fix: live YouTube search 500 on the scrape source (published_at NOT NULL)

The scrape source exposes no publish date, so its search stub carried
published_at=None. Migration 0068 (PERF-15, E2/0.68.0) made
videos.published_at NOT NULL, so insert_stubs raised IntegrityError —
outside the route's page-fetch try/except — and the whole
/api/search/youtube request 500'd ("YouTube search failed."). Every
scrape search (the default source) had been broken since 0.68.0.

- routes/search._ingest_candidates: stub published_at falls back to
  now() (mirrors stub_from_playlist_item), satisfying the constraint.
- sync.videos.apply_video_details: prefer the enrich snippet's real
  date whenever present, so the now() placeholder is corrected instead
  of becoming the video's permanent date (the guard only filled a NULL,
  which a non-null placeholder no longer is). Never clobbers a real
  date with now().

Regression tests (DB lane): dateless scrape stub ingests and takes the
real enrich date; an enrich miss keeps a non-null placeholder, unkept.

backend/app/routes/search.py

@@ -136,6 +136,10 @@ def _ingest_candidates(
 
     # 2) Video stubs, flagged via_search. Existing rows keep their flags (ON CONFLICT DO
     #    NOTHING), so an organically-synced video stays organic.
+    #    published_at is NOT NULL (migration 0068); the scrape source exposes no publish date
+    #    (only a relative "3 years ago"), so fall back to now() to satisfy the constraint — the
+    #    enrich in step 3 overwrites it with the video's real snippet date.
+    now = datetime.now(timezone.utc)
     insert_stubs(
         db,
         [
@@ -143,7 +147,7 @@ def _ingest_candidates(
                 "id": it["id"],
                 "channel_id": it["channel_id"],
                 "title": it["title"],
-                "published_at": parse_dt(it["published_at"]),
+                "published_at": parse_dt(it["published_at"]) or now,
                 "thumbnail_url": it["thumbnail_url"],
                 "via_search": True,
             }
@@ -158,7 +162,6 @@ def _ingest_candidates(
             detail_by_id = {it["id"]: it for it in yt.get_videos(ids)}
     except YouTubeError:
         detail_by_id = {}
-    now = datetime.now(timezone.utc)
     for v in videos:
         item = detail_by_id.get(v.id)
         if item is None:

backend/app/sync/videos.py

@@ -245,9 +245,15 @@ def apply_video_details(video: Video, item: dict) -> None:
         video.original_title = raw_title
         video.title = normalize_title(raw_title)
     video.description = snippet.get("description")
-    if not video.published_at:
-        # NOT NULL (0068): a fresh insert with no date yet falls back to now() (enrich normally has one).
-        video.published_at = parse_dt(snippet.get("publishedAt")) or _now()
+    # published_at is NOT NULL (0068). Prefer the enrich snippet's real date whenever it has one —
+    # this also corrects a now() placeholder written at stub time (e.g. a scrape-search stub, which
+    # carries no publish date). Only fall back to now() when the row still has nothing at all; never
+    # clobber an existing real date with now() (a re-enrich whose snippet lacks a date).
+    snippet_published = parse_dt(snippet.get("publishedAt"))
+    if snippet_published:
+        video.published_at = snippet_published
+    elif not video.published_at:
+        video.published_at = _now()
     video.thumbnail_url = best_thumbnail(snippet.get("thumbnails")) or video.thumbnail_url
     video.duration_seconds = parse_iso8601_duration(content.get("duration"))
     video.view_count = int(stats["viewCount"]) if stats.get("viewCount") else None

backend/tests/test_search_scrape_ingest.py

@@ -0,0 +1,95 @@
+"""Regression: live YouTube search ingest must survive the scrape source carrying no publish date.
+
+`videos.published_at` is NOT NULL (migration 0068, PERF-15). The scrape source exposes only a
+relative time ("3 years ago"), so its stub has `published_at=None`; the ingest used to pass that
+None straight through, which the NOT NULL constraint rejected with an IntegrityError. That error
+is raised OUTSIDE the route's page-fetch try/except, so the whole `/api/search/youtube` request
+500'd and the UI showed the generic "YouTube search failed." — every scrape search was broken.
+
+The stub now falls back to now() (a non-null placeholder), and the enrich overwrites it with the
+video's real snippet date. These DB-lane tests lock both halves: no crash on a date-less stub, and
+a placeholder never becomes the video's permanent date.
+"""
+from datetime import datetime, timezone
+
+from app import sysconfig
+from app.models import User, Video
+from app.routes.search import _ingest_candidates
+
+
+class _FakeYT:
+    """Offline stand-in for YouTubeClient. Channel enrich is a no-op (the from_search stub survives);
+    video enrich returns a real publishedAt plus a long duration so Shorts classification finalises
+    without the youtube.com/shorts probe (no network in the DB lane)."""
+
+    def __init__(self, enrich: dict[str, dict]):
+        self._enrich = enrich
+
+    def get_channels(self, ids):
+        return []
+
+    def get_videos(self, ids):
+        return [self._enrich[i] for i in ids if i in self._enrich]
+
+
+def _candidate(vid: str, ch: str) -> dict:
+    # The scrape shape (see search_scrape._video_item): published_at is ALWAYS None.
+    return {
+        "id": vid,
+        "channel_id": ch,
+        "channel_title": f"Chan {ch}",
+        "title": f"Title {vid}",
+        "published_at": None,
+        "thumbnail_url": None,
+        "live_broadcast": "none",
+    }
+
+
+def _enrich_item(vid: str, published_iso: str) -> dict:
+    # PT30M keeps the video well above shorts_probe_max_seconds so no probe fires.
+    return {
+        "id": vid,
+        "snippet": {"title": f"Title {vid}", "publishedAt": published_iso},
+        "contentDetails": {"duration": "PT30M"},
+        "statistics": {},
+        "status": {},
+    }
+
+
+def test_scrape_ingest_dateless_stub_ingests_and_takes_real_enrich_date(db):
+    # Small cap so the 1800s enrich duration keeps the shorts step offline.
+    sysconfig.set_value(db, "shorts_probe_max_seconds", 60)
+    user = User(email="searcher@example.com")
+    db.add(user)
+    db.commit()
+
+    candidates = [_candidate("vid_aaaaaaa", "UC_chan_a"), _candidate("vid_bbbbbbb", "UC_chan_a")]
+    enrich = {
+        "vid_aaaaaaa": _enrich_item("vid_aaaaaaa", "2021-02-12T19:00:07Z"),
+        "vid_bbbbbbb": _enrich_item("vid_bbbbbbb", "2023-10-30T17:00:08Z"),
+    }
+
+    kept = _ingest_candidates(db, _FakeYT(enrich), user, candidates)
+
+    # No IntegrityError, and both non-Short/non-live videos are kept in input order.
+    assert kept == ["vid_aaaaaaa", "vid_bbbbbbb"]
+    a = db.get(Video, "vid_aaaaaaa")
+    # The now() placeholder was replaced by the video's real snippet date, not left as "now".
+    assert a.published_at == datetime(2021, 2, 12, 19, 0, 7, tzinfo=timezone.utc)
+    assert a.via_search is True
+    assert a.enriched_at is not None
+
+
+def test_scrape_ingest_enrich_miss_keeps_nonnull_placeholder(db):
+    # If enrich yields nothing (proxy blip / omitted row), the stub must STILL satisfy the NOT NULL
+    # constraint (a non-null placeholder) — it simply isn't kept (enriched_at stays None).
+    user = User(email="searcher2@example.com")
+    db.add(user)
+    db.commit()
+
+    kept = _ingest_candidates(db, _FakeYT({}), user, [_candidate("vid_ccccccc", "UC_chan_c")])
+
+    assert kept == []
+    c = db.get(Video, "vid_ccccccc")
+    assert c.published_at is not None  # placeholder wrote a real value → no crash
+    assert c.enriched_at is None