fix(migration): make 0045 DDL-only; backfill image_prediction via batched task (#768)

The inline INSERT…SELECT backfill in migration 0045 wrapped the table
creation and a ~100 GB pass over image_record.tagger_predictions in one
transaction: nothing committed until the end, it was unmonitorable, and an
earlier MATERIALIZED-CTE form spilled the full 100 GB to temp on NFS. A
deploy got stuck on it for ~2h with image_prediction never appearing.

Split the concerns:
- 0045 now creates ONLY the table + indexes (instant DDL → web boots).
- New backend.app.tasks.admin.backfill_image_predictions_task copies the
  >= store-floor predictions from the JSON into image_prediction, batched by
  id window and committed per chunk: live progress, resumable (re-enqueues
  from the last committed id), idempotent (ON CONFLICT DO NOTHING). json_each
  stays in the DB executor streaming each window — no Python-side 100 GB load,
  no materialization.
- POST /api/admin/maintenance/backfill-predictions + a Maintenance-tab card
  to trigger the one-time run after upgrading.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

frontend/src/components/settings/BackfillPredictionsCard.vue

@@ -0,0 +1,50 @@
+<template>
+  <!-- #768: one-time copy of stored tagger predictions from the
+       image_record.tagger_predictions JSON into the normalized
+       image_prediction table. Migration 0045 creates the empty table; this
+       populates it for the existing library. -->
+  <v-card>
+    <v-card-title>Backfill normalized predictions</v-card-title>
+    <v-card-text>
+      <p class="text-body-2 mb-3">
+        Copies each image's stored tagger predictions into the new
+        <code>image_prediction</code> table (the source the suggestions and
+        allowlist now read from). Run this <strong>once</strong> after the
+        upgrade so existing images get their suggestions back — newly tagged
+        images populate it automatically. Batched, resumable and idempotent;
+        safe to run more than once and to leave running in the background.
+      </p>
+      <v-btn color="primary" rounded="pill" :loading="busy" @click="run">
+        <v-icon start>mdi-database-import-outline</v-icon> Backfill predictions now
+      </v-btn>
+      <span v-if="queued" class="ml-3 text-caption text-success">Queued ✓</span>
+      <QueueStatusBar queue="maintenance_long" queue-label="Maintenance (long)" />
+    </v-card-text>
+  </v-card>
+</template>
+
+<script setup>
+import { ref } from 'vue'
+
+import { useApi } from '../../composables/useApi.js'
+import { toast } from '../../utils/toast.js'
+import QueueStatusBar from './QueueStatusBar.vue'
+
+const api = useApi()
+const busy = ref(false)
+const queued = ref(false)
+
+async function run () {
+  busy.value = true
+  queued.value = false
+  try {
+    await api.post('/api/admin/maintenance/backfill-predictions')
+    queued.value = true
+    toast({ text: 'Prediction backfill queued', type: 'success' })
+  } catch (e) {
+    toast({ text: e?.body?.detail || e?.message || 'Failed to queue', type: 'error' })
+  } finally {
+    busy.value = false
+  }
+}
+</script>

frontend/src/components/settings/MaintenancePanel.vue

@@ -12,6 +12,7 @@
       <ThumbnailBackfillCard />
     </div>
     <MLThresholdSliders class="mt-4" />
+    <BackfillPredictionsCard class="mt-4" />
     <PrunePredictionsCard class="mt-4" />
     <AllowlistTable class="mt-4" />
     <AliasTable class="mt-4" />
@@ -32,6 +33,7 @@ import MLBackfillCard from './MLBackfillCard.vue'
 import CentroidRecomputeCard from './CentroidRecomputeCard.vue'
 import ThumbnailBackfillCard from './ThumbnailBackfillCard.vue'
 import MLThresholdSliders from './MLThresholdSliders.vue'
+import BackfillPredictionsCard from './BackfillPredictionsCard.vue'
 import PrunePredictionsCard from './PrunePredictionsCard.vue'
 import AllowlistTable from './AllowlistTable.vue'
 import AliasTable from './AliasTable.vue'