feat: slack parent update in indexing

backend/alembic/versions/ae6aaba063eb_perm_sync_utils.py

@@ -0,0 +1,60 @@
+"""perm-sync-utils
+
+Revision ID: ae6aaba063eb
+Revises: cec7ec36c505
+Create Date: 2025-06-11 22:18:31.061655
+
+"""
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision = "ae6aaba063eb"
+down_revision = "cec7ec36c505"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    op.add_column(
+        "document",
+        sa.Column("last_perm_synced", sa.DateTime(timezone=True), nullable=True),
+    )
+    op.create_index(
+        op.f("ix_document_last_perm_synced"),
+        "document",
+        ["last_perm_synced"],
+        unique=False,
+    )
+
+    # set parent to channel name for slack docs
+    op.add_column(
+        "document",
+        sa.Column("parent", sa.String(), nullable=True),
+    )
+    op.execute(
+        """
+        WITH slack_docs AS(
+            SELECT cc_pair.id
+            FROM document_by_connector_credential_pair AS cc_pair
+            JOIN connector ON cc_pair.connector_id = connector.id
+            WHERE connector.source = 'SLACK'
+        )
+        UPDATE document
+        SET parent = tag.tag_value
+        FROM slack_docs
+        JOIN document__tag ON slack_docs.id = document__tag.document_id
+        JOIN tag ON document__tag.tag_id = tag.id
+        WHERE
+            document.id = slack_docs.id AND
+            tag.tag_key = 'Channel'
+        """
+    )
+
+
+def downgrade() -> None:
+    op.drop_index(op.f("ix_document_last_perm_synced"), table_name="document")
+    op.drop_column("document", "last_perm_synced")
+    op.drop_column("document", "parent")

backend/ee/onyx/db/document.py

@@ -98,6 +98,7 @@ def upsert_document_external_perms(
         document.external_user_group_ids = list(prefixed_external_groups)
         document.is_public = external_access.is_public
         document.last_modified = datetime.now(timezone.utc)
+        document.last_perm_synced = datetime.now(timezone.utc)
         db_session.commit()
 
     return False

backend/onyx/connectors/models.py

@@ -145,6 +145,7 @@ class DocumentBase(BaseModel):
     primary_owners: list[BasicExpertInfo] | None = None
     # Assignee, space owner, etc.
     secondary_owners: list[BasicExpertInfo] | None = None
+    parent: str | None = None
     # title is used for search whereas semantic_identifier is used for displaying in the UI
     # different because Slack message may display as #general but general should not be part
     # of the search, at least not in the same way as a document title should be for like Confluence
@@ -204,6 +205,7 @@ class DocumentBase(BaseModel):
         else:
             size += sys.getsizeof(self.secondary_owners)
 
+        size += sys.getsizeof(self.parent)
         size += sys.getsizeof(self.title)
         size += sys.getsizeof(self.from_ingestion_api)
         size += sys.getsizeof(self.additional_info)

backend/onyx/connectors/slack/connector.py

@@ -245,6 +245,7 @@ def thread_to_doc(
         semantic_identifier=doc_sem_id,
         doc_updated_at=get_latest_message_time(thread),
         primary_owners=valid_experts,
+        parent=channel["name"],
         metadata={"Channel": channel["name"]},
         external_access=channel_access,
     )

backend/onyx/db/document.py

@@ -406,6 +406,7 @@ def upsert_documents(
                     last_modified=datetime.now(timezone.utc),
                     primary_owners=doc.primary_owners,
                     secondary_owners=doc.secondary_owners,
+                    parent=doc.parent,
                     kg_stage=KGStage.NOT_STARTED,
                     **(
                         {

backend/onyx/db/models.py

@@ -577,7 +577,7 @@ class Document(Base):
     )
     # Permission sync columns
     # Email addresses are saved at the document level for externally synced permissions
-    # This is becuase the normal flow of assigning permissions is through the cc_pair
+    # This is because the normal flow of assigning permissions is through the cc_pair
     # doesn't apply here
     external_user_emails: Mapped[list[str] | None] = mapped_column(
         postgresql.ARRAY(String), nullable=True
@@ -587,6 +587,10 @@ class Document(Base):
         postgresql.ARRAY(String), nullable=True
     )
     is_public: Mapped[bool] = mapped_column(Boolean, default=False)
+    last_perm_synced: Mapped[datetime.datetime | None] = mapped_column(
+        DateTime(timezone=True), nullable=True, index=True
+    )
+    parent: Mapped[str | None] = mapped_column(NullFilteredString, nullable=True)
 
     # tables for the knowledge graph data
     kg_stage: Mapped[KGStage] = mapped_column(

backend/onyx/document_index/interfaces.py

@@ -88,6 +88,7 @@ class DocumentMetadata:
     # Users may not be in Onyx
     primary_owners: list[str] | None = None
     secondary_owners: list[str] | None = None
+    parent: str | None = None
     from_ingestion_api: bool = False
 
     external_access: ExternalAccess | None = None

backend/onyx/indexing/indexing_pipeline.py

@@ -143,6 +143,7 @@ def _upsert_documents_in_db(
             first_link=first_link,
             primary_owners=get_experts_stores_representations(doc.primary_owners),
             secondary_owners=get_experts_stores_representations(doc.secondary_owners),
+            parent=doc.parent,
             from_ingestion_api=doc.from_ingestion_api,
             external_access=doc.external_access,
         )