bluesky.from_as1: populate blobs into external embed thumbs

for snarfed/bridgy-fed#1411

README.md

@@ -316,6 +316,7 @@ Standardize function and method names in all modules to `to_as1`, `from_as`, etc
     * Bug fix for quote posts with text content that's longer than Bluesky's limit ([snarfed/bridgy-fed#1197](https://github.com/snarfed/bridgy-fed/issues/1197)).
     * When a `flag` has multiple objects, use the first one that's an ATProto record.
     * Handle URLs more carefully, don't add link facets with invalid `uri`s.
+    * Populate `blobs` into external embed `thumb`s.
     * Bug fix: handle HTML links with `title` in `content` correctly.
     * Bug fix: handle attachments with no `id` or `url`.
   * `to_as1`:
@@ -326,11 +327,10 @@ Standardize function and method names in all modules to `to_as1`, `from_as`, etc
   * `Bluesky.get_activities`: skip unknown record types instead of raising `ValueError`.
 * `microformats2`:
   * `object_to_json`: Improve handling of items with multiple types by removing `inReplyTo` from likes, shares, etc ([snarfed/bridgy-fed#941](https://github.com/snarfed/bridgy-fed/issues/941)).
-  * `source`:
-    * `Source.postprocess_object`: add new `first_link_to_attachment` boolean kwarg to fetch and generate a preview `attachment` for the first link in the HTML `content`, if any.
-
 * `rss`:
   * Support image enclosures, both directions.
+* `source`:
+  * `Source.postprocess_object`: add new `first_link_to_attachment` boolean kwarg to fetch and generate a preview `attachment` for the first link in the HTML `content`, if any.
 
 ### 7.0 - 2024-06-24
 

granary/bluesky.py

@@ -745,7 +745,7 @@ def from_as1(obj, out_type=None, blobs=None, client=None,
         full_text = prefix + full_text
         index_offset = len(prefix)
 
-    # attachments, including quoted posts
+    # attachments to embed(s), including quoted posts
     record_embed = record_record_embed = external_embed = external_record_embed = None
 
     for att in attachments:
@@ -778,7 +778,7 @@ def from_as1(obj, out_type=None, blobs=None, client=None,
         full_text = QUOTE_RE_SUFFIX.sub('', full_text)
       else:
         # external link
-        external_record_embed = _to_external_embed(att)
+        external_record_embed = _to_external_embed(att, blobs=blobs)
         external_embed = {
           '$type': f'app.bsky.embed.external#view',
           'external': {
@@ -913,10 +913,10 @@ def from_as1(obj, out_type=None, blobs=None, client=None,
       if facet not in facets:
         facets.append(facet)
 
-    # attachments => embeds for articles/notes
+    # if we truncated this post's text, override external embed with link to
+    # original post. (if there are images, we added a link in the text instead,
+    # and this won't get used.)
     if truncated and url:
-      # override attachments with link to original post. (if there are images,
-      # we added a link in the text instead, and this won't get used.)
       external_record_embed = {
         '$type': f'app.bsky.embed.external',
         'external': {
@@ -1002,7 +1002,7 @@ def from_as1(obj, out_type=None, blobs=None, client=None,
       ret.update({
         'text': '',
         'facets': None,
-        'embed': _to_external_embed(obj, description=full_text),
+        'embed': _to_external_embed(obj, description=full_text, blobs=blobs),
       })
       if images_record_embed:
         ret['embed']['external']['thumb'] = images_record_embed['images'][0]['image']
@@ -1076,25 +1076,35 @@ def from_as1(obj, out_type=None, blobs=None, client=None,
   return ret
 
 
-def _to_external_embed(obj, description=None):
+def _to_external_embed(obj, description=None, blobs=None):
   """Converts an AS1 object to a Bluesky ``app.bsky.embed.external#external``.
 
   Args:
     obj (dict): AS1 object
-    content (str): if provided, overrides ``summary`` and ``content` in ``obj`
+    description (str): if provided, overrides ``summary`` and ``content` in ``obj`
 
   Returns:
     dict: Bluesky ``app.bsky.embed.external#external`` record
   """
   url = obj.get('url') or obj.get('id')
   assert url
+
+  thumb = None
+  if blobs:
+    for img in as1.get_objects(obj, 'image'):
+      if img_url := img.get('url') or img.get('id'):
+        if blob := blobs.get(img_url):
+          thumb = blob
+          break
+
   return {
     '$type': f'app.bsky.embed.external',
     'external': {
       '$type': f'app.bsky.embed.external#external',
       'uri': url,
       'title': obj.get('displayName') or '',  # required
       'description': description or obj.get('summary') or obj.get('content') or '',
+      'thumb': thumb,
     }
   }
 

granary/tests/test_bluesky.py

@@ -1735,6 +1735,32 @@ def test_from_as1_attachment_without_url_isnt_embed(self):
       }],
     }))
 
+  def test_from_as1_attachment_to_external_embed_thumb_blob(self):
+    self.assertEqual({
+      '$type': 'app.bsky.feed.post',
+      'text': '',
+      'createdAt': '2022-01-02T03:04:05.000Z',
+      'embed': {
+        '$type': 'app.bsky.embed.external',
+        'external': {
+          '$type': 'app.bsky.embed.external#external',
+          'uri': 'http://my/link',
+          'title': 'A link',
+          'description': 'foo bar',
+          'thumb': BLOB,
+        },
+      },
+    }, from_as1({
+      'objectType': 'note',
+      'attachments': [{
+        'objectType': 'link',
+        'url': 'http://my/link',
+        'displayName': 'A link',
+        'summary': 'foo bar',
+        'image': 'http://pic',
+      }],
+    }, blobs={'http://pic': BLOB}))
+
   def test_from_as1_note_display_name_as_embed(self):
     self.assert_equals({
       '$type': 'app.bsky.feed.post',