Forum import: archived comments posted under their original author names

One account per original author (nickname = name as signed on the old
site, group 'Original site'), display-name driver 'museum' keeps the
brackets; --replace re-imports threads without real replies.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDRjtmAiNpo5mK5N3d7QwZ

content/museum/apple-ii.md

@@ -67,7 +67,7 @@ params:
   - docs
   - videos
   - comments
-  forum_thread: 1
+  forum_thread: 7
   photos:
   - title: INSIDE AN APPLE II
     subtitle: ''

content/museum/commodore-c64.md

@@ -72,7 +72,7 @@ params:
   - docs
   - videos
   - comments
-  forum_thread: 2
+  forum_thread: 8
   photos:
   - title: COMMODORE 64
     subtitle: ''

content/museum/kaypro-pc-286i.md

@@ -67,7 +67,7 @@ params:
   tabs:
   - photos
   - comments
-  forum_thread: 5
+  forum_thread: 11
   photos:
   - title: COMPLETE SYSTEM
     subtitle: ''

content/museum/magnavox-odyssey-100.md

@@ -54,7 +54,7 @@ params:
   - links
   - docs
   - comments
-  forum_thread: 6
+  forum_thread: 12
   photos:
   - title: ODYSSEY 100
     subtitle: ''

content/museum/magnavox-odyssey.md

@@ -52,7 +52,7 @@ params:
   - links
   - videos
   - comments
-  forum_thread: 4
+  forum_thread: 10
   photos:
   - title: MAIN UNIT
     subtitle: ''

content/museum/nec-pc-6601.md

@@ -69,7 +69,7 @@ params:
   - adverts
   - links
   - comments
-  forum_thread: 3
+  forum_thread: 9
   photos:
   - title: PC-6601
     subtitle: ''

data/forum.json

@@ -1,8 +1,8 @@
 {
-    "68": 1,
+    "68": 7,
-    "98": 2,
+    "98": 8,
-    "395": 3,
+    "395": 9,
-    "883": 4,
+    "883": 10,
-    "1044": 5,
+    "1044": 11,
-    "1268": 6
+    "1268": 12
 }

forum/README.md

@@ -14,12 +14,19 @@ frame.
 
 - `../oc/scripts/06_forum.py` writes `import/machines.json`: one entry per machine with its
   archived comments, oldest first.
-- `php flarum oc:import` creates one discussion per machine: an opening post linking to the
+- `php flarum oc:import` creates one discussion per machine: an opening post by the user
-  museum page, then one post per archived comment, all by the user `Archive`, with the original
+  `Archive` linking to the museum page, then one post per archived comment, dated as the
-  author and date at the top of each post and the original date as the post date. Discussions go
+  original. Each comment is posted by an account standing for its original author, shown under
+  the name used on the old site ("darius (croatia)"); the same name on several machines is one
+  account. These accounts are in the group "Original site" (a badge on their posts), have a
+  random password and an undeliverable address, so nobody can log in as them. Discussions go
   under three tags (Computers, Consoles, Pongs) that are hidden from "All discussions".
+- Original names are stored as nicknames (Nicknames extension, which must be enabled) and shown
+  by the display-name driver `museum` of this extension, which the import selects: the stock
+  nickname driver would remove the brackets.
 - The command records what it imported (table `oc_threads`) and skips it on later runs, so it
-  can be run again when new machines arrive. It writes the map machine → discussion to
+  can be run again when new machines arrive. `--replace` deletes and re-imports the imported
+  threads that have no reply from a real member (their thread numbers change). It writes the map machine → discussion to
   `forum.json`.
 - Copied to `../data/forum.json`, that map makes `make content` give every machine a Comments
   tab pointing at its thread (`forum_thread` in the page front matter); the old comments are
@@ -72,8 +79,8 @@ are never served. This layout is tested in the local image.
    over SSH, as `http`, in the forum folder: `php flarum install --file=install.yml` (format in
    `docker/entrypoint.sh`). Do not leave the folder online uninstalled: whoever opens the
    installer first becomes its administrator.
-5. Enable the extensions Embed and "Old-Computers museum" (admin panel, or
+5. Enable the extensions Nicknames, Embed and "Old-Computers museum" (admin panel, or
-   `php flarum extension:enable flarum-embed` and `oc-museum`). Tags is enabled by default.
+   `php flarum extension:enable` with `flarum-nicknames`, `flarum-embed`, `oc-museum`). Tags is enabled by default.
 6. In the admin panel: e-mail settings (registration confirms addresses by e-mail), and a spam
    filter extension. Schedule a database dump.
 7. Import, from a shell on the server (SSH, or a one-off Task Scheduler script run as `http`),

forum/docker/entrypoint.sh

@@ -27,7 +27,7 @@ fi
 # the museum extension is mounted from the repository (forum/extension)
 composer require --no-interaction "oc/flarum-museum:*@dev" >/dev/null 2>&1 || composer require --no-interaction "oc/flarum-museum:*@dev"
 php flarum migrate
-for ext in flarum-tags flarum-markdown flarum-embed oc-museum; do php flarum extension:enable "$ext" || true; done
+for ext in flarum-tags flarum-markdown flarum-nicknames flarum-embed oc-museum; do php flarum extension:enable "$ext" || true; done
 php flarum cache:clear
 chown -R www-data:www-data storage public/assets
 exec docker-php-entrypoint "$@"

forum/extension/extend.php

@@ -2,10 +2,14 @@
 
 use Flarum\Extend;
 use Oc\Museum\Console\ImportCommand;
+use Oc\Museum\DisplayName;
 
 return [
     (new Extend\Console())->command(ImportCommand::class),
 
+    // selected by oc:import (setting display_name_driver = museum); needs the Nicknames extension
+    (new Extend\User())->displayNameDriver('museum', DisplayName::class),
+
     // threads shown inside the museum pages (flarum/embed) take the look of the static site
     (new Extend\Frontend('embed'))->css(__DIR__.'/less/embed.less'),
 ];

forum/extension/src/Console/ImportCommand.php

@@ -5,10 +5,13 @@ namespace Oc\Museum\Console;
 use Carbon\Carbon;
 use Flarum\Console\AbstractCommand;
 use Flarum\Discussion\Discussion;
+use Flarum\Group\Group;
 use Flarum\Post\CommentPost;
+use Flarum\Settings\SettingsRepositoryInterface;
 use Flarum\Tags\Tag;
 use Flarum\User\User;
 use Illuminate\Database\ConnectionInterface;
+use Oc\Museum\DisplayName;
 use Illuminate\Support\Str;
 use Symfony\Component\Console\Input\InputArgument;
 use Symfony\Component\Console\Input\InputOption;
@@ -17,18 +20,30 @@ use Symfony\Component\Console\Input\InputOption;
  * Creates one discussion per museum machine from the file written by the extraction
  * pipeline (oc/scripts/06_forum.py), with the archived visitor comments as its posts.
  *
- * Safe to run again: machines already imported (table oc_threads) are skipped.
+ * Each comment is posted by an account standing for its original author: the name as it
+ * appeared on the site ("darius (croatia)") is the account's nickname, shown instead of the
+ * generated user name. These accounts are in the group "Original site" and nobody can log in
+ * as them (random password, undeliverable address).
+ *
+ * Safe to run again: machines already imported (table oc_threads) are skipped. With
+ * --replace, imported threads that nobody has replied to yet are deleted and imported again.
  */
 class ImportCommand extends AbstractCommand
 {
     private const ARCHIVE_USER = 'Archive';
+    private const GROUP = 'Original site';
+
+    /** @var array<string, User> authors of archived comments, by lower-cased original name */
+    private array $authors = [];
+    private Group $group;
+    private int $authorSeq = 0;
     private const TAGS = [
         'Computers' => ['#9a3412', 'fas fa-desktop'],
         'Consoles' => ['#1d4ed8', 'fas fa-gamepad'],
         'Pongs' => ['#15803d', 'fas fa-table-tennis-paddle-ball'],
     ];
 
-    public function __construct(private ConnectionInterface $db)
+    public function __construct(private ConnectionInterface $db, private SettingsRepositoryInterface $settings)
     {
         parent::__construct();
     }
@@ -39,6 +54,7 @@ class ImportCommand extends AbstractCommand
             ->setDescription('Import the museum machines and their archived comments as discussions')
             ->addArgument('file', InputArgument::REQUIRED, 'machines.json written by 06_forum.py')
             ->addOption('site-url', null, InputOption::VALUE_REQUIRED, 'Address of the static site, e.g. https://example.org/oldcomputers', '')
+            ->addOption('replace', null, InputOption::VALUE_NONE, 'Delete and re-import the imported threads that have no reply from a real member')
             ->addOption('map', null, InputOption::VALUE_REQUIRED, 'Where to write the machine -> discussion map (forum.json)');
     }
 
@@ -50,9 +66,22 @@ class ImportCommand extends AbstractCommand
 
             return 1;
         }
+        if (! $this->db->getSchemaBuilder()->hasColumn('users', 'nickname')) {
+            $this->error('Enable the Nicknames extension first: php flarum extension:enable flarum-nicknames');
+
+            return 1;
+        }
+        // original names ("darius (croatia)") are not valid user names: show them as nicknames
+        $this->settings->set('display_name_driver', 'museum');
+
         $site = rtrim((string) $this->input->getOption('site-url'), '/');
         $archive = $this->archiveUser();
+        $this->group = $this->group();
+        $this->loadAuthors();
         $tags = $this->tags();
+        if ($this->input->getOption('replace')) {
+            $this->removeUntouchedThreads();
+        }
         $done = $this->db->table('oc_threads')->pluck('discussion_id', 'oc_id');
         $created = $posts = 0;
 
@@ -72,6 +101,7 @@ class ImportCommand extends AbstractCommand
         }
 
         $archive->refreshCommentCount()->refreshDiscussionCount()->save();
+        $this->refreshAuthorCounts();
         foreach ($tags as $tag) {
             $tag->discussion_count = $tag->discussions()->count();
             $tag->refreshLastPostedDiscussion()->save();
@@ -99,9 +129,7 @@ class ImportCommand extends AbstractCommand
             "Comments about the [{$machine['title']}]($page) page of the museum.\n\n"
             .($comments ? 'The first messages below were left by visitors of the original site.' : 'Be the first to comment!'));
         foreach ($comments as $comment) {
-            $header = '**'.$this->plain($comment['author']).'**'
+            $this->post($discussion, $this->author($comment['author']), Carbon::parse($comment['date']), $comment['text']);
-                .($comment['original_date'] ? ' — '.$this->plain($comment['original_date']) : '');
-            $this->post($discussion, $archive, Carbon::parse($comment['date']), $header."\n\n".$comment['text']);
         }
 
         $discussion->setFirstPost($first)->refreshCommentCount()->refreshLastPost()->refreshParticipantCount()->save();
@@ -124,9 +152,84 @@ class ImportCommand extends AbstractCommand
         return $post;
     }
 
-    private function plain(string $text): string
+    /** Account standing for the author of archived comments, created on first use. */
+    private function author(string $name): User
     {
-        return preg_replace('/([\\\\*_`\[\]@])/', '\\\\$1', trim($text));
+        $name = Str::limit(trim(preg_replace('/\s+/', ' ', $name)) ?: 'Anonymous', 150, '');
+        $key = Str::lower($name);
+        if (! isset($this->authors[$key])) {
+            $user = new User();
+            $user->username = $this->username($name);
+            $user->nickname = $name;
+            $user->email = $user->username.DisplayName::ARCHIVE_DOMAIN;
+            $user->password = Str::random(40);  // nobody logs in as an archived author
+            $user->is_email_confirmed = true;
+            $user->joined_at = Carbon::now();
+            $user->save();
+            $user->groups()->attach($this->group->id);
+            $this->authors[$key] = $user;
+        }
+
+        return $this->authors[$key];
+    }
+
+    /** Valid, unique user name: "darius (croatia)" -> "darius-croatia-a12". */
+    private function username(string $name): string
+    {
+        $base = trim(Str::limit(Str::slug($name), 20, ''), '-') ?: 'visitor';
+
+        return $base.'-a'.(++$this->authorSeq);
+    }
+
+    private function loadAuthors(): void
+    {
+        foreach ($this->group->users()->get() as $user) {
+            $this->authors[Str::lower((string) $user->nickname)] = $user;
+            $this->authorSeq = max($this->authorSeq, (int) Str::afterLast($user->username, '-a'));
+        }
+    }
+
+    private function refreshAuthorCounts(): void
+    {
+        $p = $this->db->getTablePrefix();
+        $this->db->update(
+            "update {$p}users set comment_count = (select count(*) from {$p}posts where {$p}posts.user_id = {$p}users.id and {$p}posts.type = 'comment')
+             where id in (select user_id from {$p}group_user where group_id = ?)",
+            [$this->group->id]
+        );
+    }
+
+    /** --replace: forget imported threads in which only archive accounts have posted. */
+    private function removeUntouchedThreads(): void
+    {
+        $archived = $this->group->users()->pluck('users.id')->push($this->archiveUser()->id)->all();
+        $removed = $kept = 0;
+        foreach ($this->db->table('oc_threads')->pluck('discussion_id', 'oc_id') as $ocId => $discussionId) {
+            $discussion = Discussion::find($discussionId);
+            if ($discussion && $discussion->posts()->whereNotIn('user_id', $archived)->exists()) {
+                $kept++;
+                continue;
+            }
+            $discussion?->delete();
+            $this->db->table('oc_threads')->where('oc_id', $ocId)->delete();
+            $removed++;
+        }
+        $this->info("$removed imported threads removed for re-import, $kept kept (they have replies).");
+    }
+
+    private function group(): Group
+    {
+        $group = Group::query()->where('name_singular', self::GROUP)->first();
+        if (! $group) {
+            $group = new Group();
+            $group->name_singular = self::GROUP;
+            $group->name_plural = self::GROUP;
+            $group->color = '#8a7f66';
+            $group->icon = 'fas fa-box-archive';
+            $group->save();
+        }
+
+        return $group;
     }
 
     private function archiveUser(): User

forum/extension/src/DisplayName.php

@@ -0,0 +1,27 @@
+<?php
+
+namespace Oc\Museum;
+
+use Flarum\Nicknames\NicknameDriver;
+use Flarum\User\DisplayName\DriverInterface;
+use Flarum\User\User;
+
+/**
+ * Display-name driver "museum": the authors of archived comments are shown exactly as they
+ * signed on the original site, e.g. "darius (croatia)". The stock nickname driver removes
+ * brackets from names (a protection for notification e-mails, which these accounts never
+ * trigger); every other account goes through it unchanged.
+ */
+class DisplayName implements DriverInterface
+{
+    public const ARCHIVE_DOMAIN = '@archive.invalid';
+
+    public function displayName(User $user): string
+    {
+        if ($user->nickname && str_ends_with((string) $user->email, self::ARCHIVE_DOMAIN)) {
+            return $user->nickname;
+        }
+
+        return (new NicknameDriver())->displayName($user);
+    }
+}