Merge branch 'images-from-ap' into 'master'

Extract images from ActivityPub payloads

See merge request jaywink/federation!154

federation/entities/activitypub/mappers.py

@@ -5,7 +5,8 @@ from federation.entities.activitypub.constants import NAMESPACE_PUBLIC
 from federation.entities.activitypub.entities import (
     ActivitypubFollow, ActivitypubProfile, ActivitypubAccept, ActivitypubPost, ActivitypubComment,
     ActivitypubRetraction, ActivitypubShare)
-from federation.entities.base import Follow, Profile, Accept, Post, Comment, Retraction, Share
+from federation.entities.activitypub.objects import IMAGE_TYPES
+from federation.entities.base import Follow, Profile, Accept, Post, Comment, Retraction, Share, Image
 from federation.entities.mixins import BaseEntity
 from federation.types import UserType, ReceiverVariant
 
@@ -35,6 +36,7 @@ UNDO_MAPPINGS = {
     "Announce": ActivitypubRetraction,
 }
 
+
 def element_to_objects(payload: Dict) -> List:
     """
     Transform an Element to a list of entities.
@@ -65,6 +67,9 @@ def element_to_objects(payload: Dict) -> List:
     entity._source_object = payload
     # Extract receivers
     entity._receivers = extract_receivers(payload)
+    # Extract children
+    if payload.get("object") and isinstance(payload.get("object"), dict):
+        entity._children = extract_attachments(payload.get("object"))
 
     if hasattr(entity, "post_receive"):
         entity.post_receive()
@@ -84,6 +89,25 @@ def element_to_objects(payload: Dict) -> List:
     return entities
 
 
+def extract_attachments(payload: Dict) -> List[Image]:
+    """
+    Extract images from attachments.
+
+    There could be other attachments, but currently we only extract images.
+    """
+    attachments = []
+    for item in payload.get('attachment', []):
+        # noinspection PyProtectedMember
+        if item.get("type") == "Document" and item.get("mediaType") in IMAGE_TYPES:
+            attachments.append(
+                Image(
+                    url=item.get('url'),
+                    name=item.get('name') or "",
+                )
+            )
+    return attachments
+
+
 def extract_receiver(payload: Dict, receiver: str) -> Optional[UserType]:
     """
     Transform a single receiver ID to a UserType.
@@ -239,7 +263,7 @@ def transform_attribute(key: str, value: Union[str, Dict, int], transformed: Dic
     elif key == "inReplyTo":
         transformed["target_id"] = value
     elif key == "name":
-        transformed["name"] = value
+        transformed["name"] = value or ""
     elif key == "object" and not is_object:
         if isinstance(value, dict):
             if cls == ActivitypubAccept:

federation/entities/activitypub/objects.py

@@ -2,17 +2,18 @@ import attr
 
 from federation.utils.network import fetch_content_type
 
+IMAGE_TYPES = (
+    "image/jpeg",
+    "image/png",
+    "image/gif",
+)
+
 
 @attr.s
 class ImageObject:
     """
     An Image object for AS2 serialization.
     """
-    _allowed_types = (
-        "image/jpeg",
-        "image/png",
-        "image/gif",
-    )
     url: str = attr.ib()
     type: str = attr.ib(default="Image")
     mediaType: str = attr.ib()
@@ -20,6 +21,6 @@ class ImageObject:
     @mediaType.default
     def cache_media_type(self):
         content_type = fetch_content_type(self.url)
-        if content_type in self._allowed_types:
+        if content_type in IMAGE_TYPES:
             return content_type
         return ""

federation/entities/base.py

@@ -17,21 +17,18 @@ class Accept(CreatedAtMixin, TargetIDMixin, BaseEntity):
         self._required.remove('id')
 
 
-class Image(PublicMixin, OptionalRawContentMixin, CreatedAtMixin, BaseEntity):
+class Image(OptionalRawContentMixin, CreatedAtMixin, BaseEntity):
     """Reflects a single image, possibly linked to another object."""
-    remote_path = ""
-    remote_name = ""
-    linked_type = ""
-    linked_guid = ""
+    url = ""
+    name = ""
     height = 0
     width = 0
-    url = ""
 
     _default_activity = ActivityType.CREATE
 
     def __init__(self, *args, **kwargs):
         super().__init__(*args, **kwargs)
-        self._required += ["remote_path", "remote_name"]
+        self._required += ["url"]
 
 
 class Comment(RawContentMixin, ParticipationMixin, CreatedAtMixin, RootTargetIDMixin, BaseEntity):

federation/entities/diaspora/mappers.py

@@ -2,6 +2,7 @@ import logging
 from datetime import datetime
 from typing import Callable, List
 
+# noinspection PyPackageRequirements
 from Crypto.PublicKey.RSA import RsaKey
 from lxml import etree
 
@@ -131,6 +132,7 @@ def element_to_objects(
     entity._mentions = entity.extract_mentions()
     # Do child elements
     for child in element:
+        # noinspection PyProtectedMember
         entity._children.extend(element_to_objects(child, sender, user=user))
     # Add to entities list
     entities.append(entity)
@@ -219,12 +221,7 @@ def transform_attributes(attrs, cls):
         elif key in ["target_type"] and cls == DiasporaRetraction:
             transformed["entity_type"] = DiasporaRetraction.entity_type_from_remote(value)
         elif key == "remote_photo_path":
-            transformed["remote_path"] = value
-        elif key == "remote_photo_name":
-            transformed["remote_name"] = value
-        elif key == "status_message_guid":
-            transformed["linked_guid"] = value
-            transformed["linked_type"] = "Post"
+            transformed["url"] = f"{value}{attrs.get('remote_photo_name')}"
         elif key == "author_signature":
             transformed["signature"] = value
         elif key in BOOLEAN_KEYS:

federation/tests/entities/activitypub/test_mappers.py

@@ -6,10 +6,11 @@ from federation.entities.activitypub.entities import (
     ActivitypubFollow, ActivitypubAccept, ActivitypubProfile, ActivitypubPost, ActivitypubComment,
     ActivitypubRetraction, ActivitypubShare)
 from federation.entities.activitypub.mappers import message_to_objects, get_outbound_entity
-from federation.entities.base import Accept, Follow, Profile, Post, Comment
+from federation.entities.base import Accept, Follow, Profile, Post, Comment, Image
 from federation.tests.fixtures.payloads import (
     ACTIVITYPUB_FOLLOW, ACTIVITYPUB_PROFILE, ACTIVITYPUB_PROFILE_INVALID, ACTIVITYPUB_UNDO_FOLLOW, ACTIVITYPUB_POST,
-    ACTIVITYPUB_COMMENT, ACTIVITYPUB_RETRACTION, ACTIVITYPUB_SHARE, ACTIVITYPUB_RETRACTION_SHARE)
+    ACTIVITYPUB_COMMENT, ACTIVITYPUB_RETRACTION, ACTIVITYPUB_SHARE, ACTIVITYPUB_RETRACTION_SHARE,
+    ACTIVITYPUB_POST_IMAGES)
 from federation.types import UserType, ReceiverVariant
 
 
@@ -71,25 +72,22 @@ class TestActivitypubEntityMappersReceive:
         assert post.public is True
         assert getattr(post, "target_id", None) is None
 
-    @pytest.mark.skip
     def test_message_to_objects_post_with_photos(self):
-        entities = message_to_objects(DIASPORA_POST_WITH_PHOTOS, "alice@alice.diaspora.example.org")
+        entities = message_to_objects(ACTIVITYPUB_POST_IMAGES, "https://mastodon.social/users/jaywink")
         assert len(entities) == 1
         post = entities[0]
-        assert isinstance(post, DiasporaPost)
+        assert isinstance(post, ActivitypubPost)
+        assert len(post._children) == 1
         photo = post._children[0]
-        assert isinstance(photo, DiasporaImage)
-        assert photo.remote_path == "https://alice.diaspora.example.org/uploads/images/"
-        assert photo.remote_name == "1234.jpg"
+        assert isinstance(photo, Image)
+        assert photo.url == "https://files.mastodon.social/media_attachments/files/017/642/079/original/" \
+                            "f51b0aee0ee1f2e1.jpg"
+        assert photo.name == ""
         assert photo.raw_content == ""
-        assert photo.linked_type == "Post"
-        assert photo.linked_guid == "((guidguidguidguidguidguidguid))"
-        assert photo.height == 120
-        assert photo.width == 120
-        assert photo.guid == "((guidguidguidguidguidguidguif))"
-        assert photo.handle == "alice@alice.diaspora.example.org"
-        assert photo.public == False
-        assert photo.created_at == datetime(2011, 7, 20, 1, 36, 7)
+        assert photo.height == 0
+        assert photo.width == 0
+        assert photo.guid == ""
+        assert photo.handle == ""
 
     def test_message_to_objects_comment(self):
         entities = message_to_objects(ACTIVITYPUB_COMMENT, "https://diaspodon.fr/users/jaywink")

federation/tests/entities/diaspora/test_mappers.py

@@ -51,16 +51,13 @@ class TestDiasporaEntityMappersReceive:
         assert isinstance(post, DiasporaPost)
         photo = post._children[0]
         assert isinstance(photo, DiasporaImage)
-        assert photo.remote_path == "https://alice.diaspora.example.org/uploads/images/"
-        assert photo.remote_name == "1234.jpg"
+        assert photo.url == "https://alice.diaspora.example.org/uploads/images/1234.jpg"
+        assert photo.name == ""
         assert photo.raw_content == ""
-        assert photo.linked_type == "Post"
-        assert photo.linked_guid == "((guidguidguidguidguidguidguid))"
         assert photo.height == 120
         assert photo.width == 120
         assert photo.guid == "((guidguidguidguidguidguidguif))"
         assert photo.handle == "alice@alice.diaspora.example.org"
-        assert photo.public == False
         assert photo.created_at == datetime(2011, 7, 20, 1, 36, 7)
 
     @patch("federation.entities.diaspora.mappers.DiasporaComment._validate_signatures")

federation/tests/factories/entities.py

@@ -59,12 +59,12 @@ class DiasporaPostFactory(PostFactory):
         model = DiasporaPost
 
 
-class ImageFactory(ActorIDMixinFactory, IDMixinFactory, PublicMixinFactory, factory.Factory):
+class ImageFactory(ActorIDMixinFactory, IDMixinFactory, factory.Factory):
     class Meta:
         model = Image
 
-    remote_path = factory.Faker('uri')
-    remote_name = factory.Faker('file_path', extension='jpg')
+    url = factory.Faker('uri')
+    name = factory.Faker('slug')
 
 
 class ProfileFactory(IDMixinFactory, RawContentMixinFactory, factory.Factory):

federation/tests/fixtures/payloads/activitypub.py

@@ -319,3 +319,56 @@ ACTIVITYPUB_POST_OBJECT = {
         'partOf': 'https://diaspodon.fr/users/jaywink/statuses/102356911717767237/replies',
         'items': []}},
 }
+
+ACTIVITYPUB_POST_IMAGES = {'@context': ['https://www.w3.org/ns/activitystreams',
+  {'ostatus': 'http://ostatus.org#',
+   'atomUri': 'ostatus:atomUri',
+   'inReplyToAtomUri': 'ostatus:inReplyToAtomUri',
+   'conversation': 'ostatus:conversation',
+   'sensitive': 'as:sensitive',
+   'Hashtag': 'as:Hashtag',
+   'toot': 'http://joinmastodon.org/ns#',
+   'Emoji': 'toot:Emoji',
+   'focalPoint': {'@container': '@list', '@id': 'toot:focalPoint'},
+   'blurhash': 'toot:blurhash'}],
+ 'id': 'https://mastodon.social/users/jaywink/statuses/102611770245850345/activity',
+ 'type': 'Create',
+ 'actor': 'https://mastodon.social/users/jaywink',
+ 'published': '2019-08-13T21:22:37Z',
+ 'to': ['https://www.w3.org/ns/activitystreams#Public'],
+ 'cc': ['https://mastodon.social/users/jaywink/followers'],
+ 'object': {'id': 'https://mastodon.social/users/jaywink/statuses/102611770245850345',
+  'type': 'Note',
+  'summary': None,
+  'inReplyTo': None,
+  'published': '2019-08-13T21:22:37Z',
+  'url': 'https://mastodon.social/@jaywink/102611770245850345',
+  'attributedTo': 'https://mastodon.social/users/jaywink',
+  'to': ['https://www.w3.org/ns/activitystreams#Public'],
+  'cc': ['https://mastodon.social/users/jaywink/followers'],
+  'sensitive': False,
+  'atomUri': 'https://mastodon.social/users/jaywink/statuses/102611770245850345',
+  'inReplyToAtomUri': None,
+  'conversation': 'tag:mastodon.social,2019-08-13:objectId=119290371:objectType=Conversation',
+  'content': '<p>image test</p>',
+  'contentMap': {'en': '<p>image test</p>'},
+  'attachment': [{'type': 'Document',
+    'mediaType': 'image/jpeg',
+    'url': 'https://files.mastodon.social/media_attachments/files/017/642/079/original/f51b0aee0ee1f2e1.jpg',
+    'name': None,
+    'blurhash': 'UaH1x+IpD*RktToft6s:0f%2tQj@xsWWRkNG'},
+   {'type': 'Document',
+    'mediaType': 'video/mp4',
+    'url': 'https://files.mastodon.social/media_attachments/files/017/642/084/original/e18dda257e5e7078.mp4',
+    'name': None,
+    'blurhash': 'UH9jv0ay00Rj%MM{IU%M%MWBRjofxuayM{t7'}],
+  'tag': [],
+  'replies': {'id': 'https://mastodon.social/users/jaywink/statuses/102611770245850345/replies',
+   'type': 'Collection',
+   'first': {'type': 'CollectionPage',
+    'partOf': 'https://mastodon.social/users/jaywink/statuses/102611770245850345/replies',
+    'items': []}}},
+ 'signature': {'type': 'RsaSignature2017',
+  'creator': 'https://mastodon.social/users/jaywink#main-key',
+  'created': '2019-08-13T21:22:37Z',
+  'signatureValue': 'Ia61wdHHIy9gCY5YwqlPtd80eJ2liT9Yi3yHdRdP+fQ5/9np3wHJKNPa7gdzP/BiRzh6aOa2dHWJjB8mOnHYrYBn6Fl3RlCniqousVTDue/ek0zvcFWmlhfja02meDiva+t61O/6Ul1l4tQObMorSf7GbEPePlQiozr/SR/5HIj3SDP0Y8JmlTvhSFgiH6obdroaIYEMQAoYZVcYofGeQUEhotDRp0OGQ4UaPBli4WyzVOUqHMW6pw90QQzZF9XpimwAemk9oAgPmGEPkugFeHfrWt1l84KLdwqwWD8FRIep7gCtu6MpCA8TX4JC5yJvyQ9GbZLZfJSQ6t5wSrcafw=='}}