From 3aa3e701951f0c8da4224dc745042545d689edbe Mon Sep 17 00:00:00 2001
From: 0xcardinal <0xcardinal@gmail.com>
Date: Thu, 22 Oct 2026 01:40:53 +0530
Subject: [PATCH] fix(ingest): fold attachment bodies in a batch onto their
 prompt

A batch sent to /v1/ingest/{source} went straight to record_ingests, so an
attachment body inside one (a replayed spool can carry it) was projected as
an event instead of folding onto its prompt as it does when sent alone. The
batch route now splits at attachment bodies, keeping order; each run of
other bodies stays one write transaction.
---
 backend/app/main.py                | 23 ++++++++++++++++++++++-
 backend/tests/test_ingest_batch.py | 24 ++++++++++++++++++++++++
 2 files changed, 46 insertions(+), 1 deletion(-)

diff --git a/backend/app/main.py b/backend/app/main.py
index 028f7d9..67f589e 100644
--- a/backend/app/main.py
+++ b/backend/app/main.py
@@ -959,6 +959,27 @@ async def _ingest_body(
     return body, None
 
 
+def _ingest_batch(source: str, bodies: list[dict[str, Any]]) -> list[dict[str, Any]]:
+    """Store a batch in order. Attachment bodies fold onto their prompt, as
+    they do sent alone (a replayed spool can carry one); each run of other
+    bodies between them is one write transaction."""
+    results: list[dict[str, Any]] = []
+    run: list[tuple[str, dict[str, Any]]] = []
+    for body in bodies:
+        if not body.get("_attach_to_prompt"):
+            run.append((source, body))
+            continue
+        if run:
+            results.extend(db.record_ingests(run))
+            run = []
+        sid = str(body.get("session_id") or "")
+        attached = db.attach_to_prompt(sid, body.get("text"), body.get("attachments") or [], body.get("timestamp"))
+        results.append({"ok": True, "attached": attached, "session_id": sid})
+    if run:
+        results.extend(db.record_ingests(run))
+    return results
+
+
 @app.post("/v1/ingest/{source}")
 async def ingest(source: str, request: Request) -> dict[str, Any]:
     if source not in AGENTS:
@@ -976,7 +997,7 @@ async def ingest(source: str, request: Request) -> dict[str, Any]:
         return result
 
     if isinstance(body, list):
-        results = await asyncio.to_thread(db.record_ingests, [(source, b) for b in body])
+        results = await asyncio.to_thread(_ingest_batch, source, body)
         for r in results:
             _bump_revision(r.get("session_id"))
         return {"ok": True, "results": results}
diff --git a/backend/tests/test_ingest_batch.py b/backend/tests/test_ingest_batch.py
index 9aa11c5..95e5ee9 100644
--- a/backend/tests/test_ingest_batch.py
+++ b/backend/tests/test_ingest_batch.py
@@ -45,3 +45,27 @@ def test_single_body_ingest_still_returns_its_result(fresh_db: Path):
     )
     assert resp.status_code == 200
     assert resp.json()["raw_status"] == "projected"
+
+
+def test_batch_folds_attachments_onto_their_prompt(fresh_db: Path):
+    """A replayed spool can carry an attachment body inside a batch; it must
+    fold onto its prompt as it does sent alone, not become an event."""
+    client = TestClient(main.app, base_url="http://127.0.0.1")
+    body = [
+        {"session_id": "att1", "hook_event_name": "UserPromptSubmit", "prompt": "see this",
+         "timestamp": "2026-07-01T00:00:00Z"},
+        {"session_id": "att1", "_attach_to_prompt": True, "text": "see this",
+         "attachments": [{"type": "image", "name": "shot.png"}], "timestamp": "2026-07-01T00:00:00Z"},
+        {"session_id": "att1", "hook_event_name": "UserPromptSubmit", "prompt": "next",
+         "timestamp": "2026-07-01T00:01:00Z"},
+    ]
+    resp = client.post("/v1/ingest/claude", json=body)
+    assert resp.status_code == 200
+    results = resp.json()["results"]
+    assert len(results) == 3 and results[1]["attached"]
+    with store.read() as conn:
+        rows = conn.execute(
+            "SELECT category, detail, attachments FROM events WHERE session_id = 'att1' ORDER BY ts"
+        ).fetchall()
+    assert [r["detail"] for r in rows if r["category"] == "prompt"] == ["see this", "next"]
+    assert "shot.png" in (rows[0]["attachments"] or "")