Move webmention parsing and authorship into the app

The jonnybarnes/webmentions-parser package is now archived, and its
guzzlehttp/guzzle ^6|^7 constraint was the only thing holding Guzzle
back from 8. Port the parts we use into App\Services\WebMention and
drop the package.

Parser keeps getMentionType and the checkInReplyTo/LikeOf/RepostOf
checks unchanged. Authorship follows the same authorship algorithm,
with some fixes:

- fetch the author page with Laravel's Http client, so tests can fake
  it, and treat connection failures as AuthorshipParserException too
- keep state in locals so one instance can be reused safely
- make the url == uid == author-page check actually compare the
  h-card's uid, rather than an undefined variable

Add direct tests for both, and mark the library as archived on the
projects page.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

app/Exceptions/AuthorshipParserException.php

@@ -0,0 +1,9 @@
+<?php
+
+declare(strict_types=1);
+
+namespace App\Exceptions;
+
+use Exception;
+
+class AuthorshipParserException extends Exception {}

app/Exceptions/InvalidMentionException.php

@@ -0,0 +1,9 @@
+<?php
+
+declare(strict_types=1);
+
+namespace App\Exceptions;
+
+use Exception;
+
+class InvalidMentionException extends Exception {}

app/Jobs/ProcessLike.php

@@ -4,7 +4,9 @@ declare(strict_types=1);
 
 namespace App\Jobs;
 
+use App\Exceptions\AuthorshipParserException;
 use App\Models\Like;
+use App\Services\WebMention\Authorship;
 use Illuminate\Bus\Queueable;
 use Illuminate\Contracts\Queue\ShouldQueue;
 use Illuminate\Foundation\Bus\Dispatchable;
@@ -12,8 +14,6 @@ use Illuminate\Queue\InteractsWithQueue;
 use Illuminate\Queue\SerializesModels;
 use Illuminate\Support\Arr;
 use Illuminate\Support\Facades\Http;
-use Jonnybarnes\WebmentionsParser\Authorship;
-use Jonnybarnes\WebmentionsParser\Exceptions\AuthorshipParserException;
 
 class ProcessLike implements ShouldQueue
 {

app/Jobs/ProcessWebMention.php

@@ -4,17 +4,17 @@ declare(strict_types=1);
 
 namespace App\Jobs;
 
+use App\Exceptions\InvalidMentionException;
 use App\Exceptions\RemoteContentNotFoundException;
 use App\Models\Note;
 use App\Models\WebMention;
+use App\Services\WebMention\Parser;
 use Illuminate\Bus\Queueable;
 use Illuminate\Contracts\Queue\ShouldQueue;
 use Illuminate\Http\Client\ConnectionException;
 use Illuminate\Queue\InteractsWithQueue;
 use Illuminate\Queue\SerializesModels;
 use Illuminate\Support\Facades\Http;
-use Jonnybarnes\WebmentionsParser\Exceptions\InvalidMentionException;
-use Jonnybarnes\WebmentionsParser\Parser;
 use Mf2;
 
 class ProcessWebMention implements ShouldQueue

app/Jobs/SaveProfileImage.php

@@ -4,6 +4,8 @@ declare(strict_types=1);
 
 namespace App\Jobs;
 
+use App\Exceptions\AuthorshipParserException;
+use App\Services\WebMention\Authorship;
 use Illuminate\Bus\Queueable;
 use Illuminate\Contracts\Queue\ShouldQueue;
 use Illuminate\Http\Client\ConnectionException;
@@ -12,8 +14,6 @@ use Illuminate\Queue\InteractsWithQueue;
 use Illuminate\Queue\SerializesModels;
 use Illuminate\Support\Arr;
 use Illuminate\Support\Facades\Http;
-use Jonnybarnes\WebmentionsParser\Authorship;
-use Jonnybarnes\WebmentionsParser\Exceptions\AuthorshipParserException;
 
 class SaveProfileImage implements ShouldQueue
 {

app/Models/WebMention.php

@@ -4,6 +4,7 @@ declare(strict_types=1);
 
 namespace App\Models;
 
+use App\Services\WebMention\Authorship;
 use App\Traits\FilterHtml;
 use Exception;
 use Illuminate\Database\Eloquent\Attributes\Guarded;
@@ -13,7 +14,6 @@ use Illuminate\Database\Eloquent\Factories\HasFactory;
 use Illuminate\Database\Eloquent\Model;
 use Illuminate\Database\Eloquent\Relations\MorphTo;
 use Illuminate\Filesystem\Filesystem;
-use Jonnybarnes\WebmentionsParser\Authorship;
 
 #[Table('webmentions')]
 #[Guarded(['id'])]

app/Services/WebMention/Authorship.php

@@ -0,0 +1,152 @@
+<?php
+
+declare(strict_types=1);
+
+namespace App\Services\WebMention;
+
+use App\Exceptions\AuthorshipParserException;
+use Illuminate\Http\Client\ConnectionException;
+use Illuminate\Support\Facades\Http;
+
+use function Mf2\parse;
+
+/**
+ * Determine the author of a post from its parsed microformats, following
+ * the IndieWeb authorship algorithm: https://indieweb.org/authorship
+ *
+ * Originally from the jonnybarnes/webmentions-parser package.
+ */
+class Authorship
+{
+    /**
+     * Find the author of the first h-entry (or h-feed) in the microformats.
+     *
+     * Returns the author's h-card, or their name as a string if that's all
+     * we have, or false if the author can't be determined.
+     *
+     * @throws AuthorshipParserException
+     */
+    public function findAuthor(array $mf, bool $permalink = true): array|string|false
+    {
+        $hEntry = null;
+        $hFeed = null;
+        foreach ($mf['items'] ?? [] as $item) {
+            foreach ($item['type'] ?? [] as $type) {
+                if ($type === 'h-entry') {
+                    $hEntry = $item;
+                } elseif ($type === 'h-feed') {
+                    $hFeed = $item;
+                }
+            }
+        }
+
+        if ($hEntry === null && $hFeed === null) {
+            throw new AuthorshipParserException('No h-entry or h-feed found');
+        }
+
+        // If the h-entry has an author property use that
+        $author = $hEntry['properties']['author'] ?? null;
+
+        // A parent h-feed with an h-card takes precedence
+        foreach ($hFeed['children'] ?? [] as $child) {
+            if (($child['type'][0] ?? null) === 'h-card') {
+                $author = $child;
+            }
+        }
+
+        $authorPage = null;
+
+        if ($author !== null) {
+            // Flatten a single string entry
+            if (is_array($author) && count($author) === 1 && is_string($author[0] ?? null)) {
+                $author = $author[0];
+            }
+
+            // A URL author is the author-page, anything else is the author itself
+            if (is_string($author) && filter_var($author, FILTER_VALIDATE_URL)) {
+                $authorPage = $author;
+            } else {
+                return $this->normalise($author);
+            }
+        }
+
+        // If no author-page and h-entry is a permalink then use the rel-author link
+        if ($authorPage === null && $permalink && isset($mf['rels']['author'])) {
+            $authorPage = is_array($mf['rels']['author']) ? $mf['rels']['author'][0] : $mf['rels']['author'];
+        }
+
+        if ($authorPage === null) {
+            return false;
+        }
+
+        return $this->findAuthorOnPage($authorPage);
+    }
+
+    /**
+     * Fetch the author-page and look for a representative h-card on it.
+     *
+     * @throws AuthorshipParserException
+     */
+    private function findAuthorOnPage(string $authorPage): array|false
+    {
+        try {
+            $response = Http::get($authorPage);
+        } catch (ConnectionException) {
+            throw new AuthorshipParserException('Unable to connect to the author’s page');
+        }
+
+        if ($response->failed()) {
+            throw new AuthorshipParserException('Unable to get the content from the author’s page');
+        }
+
+        $authorMf2 = parse($response->body(), $authorPage);
+        $hCards = array_filter(
+            $authorMf2['items'] ?? [],
+            fn (array $item): bool => in_array('h-card', $item['type'] ?? [], true)
+        );
+
+        // Use the first h-card where url == uid == author-page
+        foreach ($hCards as $hCard) {
+            if (
+                in_array($authorPage, $hCard['properties']['url'] ?? [], true)
+                && in_array($authorPage, $hCard['properties']['uid'] ?? [], true)
+            ) {
+                return $this->normalise($hCard);
+            }
+        }
+
+        // Else the first h-card with a url that matches a rel-me link on the page
+        $relMeLinks = $authorMf2['rels']['me'] ?? [];
+        foreach ($hCards as $hCard) {
+            foreach ($hCard['properties']['url'] ?? [] as $url) {
+                if (in_array($url, $relMeLinks, true)) {
+                    return $this->normalise($hCard);
+                }
+            }
+        }
+
+        // Else the first h-card with url == author-page
+        foreach ($hCards as $hCard) {
+            if (in_array($authorPage, $hCard['properties']['url'] ?? [], true)) {
+                return $this->normalise($hCard);
+            }
+        }
+
+        return false;
+    }
+
+    private function normalise(array|string $author): array|string
+    {
+        if (is_array($author) && array_key_exists(0, $author)) {
+            $author = $author[0];
+        }
+
+        if (is_string($author)) {
+            return $author;
+        }
+
+        unset($author['value'], $author['children']);
+
+        return $author;
+    }
+}

app/Services/WebMention/Parser.php

@@ -0,0 +1,86 @@
+<?php
+
+declare(strict_types=1);
+
+namespace App\Services\WebMention;
+
+use App\Exceptions\InvalidMentionException;
+
+/**
+ * Inspect the parsed microformats of a webmention source.
+ *
+ * Originally from the jonnybarnes/webmentions-parser package.
+ */
+class Parser
+{
+    /**
+     * Return the type of mention.
+     *
+     * @throws InvalidMentionException when the mention type can't be determined
+     */
+    public function getMentionType(array $microformats): string
+    {
+        foreach (['in-reply-to', 'like-of', 'repost-of'] as $type) {
+            if ($this->arrayKeyExistsRecursive($type, $microformats)) {
+                return $type;
+            }
+        }
+
+        throw new InvalidMentionException;
+    }
+
+    /**
+     * Check a reply is to the intended target.
+     */
+    public function checkInReplyTo(array $microformats, string $target): bool
+    {
+        foreach ($microformats['items'] ?? [] as $item) {
+            $inReplyTo = $item['properties']['in-reply-to'] ?? null;
+            if ($inReplyTo === null) {
+                continue;
+            }
+
+            // The reply context may be an h-cite rather than a plain URL
+            if (is_array($inReplyTo[0]) && ($inReplyTo[0]['properties']['url'][0] ?? null) === $target) {
+                return true;
+            }
+
+            if (in_array($target, $inReplyTo, true)) {
+                return true;
+            }
+        }
+
+        return false;
+    }
+
+    /**
+     * Check the microformats contain a like of the target.
+     */
+    public function checkLikeOf(array $microformats, string $target): bool
+    {
+        return in_array($target, $microformats['items'][0]['properties']['like-of'] ?? [], true);
+    }
+
+    /**
+     * Check the microformats contain a repost of the target.
+     */
+    public function checkRepostOf(array $microformats, string $target): bool
+    {
+        return in_array($target, $microformats['items'][0]['properties']['repost-of'] ?? [], true);
+    }
+
+    private function arrayKeyExistsRecursive(string $needle, array $haystack): bool
+    {
+        if (array_key_exists($needle, $haystack)) {
+            return true;
+        }
+
+        foreach ($haystack as $value) {
+            if (is_array($value) && $this->arrayKeyExistsRecursive($needle, $value)) {
+                return true;
+            }
+        }
+
+        return false;
+    }
+}