-Implemented YouTube timestamp correction and added the admin Upkeeping section.

What changed: Added shared YouTube publish-date scraping helpers in [includes/db.php](/Users/tyemeclifford/Documents/GH/mediaplayer/includes/db.php). New YouTube imports now set videos.created_at to the original YouTube publish date when available. Manual Add/Edit saves now also try to correct timestamps for YouTube-backed videos. Added Admin Dashboard → Upkeeping → Correct YouTube timestamps in [admin/index.php](/Users/tyemeclifford/Documents/GH/mediaplayer/admin/index.php), with batch size control for backfilling existing records. Renamed the admin table column from Added to Date, since it may now reflect original publish date. Documented the behavior in [README.md](/Users/tyemeclifford/Documents/GH/mediaplayer/README.md).
2026-06-24 13:40:03 -04:00
parent 99492287e7
commit dc71d26c2a
6 changed files with 205 additions and 8 deletions
@@ -84,6 +84,8 @@ If no thumbnail image is uploaded, the add/edit forms can auto-grab one from the

 Use **Admin → YouTube Importer** to paste a YouTube channel URL, `@handle`, or `UC...` channel ID. The importer scrapes the channel's `/videos` page, extracts video data from `ytInitialData`, previews the results, and imports selected videos as remote YouTube embed sources. Duplicate YouTube watch URLs are skipped.

+Imported YouTube videos use the video's original publish date for the database `created_at` timestamp when YouTube exposes it. The admin dashboard also includes **Upkeeping → Correct YouTube timestamps** to backfill existing YouTube entries in batches.
+
 ### Public Search

 The full catalogue page includes a public search for published videos only. Search matches video title, description, and slug; draft videos remain hidden.
@@ -119,6 +119,7 @@ if ($_SERVER['REQUEST_METHOD'] === 'POST') {
                        $db->prepare("UPDATE videos SET thumbnail=? WHERE id=?")->execute([$thumbnail, $video_id]);
                    }
                }
+                media_apply_youtube_timestamp($video_id);

                header('Location: edit.php?id=' . $video_id . '&saved=1');
                exit;
@@ -190,6 +190,8 @@ if ($_SERVER['REQUEST_METHOD'] === 'POST') {
                            }
                        }

+                        media_apply_youtube_timestamp($id);
+
                        header('Location: edit.php?id=' . $id . '&saved=1');
                        exit;
                    }
@@ -39,6 +39,16 @@ if ($_SERVER['REQUEST_METHOD'] === 'POST') {
        $db->prepare("UPDATE videos SET published=? WHERE id=?")->execute([$cur ? 0 : 1, $id]);
        $msg = 'Video ' . ($cur ? 'unpublished' : 'published') . '.';
    }
+
+    if ($action === 'correct_youtube_timestamps') {
+        $limit = max(1, min(100, (int)($_POST['timestamp_limit'] ?? 25)));
+        $result = media_correct_youtube_timestamps($limit);
+        $msg = 'Upkeeping checked ' . $result['checked'] . ' YouTube video' . ($result['checked'] === 1 ? '' : 's') . ': ';
+        $msg .= $result['corrected'] . ' corrected';
+        if ($result['already_correct']) $msg .= ', ' . $result['already_correct'] . ' already correct';
+        if ($result['not_found']) $msg .= ', ' . $result['not_found'] . ' without a readable original date';
+        $msg .= '.';
+    }
 }

 // Search + pagination
@@ -73,6 +83,7 @@ $list_stmt->bindValue(':limit',  $per_page, PDO::PARAM_INT);
 $list_stmt->bindValue(':offset', $offset,   PDO::PARAM_INT);
 $list_stmt->execute();
 $videos = $list_stmt->fetchAll();
+$youtube_timestamp_candidates = media_youtube_timestamp_candidate_count();

 $qs = fn($p) => '?' . http_build_query(array_filter(['q' => $search, 'page' => $p]));

@@ -130,6 +141,12 @@ render_head('Admin — TyClifford.com', '

  .admin-footer-bar { display:flex; align-items:center; justify-content:space-between; gap:1rem; flex-wrap:wrap; }
  .rec-count { font-family:"Share Tech Mono",monospace; font-size:.63rem; color:var(--text-3); letter-spacing:.06em; }
+  .section-heading { font-family:"Share Tech Mono",monospace; font-size:.65rem; letter-spacing:.15em; text-transform:uppercase; color:var(--text-3); margin-bottom:.75rem; }
+  .hint { font-size:.72rem; color:var(--text-3); margin-top:.2rem; line-height:1.5; }
+  .upkeep-row { display:flex; align-items:flex-end; justify-content:space-between; gap:1rem; flex-wrap:wrap; }
+  .upkeep-copy { max-width:620px; }
+  .upkeep-form { display:flex; align-items:flex-end; gap:.65rem; flex-wrap:wrap; }
+  .upkeep-form .form-group { min-width:120px; }
 ');
 ?>
 <main class="page" role="main">
@@ -183,6 +200,29 @@ render_head('Admin — TyClifford.com', '
    <!-- Main content -->
    <div class="admin-main">

+      <div class="card">
+        <p class="section-heading">Upkeeping</p>
+        <div class="upkeep-row">
+          <div class="upkeep-copy">
+            <h2 class="card-title">Correct YouTube timestamps</h2>
+            <p class="hint">
+              Fetches original YouTube publish dates and updates the database timestamp shown in the admin list and public catalogue.
+              <?= $youtube_timestamp_candidates ?> YouTube-backed video<?= $youtube_timestamp_candidates === 1 ? '' : 's' ?> currently match this upkeep task.
+            </p>
+          </div>
+          <form class="upkeep-form" method="post" action="index.php">
+            <input type="hidden" name="action" value="correct_youtube_timestamps">
+            <div class="form-group">
+              <label class="form-label" for="timestamp_limit">Batch Size</label>
+              <input class="form-input" type="number" id="timestamp_limit" name="timestamp_limit" value="25" min="1" max="100">
+            </div>
+            <button class="btn btn-secondary btn-sm" type="submit">
+              Correct Dates
+            </button>
+          </form>
+        </div>
+      </div>
+
      <div class="admin-toolbar">
        <form class="search-form" method="get" action="index.php">
          <input class="form-input" type="search" name="q"
@@ -214,7 +254,7 @@ render_head('Admin — TyClifford.com', '
              <th>Duration</th>
              <th>Sources</th>
              <th>Status</th>
-              <th>Added</th>
+              <th>Date</th>
              <th class="actions-cell">Actions</th>
            </tr>
          </thead>
@@ -486,10 +486,8 @@ function yt_import_videos(PDO $db, array $videos, array $selected_ids, bool $pub
        try {
            $slug = make_slug($video['title']);
            $thumbnail = yt_download_thumbnail($video['thumbnail_url'], $slug);
-            $db->prepare("
-                INSERT INTO videos (title, description, slug, thumbnail, duration, sort_order, published)
-                VALUES (?, ?, ?, ?, ?, ?, ?)
-            ")->execute([
+            $published_at = media_youtube_publish_datetime($video['watch_url']);
+            $video_params = [
                $video['title'],
                $video['description'],
                $slug,
@@ -497,7 +495,19 @@ function yt_import_videos(PDO $db, array $videos, array $selected_ids, bool $pub
                (int)$video['duration'],
                0,
                $published ? 1 : 0,
-            ]);
+            ];
+            if ($published_at !== '') {
+                $video_params[] = $published_at;
+                $db->prepare("
+                    INSERT INTO videos (title, description, slug, thumbnail, duration, sort_order, published, created_at, updated_at)
+                    VALUES (?, ?, ?, ?, ?, ?, ?, ?, datetime('now'))
+                ")->execute($video_params);
+            } else {
+                $db->prepare("
+                    INSERT INTO videos (title, description, slug, thumbnail, duration, sort_order, published)
+                    VALUES (?, ?, ?, ?, ?, ?, ?)
+                ")->execute($video_params);
+            }
            $video_id = (int)$db->lastInsertId();
            $db->prepare("
                INSERT INTO video_sources (video_id, label, source_type, file_path, source_url, mime_type, quality, sort_order)
@@ -306,13 +306,17 @@ function media_append_query(string $base, array $params): string {
    return $base . (strpos($base, '?') === false ? '?' : '&') . $query;
 }

-function media_http_get(string $url, int $timeout = 12): array {
+function media_http_get(string $url, int $timeout = 12, string $accept = ''): array {
    $url = normalize_media_url($url);
    if ($url === '') return ['body' => '', 'code' => 0, 'content_type' => '', 'error' => 'Invalid URL.'];

+    if ($accept === '') {
+        $accept = 'image/avif,image/webp,image/png,image/jpeg,image/*;q=0.9,application/json;q=0.8,text/html;q=0.7,*/*;q=0.5';
+    }
    $headers = [
        'User-Agent: Mozilla/5.0 (compatible; TyCliffordMedia/1.0)',
-        'Accept: image/avif,image/webp,image/png,image/jpeg,image/*;q=0.9,application/json;q=0.8,text/html;q=0.7,*/*;q=0.5',
+        'Accept: ' . $accept,
+        'Accept-Language: en-US,en;q=0.9',
    ];

    if (function_exists('curl_init')) {
@@ -411,6 +415,144 @@ function media_youtube_id_from_url(string $url): string {
    return media_clean_token($id);
 }

+function media_normalize_datetime(string $value): string {
+    $value = trim(html_entity_decode($value, ENT_QUOTES | ENT_HTML5, 'UTF-8'));
+    if ($value === '') return '';
+
+    if (preg_match('/^(\d{4}-\d{2}-\d{2})/', $value, $m)) {
+        return $m[1] . ' 00:00:00';
+    }
+
+    try {
+        $date = new DateTimeImmutable($value);
+        return $date->format('Y-m-d H:i:s');
+    } catch (Throwable $e) {
+        return '';
+    }
+}
+
+function media_youtube_publish_datetime_from_html(string $html): string {
+    $patterns = [
+        '/<meta[^>]+itemprop=["\']datePublished["\'][^>]+content=["\']([^"\']+)["\']/i',
+        '/<meta[^>]+content=["\']([^"\']+)["\'][^>]+itemprop=["\']datePublished["\']/i',
+        '/"datePublished"\s*:\s*"([^"]+)"/',
+        '/"publishDate"\s*:\s*"([^"]+)"/',
+        '/"uploadDate"\s*:\s*"([^"]+)"/',
+    ];
+
+    foreach ($patterns as $pattern) {
+        if (!preg_match($pattern, $html, $match)) continue;
+        $value = json_decode('"' . str_replace('"', '\"', $match[1]) . '"');
+        $date = media_normalize_datetime(is_string($value) ? $value : $match[1]);
+        if ($date !== '') return $date;
+    }
+
+    return '';
+}
+
+function media_youtube_publish_datetime(string $url): string {
+    $id = media_youtube_id_from_url($url);
+    if ($id === '') return '';
+
+    $watch_url = 'https://www.youtube.com/watch?v=' . rawurlencode($id);
+    $fetch = media_http_get($watch_url, 15, 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8');
+    if ($fetch['body'] === '' || $fetch['error'] !== '') return '';
+
+    return media_youtube_publish_datetime_from_html($fetch['body']);
+}
+
+function media_youtube_source_url_for_video(int $video_id): string {
+    $stmt = get_db()->prepare("
+        SELECT source_url
+        FROM video_sources
+        WHERE video_id=? AND source_type='remote' AND source_url!=''
+        ORDER BY sort_order ASC, id ASC
+    ");
+    $stmt->execute([$video_id]);
+
+    foreach ($stmt->fetchAll() as $source) {
+        $url = normalize_media_url($source['source_url'] ?? '');
+        if ($url !== '' && media_youtube_id_from_url($url) !== '') return $url;
+    }
+
+    return '';
+}
+
+function media_apply_youtube_timestamp(int $video_id): array {
+    $url = media_youtube_source_url_for_video($video_id);
+    if ($url === '') return ['status' => 'no_youtube', 'date' => '', 'url' => ''];
+
+    $published_at = media_youtube_publish_datetime($url);
+    if ($published_at === '') return ['status' => 'not_found', 'date' => '', 'url' => $url];
+
+    $stmt = get_db()->prepare("SELECT created_at FROM videos WHERE id=?");
+    $stmt->execute([$video_id]);
+    $current = (string)$stmt->fetchColumn();
+
+    if (substr($current, 0, 10) === substr($published_at, 0, 10)) {
+        return ['status' => 'already_correct', 'date' => $published_at, 'url' => $url];
+    }
+
+    get_db()->prepare("UPDATE videos SET created_at=?, updated_at=datetime('now') WHERE id=?")
+            ->execute([$published_at, $video_id]);
+
+    return ['status' => 'corrected', 'date' => $published_at, 'url' => $url];
+}
+
+function media_youtube_timestamp_candidates(int $limit): array {
+    $limit = max(1, min(100, $limit));
+    $stmt = get_db()->prepare("
+        SELECT DISTINCT v.id
+        FROM videos v
+        INNER JOIN video_sources vs ON vs.video_id = v.id
+        WHERE vs.source_type='remote'
+          AND (
+            vs.source_url LIKE '%youtube.com%'
+            OR vs.source_url LIKE '%youtu.be%'
+            OR vs.source_url LIKE '%youtube-nocookie.com%'
+          )
+        ORDER BY v.id DESC
+        LIMIT :limit
+    ");
+    $stmt->bindValue(':limit', $limit, PDO::PARAM_INT);
+    $stmt->execute();
+    return array_map('intval', array_column($stmt->fetchAll(), 'id'));
+}
+
+function media_youtube_timestamp_candidate_count(): int {
+    return (int)get_db()->query("
+        SELECT COUNT(DISTINCT v.id)
+        FROM videos v
+        INNER JOIN video_sources vs ON vs.video_id = v.id
+        WHERE vs.source_type='remote'
+          AND (
+            vs.source_url LIKE '%youtube.com%'
+            OR vs.source_url LIKE '%youtu.be%'
+            OR vs.source_url LIKE '%youtube-nocookie.com%'
+          )
+    ")->fetchColumn();
+}
+
+function media_correct_youtube_timestamps(int $limit = 25): array {
+    $result = [
+        'checked' => 0,
+        'corrected' => 0,
+        'already_correct' => 0,
+        'not_found' => 0,
+        'no_youtube' => 0,
+    ];
+
+    foreach (media_youtube_timestamp_candidates($limit) as $video_id) {
+        $result['checked']++;
+        $applied = media_apply_youtube_timestamp($video_id);
+        $status = $applied['status'] ?? 'not_found';
+        if (!isset($result[$status])) $status = 'not_found';
+        $result[$status]++;
+    }
+
+    return $result;
+}
+
 function media_provider_oembed_endpoint(string $url, string $host): string {
    if (media_host_matches($host, 'youtube.com') || media_host_matches($host, 'youtu.be') || media_host_matches($host, 'youtube-nocookie.com')) {
        return 'https://www.youtube.com/oembed?format=json&url=' . rawurlencode($url);