<?php
declare(strict_types=1);

require_once __DIR__ . '/../includes/bootstrap.php';

dbnToolsRequireMethod('GET');
dbnToolsRequireAuth();

try {
    $ragDb = dbnToolsRagDb();
    $bnlDb = dbnToolsDb();

    $category   = isset($_GET['category'])    && $_GET['category']    !== '' ? trim((string)$_GET['category'])    : null;
    $sourceName = isset($_GET['source_name']) && $_GET['source_name'] !== '' ? trim((string)$_GET['source_name']) : null;
    $titleFilter = isset($_GET['title'])      && $_GET['title']       !== '' ? trim((string)$_GET['title'])       : null;
    $offset     = max(0, (int)($_GET['offset'] ?? 0));
    $limit      = max(1, min(50, (int)($_GET['limit'] ?? 20)));
    $orderBy    = match($_GET['sort'] ?? 'newest') {
        'oldest' => 'd.updated_at ASC',
        'alpha'  => 'd.title ASC',
        'chunks' => 'chunk_count DESC',
        default  => 'd.updated_at DESC',
    };

    // Build WHERE clause
    $where  = ["d.corpus_id = 1", "d.status = 'ready'"];
    $params = [];

    if ($category !== null) {
        $where[]  = 'd.category = ?';
        $params[] = $category;
    }

    if ($titleFilter !== null) {
        $where[]  = 'd.title LIKE ?';
        $params[] = '%' . str_replace(['%', '_'], ['\\%', '\\_'], $titleFilter) . '%';
    }

    if ($sourceName !== null) {
        // Filter by source via a JOIN to corpus_sources on category match
        // or by matching the scraper's URL pattern in source_url
        // We join bnl_admin.corpus_sources — but that's a different DB.
        // Simplest: filter documents whose source_url LIKE the source's url.
        // Fetch the source URL from bnl_admin first.
        $srcStmt = $bnlDb->prepare(
            "SELECT url FROM corpus_sources WHERE corpus_id = 1 AND name = ? LIMIT 1"
        );
        $srcStmt->execute([$sourceName]);
        $srcRow = $srcStmt->fetch(PDO::FETCH_ASSOC);
        if ($srcRow && !empty($srcRow['url'])) {
            $parsed = parse_url($srcRow['url']);
            $host   = $parsed['host'] ?? '';
            if ($host !== '') {
                $where[]  = "d.source_url LIKE ?";
                $params[] = '%' . $host . '%';
            }
        }
    }

    $whereStr = implode(' AND ', $where);

    // Total count
    $countParams = $params;
    $countStmt   = $ragDb->prepare("SELECT COUNT(*) FROM documents d WHERE $whereStr");
    $countStmt->execute($countParams);
    $total = (int)$countStmt->fetchColumn();

    // Paginated rows — LIMIT/OFFSET interpolated as ints (MariaDB rejects bound params here)
    $dataStmt = $ragDb->prepare(
        "SELECT d.id, d.title, d.category, d.source_url, d.language, d.updated_at,
                COUNT(c.id) AS chunk_count
         FROM documents d
         LEFT JOIN chunks c ON c.document_id = d.id
         WHERE $whereStr
         GROUP BY d.id
         ORDER BY $orderBy
         LIMIT $limit OFFSET $offset"
    );
    $dataStmt->execute($params);
    $documents = $dataStmt->fetchAll(PDO::FETCH_ASSOC);

    // Normalise chunk_count to int
    foreach ($documents as &$doc) {
        $doc['chunk_count'] = (int)$doc['chunk_count'];
    }
    unset($doc);

    dbnToolsRespond([
        'ok'        => true,
        'documents' => $documents,
        'total'     => $total,
        'offset'    => $offset,
        'limit'     => $limit,
        'filter'    => [
            'category'    => $category,
            'source_name' => $sourceName,
            'title'       => $titleFilter,
            'sort'        => $_GET['sort'] ?? 'newest',
        ],
    ]);
} catch (Throwable $e) {
    dbnToolsError('Could not load documents: ' . $e->getMessage(), 500, 'documents_error');
}