conn->error);
}
if (!empty($params)) {
$stmt->bind_param($types, ...$params);
}
$stmt->execute();
$result = $stmt->get_result();
if ($result === false) {
// Non-exception mysqli mode: execute/get_result return false instead of
// throwing. Treat as a query failure so we don't fatal on $result below.
throw new RuntimeException('query failed: ' . $conn->error);
}
} catch (Throwable $e) {
error_log('check_duplicates: ' . $e->getMessage());
ResponseHelper::success(['duplicates' => []]);
}
while ($row = $result->fetch_assoc()) {
// Calculate similarity score
$similarity = 0;
// Check for exact substring match
if (stripos($row['title'], $title) !== false) {
$similarity = 90;
// Check SOUNDEX match
} elseif (soundex($row['title']) === $soundexTitle) {
$similarity = 70;
// Check word overlap
} else {
$titleWords = array_map('strtolower', preg_split('/\s+/', $title));
$rowWords = array_map('strtolower', preg_split('/\s+/', $row['title']));
$matchingWords = array_intersect($titleWords, $rowWords);
$similarity = (count($matchingWords) / max(count($titleWords), 1)) * 60;
}
if ($similarity >= 30) {
$duplicates[] = [
'ticket_id' => $row['ticket_id'],
'title' => $row['title'],
'status' => $row['status'],
'priority' => $row['priority'],
'created_at' => $row['created_at'],
'similarity' => round($similarity)
];
}
}
$stmt->close();
// Sort by similarity descending
usort($duplicates, function ($a, $b) {
return $b['similarity'] - $a['similarity'];
});
// Limit to top 5
$duplicates = array_slice($duplicates, 0, 5);
ResponseHelper::success(['duplicates' => $duplicates]);
ResponseHelper::success(['duplicates' => SimilarTicketService::find($conn, $currentUser, $title)]);