From 5e118fdbd0209903da0e2da210ba60dadd9d3a64 Mon Sep 17 00:00:00 2001
From: Ulf Rompe <ulf@rompe.org>
Date: Fri, 1 Aug 2025 17:33:22 +0200
Subject: [PATCH 1/3] Improve Mastodon profile link detection

- Use a regular expression instead of substrings
- Detect profile URL schema for Mastodon, Akkoma, Friendica and others
- Add more uids to `test.ts` to provide some examples

Fixes #10
---
 src/Mastodon.ts | 16 ++++++----------
 src/test.ts     | 19 +++++++++++++++----
 2 files changed, 21 insertions(+), 14 deletions(-)

diff --git a/src/Mastodon.ts b/src/Mastodon.ts
index 755725e..39b57e4 100644
--- a/src/Mastodon.ts
+++ b/src/Mastodon.ts
@@ -186,16 +186,12 @@ ${text.split("\n").map(txt => "  > " + txt).join("\n")}`)
         return overview.length + rest.join("\n").length + 1
     }
 
-    private static notMastodon = ["wiki.openstreetmap.org", "hdyc.neis-one.org", "matrix.to","facebook.com","github", "tasks.hotosm.org","community.openstreetmap.org","mapillary.com","wikimedia.org","wikipedia.org","wikidata.org"]
-    private static isMastodon = ["en.osm.town", ".social", "mstdn", "mastodon", "mapstodon","masto"]
+    // Fediverse profiles look like this:
+    // Mastodon, Akkoma, etc.: https://foo.bar/@user
+    // Friendica: https://foo.bar/profile/user
+    private static isProbablyMastodon = new RegExp("^https?://[^/]+/(profile/|@)[^/]+/?$")
 
     static isProbablyMastodonLink(link: string) {
-        if (this.isMastodon.some(white => link.indexOf(white) >= 0)) {
-            return true
-        }
-        if (this.notMastodon.some(black => link.indexOf(black) >= 0)) {
-            return false
-        }
-        return true; // probably?
+        return this.isProbablyMastodon.test(link)
     }
-}
\ No newline at end of file
+}
diff --git a/src/test.ts b/src/test.ts
index 7996b27..9b1731b 100644
--- a/src/test.ts
+++ b/src/test.ts
@@ -3,10 +3,21 @@ import MastodonPoster from "./Mastodon";
 
 console.log("Hello world")
 
+const uids = [
+    1214300,  // tornooc
+    4685130,  // contrapunctus
+    8404193,  // queerthoughts
+    153277,  // rompe
+]
 
-new OsmUserInfo(1214300).getMeLinks().then(links => {
-    console.log("Got links:", links);
+for (const uid of uids) {
+  new OsmUserInfo(uid).getMeLinks().then((links) => {
+    console.log("Got links for uid", uid, ":", links);
     for (const link of links) {
-        console.log(link, MastodonPoster.isProbablyMastodonLink(link) ? "Mastodon" : "NOT mastodon")
+      console.log(
+        link,
+        MastodonPoster.isProbablyMastodonLink(link) ? "Mastodon" : "NOT mastodon"
+      );
     }
-})
\ No newline at end of file
+  });
+}

From 0ed685baa0fbb568ab6a93b46611d36e78f52a2b Mon Sep 17 00:00:00 2001
From: Ulf Rompe <ulf@rompe.org>
Date: Fri, 1 Aug 2025 17:52:13 +0200
Subject: [PATCH 2/3] Add hector to test.ts as he has good test links

---
 src/test.ts | 1 +
 1 file changed, 1 insertion(+)

diff --git a/src/test.ts b/src/test.ts
index 9b1731b..0a6d1d6 100644
--- a/src/test.ts
+++ b/src/test.ts
@@ -8,6 +8,7 @@ const uids = [
     4685130,  // contrapunctus
     8404193,  // queerthoughts
     153277,  // rompe
+    393359,  // hector
 ]
 
 for (const uid of uids) {

From 057a2d2f841a9f923ee5458793f661506a9bf604 Mon Sep 17 00:00:00 2001
From: Ulf Rompe <ulf@rompe.org>
Date: Wed, 6 Aug 2025 17:59:35 +0200
Subject: [PATCH 3/3] Detect Pleroma and Pixelfed profile links

- Re-introduce the server blacklist unchanged
- Incorporate it into the Fediverse profile regex

These two platforms sadly have a non-unique URL scheme,
so it may be better to keep the blacklist in place to filter out
all the github and codeberg URLs.
---
 src/Mastodon.ts | 4 +++-
 src/test.ts     | 1 +
 2 files changed, 4 insertions(+), 1 deletion(-)

diff --git a/src/Mastodon.ts b/src/Mastodon.ts
index 39b57e4..0f907d6 100644
--- a/src/Mastodon.ts
+++ b/src/Mastodon.ts
@@ -186,10 +186,12 @@ ${text.split("\n").map(txt => "  > " + txt).join("\n")}`)
         return overview.length + rest.join("\n").length + 1
     }
 
+    private static notMastodon = ["wiki.openstreetmap.org", "hdyc.neis-one.org", "matrix.to","facebook.com","github", "tasks.hotosm.org","community.openstreetmap.org","mapillary.com","wikimedia.org","wikipedia.org","wikidata.org"]
     // Fediverse profiles look like this:
     // Mastodon, Akkoma, etc.: https://foo.bar/@user
     // Friendica: https://foo.bar/profile/user
-    private static isProbablyMastodon = new RegExp("^https?://[^/]+/(profile/|@)[^/]+/?$")
+    // Pixelfed, Pleroma: https://foo.bar/user
+    private static isProbablyMastodon = new RegExp(`^https?://(?!.*(?:${MastodonPoster.notMastodon.join('|')}))[a-zA-Z0-9.-]+/(profile/|@?)[a-zA-Z0-9_.-]+/?$`)
 
     static isProbablyMastodonLink(link: string) {
         return this.isProbablyMastodon.test(link)
diff --git a/src/test.ts b/src/test.ts
index 0a6d1d6..8e83057 100644
--- a/src/test.ts
+++ b/src/test.ts
@@ -9,6 +9,7 @@ const uids = [
     8404193,  // queerthoughts
     153277,  // rompe
     393359,  // hector
+    14124839,  // ilja has a Pleroma link
 ]
 
 for (const uid of uids) {