bluesky.from_as1: mention facet bug fix for mentions at beginning of …

…content

fixes snarfed/bridgy-fed#957 (again)

granary/bluesky.py

@@ -644,16 +644,18 @@ def from_as1(obj, out_type=None, blobs=None, client=None):
         # can't use \b for word boundaries here because that only includes
         # alphanumerics, and Bluesky hashtags can include emoji
         prefix = '#' if type == 'hashtag' else '@' if type == 'mention' else ''
-        match = re.search(fr'[\s^]({prefix}{name})[\s$]', text)
+        # can't use \b at beginning/end because # and @ and emoji aren't
+        # word-constituent chars
+        match = re.search(fr'(^|\s)({prefix}{name})($|\s)', text)
         if not match and type == 'mention' and '@' in name:
           # try without @[server] suffix
           username = name.split('@')[0]
-          match = re.search(fr'[\s^]({prefix}{username})[\s$]', text)
+          match = re.search(fr'(^|\s)({prefix}{username})($|\s)', text)
 
         if match:
           facet['index'] = {
-            'byteStart': len(full_text[:match.start(1)].encode()),
-            'byteEnd': len(full_text[:match.end(1)].encode()),
+            'byteStart': len(full_text[:match.start(2)].encode()),
+            'byteEnd': len(full_text[:match.end(2)].encode()),
           }
 
       # skip or trim this facet if it's off the end of content that got truncated
@@ -926,6 +928,7 @@ def to_as1(obj, type=None, uri=None, repo_did=None, repo_handle=None,
     ret.update({
       'url': util.dedupe_urls(urls),
       'displayName': obj.get('displayName'),
+      # TODO: for app.bsky.feed.generator, use descriptionFacets
       'summary': util.linkify(obj.get('description') or '', pretty=True),
       'image': images,
       'published': obj.get('createdAt'),

granary/tests/test_bluesky.py

@@ -863,8 +863,36 @@ def test_from_as1_tag_mention_at_char_html_content_guess_index(self):
     note = copy.deepcopy(NOTE_AS_TAG_MENTION_URL)
     note['content'] = '<p>foo <a href="https://bsky.app/...">@you.com</a> bar</p>'
     note['tags'][0]['displayName'] = '@you.com'
+    del note['tags'][0]['startIndex']
+    del note['tags'][0]['length']
     self.assert_equals(POST_BSKY_FACET_MENTION, from_as1(note))
 
+  def test_from_as1_tag_mention_at_beginning(self):
+    self.assert_equals({
+      '$type': 'app.bsky.feed.post',
+      'createdAt': '2022-01-02T03:04:05.000Z',
+      'text': '@shreyanjain.net hello there',
+      'facets': [{
+        '$type': 'app.bsky.richtext.facet',
+        'features': [{
+          '$type': 'app.bsky.richtext.facet#mention',
+          'did': 'did:plc:foo',
+        }],
+        'index': {
+          'byteStart': 0,
+          'byteEnd': 16,
+        },
+      }],
+    }, from_as1({
+      'objectType' : 'note',
+      'content' : '<p><span class=\'h-card\' translate=\'no\'><a href=\'https://bsky.brid.gy/r/https://bsky.app/profile/shreyanjain.net\' class=\'u-url mention\'>@<span>shreyanjain.net</span></a></span> hello there</p>',
+      'tags' : [{
+        'objectType': 'mention',
+        'displayName': '@[email protected]',
+        'url': 'did:plc:foo',
+      }],
+    }))
+
   def test_from_as1_tag_mention_not_bluesky(self):
     self.assert_equals({
       '$type': 'app.bsky.feed.post',