From 13be9085b559466962b46efb1ca75d089a7f52a7 Mon Sep 17 00:00:00 2001
From: Bryan Van Deusen <bryan.vandeusen@fabledsword.com>
Date: Fri, 15 May 2026 22:23:32 -0400
Subject: [PATCH] fix(integration): asyncio auto-mode, per-test truncation,
 artist images IN-subquery

---
 backend/app/services/artist_service.py | 16 ++++++++++------
 pyproject.toml                         |  5 +++++
 tests/conftest.py                      | 26 ++++++++++++++++++++++++++
 3 files changed, 41 insertions(+), 6 deletions(-)

diff --git a/backend/app/services/artist_service.py b/backend/app/services/artist_service.py
index 9aeaa8e..c195fcb 100644
--- a/backend/app/services/artist_service.py
+++ b/backend/app/services/artist_service.py
@@ -157,12 +157,16 @@ class ArtistService:
         if artist is None:
             return None
 
-        stmt = (
-            select(ImageRecord)
-            .join(ImageProvenance, ImageProvenance.image_record_id == ImageRecord.id)
-            .join(Source, Source.id == ImageProvenance.source_id)
-            .where(Source.artist_id == artist.id)
-            .distinct()
+        # Dedupe via IN-subquery rather than JOIN + DISTINCT: an image can
+        # have several provenance rows for one artist, and SELECT DISTINCT
+        # over ImageRecord fails in Postgres because its json columns have
+        # no equality operator.
+        stmt = select(ImageRecord).where(
+            ImageRecord.id.in_(
+                select(ImageProvenance.image_record_id)
+                .join(Source, Source.id == ImageProvenance.source_id)
+                .where(Source.artist_id == artist.id)
+            )
         )
         if cursor:
             cur_ts, cur_id = decode_cursor(cursor)
diff --git a/pyproject.toml b/pyproject.toml
index eb26856..18206b0 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -9,6 +9,11 @@ where = ["."]
 include = ["backend*"]
 
 [tool.pytest.ini_options]
+# auto mode: async test fns and async fixtures don't need explicit
+# @pytest.mark.asyncio / @pytest_asyncio.fixture decoration. Required
+# because the test_api_* fixtures use plain @pytest.fixture on async
+# app/client; strict mode errors on those under pytest-asyncio 1.x.
+asyncio_mode = "auto"
 # The fast backend job runs `-m "not integration"`; the dedicated CI
 # `integration` job (pgvector Postgres + Redis service containers) runs
 # `-m integration`. Both also runnable locally via docker-compose.
diff --git a/tests/conftest.py b/tests/conftest.py
index 73b3ef9..619f4b0 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -18,6 +18,8 @@ from sqlalchemy.ext.asyncio import (
 )
 from sqlalchemy.orm import sessionmaker
 
+from backend.app.models import Base
+
 
 def _async_database_url() -> str:
     user = os.environ.get("DB_USER", "fabledcurator")
@@ -73,3 +75,27 @@ def db_sync(sync_engine):
             yield session
         finally:
             session.rollback()
+
+
+@pytest.fixture(autouse=True)
+def _truncate_after_integration(request):
+    """Integration tests exercise app/celery code that commits on its own
+    connection, which the rollback fixtures above can't undo — so data
+    would leak between tests. Truncate every model table after each
+    integration-marked test. Scoped to the integration marker so the
+    DB-less fast unit job never connects here.
+    """
+    yield
+    if request.node.get_closest_marker("integration") is None:
+        return
+    tables = ", ".join(t.name for t in Base.metadata.sorted_tables)
+    if not tables:
+        return
+    eng = create_engine(_sync_database_url(), future=True)
+    try:
+        with eng.begin() as conn:
+            conn.exec_driver_sql(
+                f"TRUNCATE {tables} RESTART IDENTITY CASCADE"
+            )
+    finally:
+        eng.dispose()