Add periodic cleanup of orphaned drafts: cleanup job, just command, deploy cron

TODO.md

@@ -1,19 +1,17 @@
 # TODO
 
-> Last updated: 2026-06-11 12:10
+> Last updated: 2026-06-11
 > Context: Form Accessibility & Resilience improvements for XAMXAM thesis submission platform
 
-## In Progress
-
 ## Pending
 
 - [ ] #aria-test-manual Test WCAG changes with VoiceOver and NVDA on full add/edit/partage form flows
 - [ ] #nojs-upload-test Test end-to-end: submit partage form with JS disabled, verify files arrive via `$_FILES`
-- [ ] #cleanup-drafts Add periodic cleanup job for orphaned drafts (`just cleanup-drafts`)
-- [ ] #form-setup-helper Add `ThesisFormSetup` helper class to reduce bootstrap duplication across add/edit/partage `(partage/index.php)` `(admin/add.php)` `(admin/edit.php)`
 
 ## Completed
 
+- [x] #cleanup-drafts Add periodic cleanup job for orphaned drafts (`Database.php`, `justfile`, `deploy/xamxam-cleanup.cron`, `scripts/cleanup-drafts.php`) ✓
+- [x] #form-setup-helper Add `FormBootstrap` helper class to reduce bootstrap duplication across add/edit/partage `(admin/add.php)` `(admin/edit.php)` ✓
 - [x] #two-phase-commit Add two-phase commit: INSERT thesis `status='draft'`, COMMIT, move files, UPDATE to `active` `(ThesisCreateController.php)` ✓
 - [x] #filepond-preserve Preserve FilePond temp file IDs on partage validation redirect `(partage/index.php)` `(FilepondHandler.php)` ✓
 - [x] #refactor-partage Extract partage form page chrome to `templates/partage/form-page.php` `(partage/index.php)` ✓

app/src/Database.php

@@ -2307,6 +2307,59 @@ class Database
         return $newId;
     }
 
+    /**
+     * Find and optionally delete orphaned draft theses older than a threshold.
+     *
+     * Draft theses are created with status='draft' before file operations
+     * (two-phase commit). If the file phase throws after COMMIT, the draft
+     * remains orphaned indefinitely — no files are attached but the row blocks
+     * the identifier number.
+     *
+     * @param int    $olderThanHours  Drafts older than this many hours are candidates.
+     * @param bool   $dryRun          When true, only list candidates without deleting.
+     * @return array{deleted: int, candidates: array}  Deleted count + list of IDs found.
+     */
+    public function cleanupOrphanedDrafts(int $olderThanHours = 24, bool $dryRun = true): array
+    {
+        $cutoff = date('Y-m-d H:i:s', strtotime("-{$olderThanHours} hours"));
+
+        // Draft theses with no files attached = orphaned submissions.
+        $stmt = $this->pdo->prepare(
+            "SELECT t.id, t.identifier, t.title, t.submitted_at
+             FROM theses t
+             WHERE t.status = 'draft'
+               AND t.deleted_at IS NULL
+               AND t.submitted_at < ?
+               AND NOT EXISTS (
+                   SELECT 1 FROM thesis_files tf WHERE tf.thesis_id = t.id
+               )
+             ORDER BY t.submitted_at ASC"
+        );
+        $stmt->execute([$cutoff]);
+        $candidates = $stmt->fetchAll();
+
+        if ($dryRun) {
+            return ['deleted' => 0, 'candidates' => $candidates];
+        }
+
+        $deleted = 0;
+        require_once __DIR__ . '/Audit.php';
+        $actor = Audit::actor();
+        foreach ($candidates as $row) {
+            $id = (int)$row['id'];
+            $old = $this->fetchRow('theses', $id);
+            // Hard-delete the orphan row and its junction records.
+            // Cascade covers thesis_authors, thesis_tags, thesis_files (empty),
+            // thesis_supervisors, thesis_languages, thesis_formats.
+            $this->pdo->prepare('DELETE FROM theses WHERE id = ?')->execute([$id]);
+            Audit::log($this, $actor, 'DELETE', 'theses', $id, $old, null);
+            $deleted++;
+            error_log("[cleanup-drafts] Deleted orphaned draft thesis {$id} ({$row['identifier']}) — submitted {$row['submitted_at']}");
+        }
+
+        return ['deleted' => $deleted, 'candidates' => $candidates];
+    }
+
     /**
      * Soft-delete a single thesis (sets deleted_at).
      */

deploy/xamxam-cleanup.cron

@@ -0,0 +1,7 @@
+# XAMXAM — orphaned draft cleanup cron job
+# Installed to /etc/cron.d/xamxam-cleanup (system cron format: minute hour dom month dow user command)
+#
+# Deletes draft theses older than 24h that have no attached files.
+# Runs every 4 hours — drafts only become eligible after 24h, so this is ample.
+# The script is a dry-run unless --no-dry-run is passed.
+0  */4  *  *  *  www-data  php /var/www/xamxam/scripts/cleanup-drafts.php --no-dry-run >> /var/log/xamxam-cleanup.log 2>&1

justfile

@@ -319,6 +319,27 @@ deploy-list-backups:
     # List all existing backups on the server (most recent last).
     ssh xamxam "ls -lth /var/backups/xamxam/ 2>/dev/null || echo 'No backups yet.'"
 
+[group('deploy')]
+deploy-cleanup-cron:
+    # Install cron job for orphaned draft cleanup (every 4 hours, 24h threshold).
+    # Creates /etc/cron.d/xamxam-cleanup and log file on the server.
+    @echo "📋 Installing draft cleanup cron job…"
+    rsync -v scripts/cleanup-drafts.php xamxam:/var/www/xamxam/scripts/cleanup-drafts.php
+    ssh xamxam "chown www-data:xamxam /var/www/xamxam/scripts/cleanup-drafts.php && chmod 755 /var/www/xamxam/scripts/cleanup-drafts.php"
+    rsync -v deploy/xamxam-cleanup.cron xamxam:/tmp/xamxam-cleanup.cron
+    ssh -t xamxam "sudo install -o root -g root -m 644 /tmp/xamxam-cleanup.cron /etc/cron.d/xamxam-cleanup && rm -f /tmp/xamxam-cleanup.cron"
+    ssh -t xamxam "sudo touch /var/log/xamxam-cleanup.log && sudo chown www-data:www-data /var/log/xamxam-cleanup.log && sudo chmod 644 /var/log/xamxam-cleanup.log"
+    @echo "✅ Cleanup cron installed."
+    @echo "   Cron file: /etc/cron.d/xamxam-cleanup"
+    @echo "   Script:    /var/www/xamxam/scripts/cleanup-drafts.php"
+    @echo "   Log file:  /var/log/xamxam-cleanup.log"
+    @echo ""
+    @echo "Verify with: just deploy-check-cleanup-log"
+
+[group('deploy')]
+deploy-check-cleanup-log:
+    ssh xamxam "tail -20 /var/log/xamxam-cleanup.log 2>/dev/null || echo '(log file empty or missing — will be created on first cron run)'"
+
 [group('deploy')]
 test-restore remote_gz_path:
     # Test-restore a production backup snapshot to a local temp DB and verify.
@@ -351,8 +372,8 @@ deploy-migrate-storage dry_run='' target_host='xamxam':
     ssh {{target_host}} 'rm -f /var/www/xamxam/migrate-storage-paths.php'
 
 [group('deploy')]
-deploy-all-first: deploy deploy-backup
-    # One-shot: full initial deploy including backup cron.
+deploy-all-first: deploy deploy-backup deploy-cleanup-cron
+    # One-shot: full initial deploy including backup and cleanup cron jobs.
 
 # ============================================================================
 # Testing
@@ -434,3 +455,10 @@ clean:
     @rm -f app/error.log
     @rm -rf app/storage/cache/rate_limit/*
     @rm -f /tmp/xamxam-*.log /tmp/xamxam-*.pid
+
+[group('utils')]
+cleanup-drafts dry_run='':
+    # List (dry-run) or delete orphaned draft theses older than 24h.
+    # Pass --no-dry-run to actually delete:
+    #   just cleanup-drafts --no-dry-run
+    @php -r 'define("APP_ROOT", getcwd()."/app");require APP_ROOT."/src/Database.php";$db=new Database();$dry="{{dry_run}}"!=="--no-dry-run";$res=$db->cleanupOrphanedDrafts(24,$dry);$c=count($res["candidates"]);if($c===0){echo"✅ No orphaned drafts found.\n";}elseif($dry){echo"🔍 Found {$c} orphaned draft(s):\n";foreach($res["candidates"]as$row){printf("  → #%d %s \"%s\" (submitted %s)\n",$row["id"],$row["identifier"],$row["title"],$row["submitted_at"]);}echo"\nRun \"just cleanup-drafts --no-dry-run\" to delete them.\n";}else{echo"🗑 Deleted {$res["deleted"]} orphaned draft(s).\n";}'

scripts/cleanup-drafts.php

@@ -0,0 +1,55 @@
+#!/usr/bin/env php
+<?php
+/**
+ * cleanup-drafts.php — Delete orphaned draft theses older than 24h.
+ *
+ * Draft theses are created with status='draft' during the two-phase commit
+ * in ThesisCreateController. If the file phase throws after COMMIT, the
+ * draft remains orphaned — no files attached, but blocks the identifier.
+ *
+ * Usage:
+ *   php scripts/cleanup-drafts.php              # dry-run (list candidates)
+ *   php scripts/cleanup-drafts.php --no-dry-run # actually delete
+ *
+ * Exit codes: 0 on success, 1 on error.
+ */
+
+declare(strict_types=1);
+
+$root = dirname(__DIR__);
+define('APP_ROOT', $root . '/app');
+
+require_once APP_ROOT . '/src/Database.php';
+
+$dryRun = !in_array('--no-dry-run', $argv, true);
+
+try {
+    $db = new Database();
+    $result = $db->cleanupOrphanedDrafts(24, $dryRun);
+} catch (Exception $e) {
+    error_log('[cleanup-drafts] Error: ' . $e->getMessage());
+    exit(1);
+}
+
+$count = count($result['candidates']);
+
+if ($count === 0) {
+    exit(0); // nothing to do — quiet exit
+}
+
+if ($dryRun) {
+    foreach ($result['candidates'] as $row) {
+        printf(
+            "DRY-RUN → #%d %s \"%s\" (submitted %s)\n",
+            $row['id'],
+            $row['identifier'],
+            $row['title'],
+            $row['submitted_at']
+        );
+    }
+    echo "Found {$count} orphaned draft(s). Re-run with --no-dry-run to delete.\n";
+    exit(0);
+}
+
+printf("Deleted %d orphaned draft(s).\n", $result['deleted']);
+exit(0);