From 4f16e3b940ffcf098a5fe52f340e47807341ae8f Mon Sep 17 00:00:00 2001
From: Thomas Sileo <t@a4.io>
Date: Fri, 15 Jul 2022 08:40:07 +0200
Subject: [PATCH] Start support for processing transient object

---
 app/boxes.py               | 18 ++++++++++++++++++
 app/incoming_activities.py | 18 ++++++++++++++++--
 app/main.py                |  2 +-
 3 files changed, 35 insertions(+), 3 deletions(-)

diff --git a/app/boxes.py b/app/boxes.py
index bf42a5e..a7f9261 100644
--- a/app/boxes.py
+++ b/app/boxes.py
@@ -836,6 +836,20 @@ async def _process_note_object(
         db_session.add(notif)
 
 
+async def _process_transient_object(
+    db_session: AsyncSession,
+    raw_object: ap.RawObject,
+    from_actor: models.Actor,
+) -> None:
+    ap_type = raw_object["type"]
+    if ap_type in ["Add", "Remove"]:
+        logger.info(f"Dropping unsupported {ap_type} object")
+    else:
+        logger.warning(f"Received unknown {ap_type} object")
+
+    return None
+
+
 async def save_to_inbox(
     db_session: AsyncSession,
     raw_object: ap.RawObject,
@@ -847,6 +861,10 @@ async def save_to_inbox(
         logger.exception("Failed to fetch actor")
         return
 
+    if "id" not in raw_object:
+        await _process_transient_object(db_session, raw_object, actor)
+        return None
+
     raw_object_id = ap.get_id(raw_object)
 
     # Ensure forwarded activities have a valid LD sig
diff --git a/app/incoming_activities.py b/app/incoming_activities.py
index 1f880da..91b59f0 100644
--- a/app/incoming_activities.py
+++ b/app/incoming_activities.py
@@ -9,6 +9,7 @@ from sqlalchemy import select
 
 from app import activitypub as ap
 from app import httpsig
+from app import ldsig
 from app import models
 from app.boxes import save_to_inbox
 from app.database import AsyncSession
@@ -22,10 +23,23 @@ async def new_ap_incoming_activity(
     db_session: AsyncSession,
     httpsig_info: httpsig.HTTPSigInfo,
     raw_object: ap.RawObject,
-) -> models.IncomingActivity:
+) -> models.IncomingActivity | None:
+    ap_id: str
+    if "id" not in raw_object:
+        if "@context" not in raw_object:
+            logger.warning(f"Dropping invalid object: {raw_object}")
+            return None
+        else:
+            # This is a transient object, Build the JSON LD hash as the ID
+            ap_id = ldsig._doc_hash(raw_object)
+    else:
+        ap_id = ap.get_id(raw_object)
+
+    # TODO(ts): dedup first
+
     incoming_activity = models.IncomingActivity(
         sent_by_ap_actor_id=httpsig_info.signed_by_ap_actor_id,
-        ap_id=ap.get_id(raw_object),
+        ap_id=ap_id,
         ap_object=raw_object,
     )
     db_session.add(incoming_activity)
diff --git a/app/main.py b/app/main.py
index 5f5f6b8..e9d4095 100644
--- a/app/main.py
+++ b/app/main.py
@@ -72,7 +72,7 @@ _RESIZED_CACHE: MutableMapping[tuple[str, int], tuple[bytes, str, Any]] = LFUCac
 # TODO(ts):
 #
 # Next:
-# - incoming activity worker
+# - fix stream (only content from follows + mention, and dedup shares)
 # - handle remove activity
 # - retries httpx?
 # - DB models for webmentions