Process mention on annotation creation #9322

Author: mtomilov

docs/_extra/api-reference/schemas/annotation.yaml

@@ -199,3 +199,24 @@ Annotation:
               description: The annotation creator's display name
               example: "Felicity Nunsun"
             - type: null
+    mentions:
+      type: array
+      items:
+        type: object
+        properties:
+          userid:
+            type: string
+            pattern: "acct:^[A-Za-z0-9._]{3,30}@.*$"
+            description: user account ID in the format `"acct:<username>@<authority>"`
+            example: "acct:[email protected]"
+          username:
+            type: string
+            description: The username of the user at the time of the mention
+          display_name:
+            type: string
+            description: The display name of the user
+          link:
+            type: string
+            format: uri
+            description: The link to the user profile
+      description: An array of user mentions the annotation text

h/models/mention.py

@@ -6,7 +6,7 @@
 from h.models import helpers
 
 
-class Mention(Base, Timestamps):  # pragma: nocover
+class Mention(Base, Timestamps):
     __tablename__ = "mention"
 
     id: Mapped[int] = mapped_column(sa.Integer, autoincrement=True, primary_key=True)
@@ -17,7 +17,9 @@ class Mention(Base, Timestamps):  # pragma: nocover
         nullable=False,
     )
     """FK to annotation.id"""
-    annotation = sa.orm.relationship("Annotation", back_populates="mentions")
+    annotation = sa.orm.relationship(
+        "Annotation", back_populates="mentions", uselist=False
+    )
 
     user_id: Mapped[int] = mapped_column(
         sa.Integer,

h/presenters/mention_json.py

@@ -0,0 +1,18 @@
+from typing import Any
+
+from h.models import Mention
+
+
+class MentionJSONPresenter:
+    """Present a mention in the JSON format returned by API requests."""
+
+    def __init__(self, mention: Mention):
+        self._mention = mention
+
+    def asdict(self) -> dict[str, Any]:
+        return {
+            "userid": self._mention.user.userid,
+            "username": self._mention.username,
+            "display_name": self._mention.user.display_name,
+            "link": self._mention.user.uri,
+        }

h/services/__init__.py

@@ -11,6 +11,7 @@
     BulkLMSStatsService,
 )
 from h.services.job_queue import JobQueueService
+from h.services.mention import MentionService
 from h.services.subscription import SubscriptionService
 
 
@@ -42,6 +43,7 @@ def includeme(config):  # pragma: no cover
     config.register_service_factory(
         "h.services.annotation_write.service_factory", iface=AnnotationWriteService
     )
+    config.register_service_factory("h.services.mention.factory", iface=MentionService)
 
     # Other services
     config.register_service_factory(

h/services/annotation_json.py

@@ -2,9 +2,12 @@
 
 from h.models import Annotation, User
 from h.presenters import DocumentJSONPresenter
+from h.presenters.mention_json import MentionJSONPresenter
 from h.security import Identity, identity_permits
 from h.security.permissions import Permission
+from h.services import MentionService
 from h.services.annotation_read import AnnotationReadService
+from h.services.feature import FeatureService
 from h.services.flag import FlagService
 from h.services.links import LinksService
 from h.services.user import UserService
@@ -22,6 +25,8 @@ def __init__(
         links_service: LinksService,
         flag_service: FlagService,
         user_service: UserService,
+        mention_service: MentionService,
+        feature_service: FeatureService,
     ):
         """
         Instantiate the service.
@@ -30,11 +35,14 @@ def __init__(
         :param links_service: LinksService instance
         :param flag_service: FlagService instance
         :param user_service: UserService instance
+        :param mention_service: MentionService instance
         """
         self._annotation_read_service = annotation_read_service
         self._links_service = links_service
         self._flag_service = flag_service
         self._user_service = user_service
+        self._mention_service = mention_service
+        self._feature_service = feature_service
 
     def present(self, annotation: Annotation):
         """
@@ -73,6 +81,11 @@ def present(self, annotation: Annotation):
                 "links": self._links_service.get_all(annotation),
             }
         )
+        if self._feature_service.enabled("at_mentions"):  # pragma: no cover
+            model["mentions"] = [
+                MentionJSONPresenter(mention).asdict()
+                for mention in annotation.mentions
+            ]
 
         model.update(user_info(self._user_service.fetch(annotation.userid)))
 
@@ -151,6 +164,8 @@ def present_all_for_user(self, annotation_ids, user: User):
                 # which ultimately depends on group permissions, causing a
                 # group lookup for every annotation without this
                 Annotation.group,
+                # Optimise access to the mentions
+                Annotation.mentions,
             ],
         )
 
@@ -184,4 +199,6 @@ def factory(_context, request):
         links_service=request.find_service(name="links"),
         flag_service=request.find_service(name="flag"),
         user_service=request.find_service(name="user"),
+        mention_service=request.find_service(MentionService),
+        feature_service=request.find_service(name="feature"),
     )

h/services/annotation_write.py

@@ -13,6 +13,7 @@
 from h.services.annotation_metadata import AnnotationMetadataService
 from h.services.annotation_read import AnnotationReadService
 from h.services.job_queue import JobQueueService
+from h.services.mention import MentionService
 from h.traversal.group import GroupContext
 from h.util.group_scope import url_in_scope
 
@@ -29,12 +30,16 @@ def __init__(
         queue_service: JobQueueService,
         annotation_read_service: AnnotationReadService,
         annotation_metadata_service: AnnotationMetadataService,
+        mention_service: MentionService,
+        feature_service,
     ):
         self._db = db_session
         self._has_permission = has_permission
         self._queue_service = queue_service
         self._annotation_read_service = annotation_read_service
         self._annotation_metadata_service = annotation_metadata_service
+        self._mention_service = mention_service
+        self._feature_service = feature_service
 
     def create_annotation(self, data: dict) -> Annotation:
         """
@@ -88,6 +93,9 @@ def create_annotation(self, data: dict) -> Annotation:
             schedule_in=60,
         )
 
+        if self._feature_service.enabled("at_mentions"):  # pragma: no cover
+            self._mention_service.update_mentions(annotation)
+
         return annotation
 
     def update_annotation(
@@ -281,4 +289,6 @@ def service_factory(_context, request) -> AnnotationWriteService:
         queue_service=request.find_service(name="queue_service"),
         annotation_read_service=request.find_service(AnnotationReadService),
         annotation_metadata_service=request.find_service(AnnotationMetadataService),
+        mention_service=request.find_service(MentionService),
+        feature_service=request.find_service(name="feature"),
     )

h/services/html.py

@@ -0,0 +1,20 @@
+from html.parser import HTMLParser
+
+
+class LinkParser(HTMLParser):
+    def __init__(self):
+        super().__init__()
+        self._links = []
+
+    def handle_starttag(self, tag, attrs):
+        if tag == "a":
+            self._links.append(dict(attrs))
+
+    def get_links(self) -> list[dict]:
+        return self._links
+
+
+def parse_html_links(html: str) -> list[dict]:
+    parser = LinkParser()
+    parser.feed(html)
+    return parser.get_links()

h/services/mention.py

@@ -0,0 +1,79 @@
+import logging
+from collections import OrderedDict
+
+from sqlalchemy import delete
+from sqlalchemy.orm import Session
+
+from h.models import Annotation, Mention
+from h.services.html import parse_html_links
+from h.services.user import UserService
+
+MENTION_ATTRIBUTE = "data-hyp-mention"
+MENTION_USERID = "data-userid"
+MENTION_LIMIT = 5
+
+logger = logging.getLogger(__name__)
+
+
+class MentionService:
+    """A service for managing user mentions."""
+
+    def __init__(self, session: Session, user_service: UserService):
+        self._session = session
+        self._user_service = user_service
+
+    def update_mentions(self, annotation: Annotation) -> None:
+        self._session.flush()
+
+        # Only shared annotations can have mentions
+        if not annotation.shared:
+            return
+        mentioning_user = self._user_service.fetch(annotation.userid)
+        # NIPSA users do not send mentions
+        if mentioning_user.nipsa:
+            return
+
+        mentioned_userids = OrderedDict.fromkeys(self._parse_userids(annotation.text))
+        mentioned_users = self._user_service.fetch_all(mentioned_userids)
+        self._session.execute(
+            delete(Mention).where(Mention.annotation_id == annotation.id)
+        )
+
+        for i, user in enumerate(mentioned_users):
+            if i >= MENTION_LIMIT:
+                logger.warning(
+                    "Annotation %s has more than %s mentions",
+                    annotation.id,
+                    MENTION_LIMIT,
+                )
+                break
+            # NIPSA users do not receive mentions
+            if user.nipsa:
+                continue
+            # Only allow mentions if the annotation is in the public group
+            # or the annotation is in one of mentioned user's groups
+            if not (
+                annotation.groupid == "__world__" or annotation.group in user.groups
+            ):
+                continue
+
+            mention = Mention(
+                annotation_id=annotation.id, user_id=user.id, username=user.username
+            )
+            self._session.add(mention)
+
+    @staticmethod
+    def _parse_userids(text: str) -> list[str]:
+        links = parse_html_links(text)
+        return [
+            user_id
+            for link in links
+            if MENTION_ATTRIBUTE in link and (user_id := link.get(MENTION_USERID))
+        ]
+
+
+def factory(_context, request) -> MentionService:
+    """Return a MentionService instance for the passed context and request."""
+    return MentionService(
+        session=request.db, user_service=request.find_service(name="user")
+    )

tests/common/factories/__init__.py

@@ -16,6 +16,7 @@
 from tests.common.factories.group import Group, OpenGroup, RestrictedGroup
 from tests.common.factories.group_scope import GroupScope
 from tests.common.factories.job import ExpungeUserJob, Job, SyncAnnotationJob
+from tests.common.factories.mention import Mention
 from tests.common.factories.organization import Organization
 from tests.common.factories.setting import Setting
 from tests.common.factories.subscriptions import Subscriptions

tests/common/factories/mention.py

@@ -0,0 +1,17 @@
+import factory
+
+from h import models
+
+from .annotation import Annotation
+from .base import ModelFactory
+from .user import User
+
+
+class Mention(ModelFactory):
+    class Meta:
+        model = models.Mention
+        sqlalchemy_session_persistence = "flush"
+
+    annotation = factory.SubFactory(Annotation)
+    user = factory.SubFactory(User)
+    username = factory.LazyAttribute(lambda obj: obj.user.username)

tests/common/fixtures/services.py

@@ -2,6 +2,7 @@
 
 import pytest
 
+from h.services import MentionService
 from h.services.analytics import AnalyticsService
 from h.services.annotation_delete import AnnotationDeleteService
 from h.services.annotation_json import AnnotationJSONService
@@ -19,6 +20,7 @@
     BulkLMSStatsService,
 )
 from h.services.developer_token import DeveloperTokenService
+from h.services.feature import FeatureService
 from h.services.flag import FlagService
 from h.services.group import GroupService
 from h.services.group_create import GroupCreateService
@@ -84,6 +86,8 @@
     "user_signup_service",
     "user_unique_service",
     "user_update_service",
+    "mention_service",
+    "feature_service",
 )
 
 
@@ -306,3 +310,13 @@ def user_unique_service(mock_service):
 @pytest.fixture
 def user_update_service(mock_service):
     return mock_service(UserUpdateService, name="user_update")
+
+
+@pytest.fixture
+def mention_service(mock_service):
+    return mock_service(MentionService)
+
+
+@pytest.fixture
+def feature_service(mock_service):
+    return mock_service(FeatureService, name="feature")

tests/unit/h/models/mention_test.py

@@ -0,0 +1,8 @@
+def test_repr(factories):
+    annotation = factories.Annotation()
+    mention = factories.Mention(annotation=annotation)
+
+    assert (
+        repr(mention)
+        == f"Mention(id={mention.id}, annotation_id={mention.annotation.id!r}, user_id={mention.user.id})"
+    )

tests/unit/h/presenters/mention_json_test.py

@@ -0,0 +1,26 @@
+import pytest
+
+from h.models import Mention
+from h.presenters.mention_json import MentionJSONPresenter
+
+
+class TestMentionJSONPresenter:
+    def test_as_dict(self, user, annotation):
+        mention = Mention(annotation=annotation, user=user, username=user.username)
+
+        data = MentionJSONPresenter(mention).asdict()
+
+        assert data == {
+            "userid": user.userid,
+            "username": user.username,
+            "display_name": user.display_name,
+            "link": user.uri,
+        }
+
+    @pytest.fixture
+    def user(self, factories):
+        return factories.User.build()
+
+    @pytest.fixture
+    def annotation(self, factories):
+        return factories.Annotation.build()