fix: make asynq consumer actually claim and heartbeat translation tasks

ClaimNextTranslationTask and HeartbeatTask were no-ops in the asynq
Consumer, so translation tasks created in PocketBase were never picked
up by the runner. Translation tasks live in PocketBase (not Redis),
so they must be claimed/heartbeated via the underlying pb consumer.
ReapStaleTasks is also delegated so stale translation tasks get reset.

Also removes the LibreTranslate healthcheck from homelab/runner
docker-compose.yml and relaxes depends_on to service_started — the
healthcheck was blocking runner startup until models loaded (~2 min)
and the models are already pre-downloaded in the volume.

backend/internal/asynqqueue/consumer.go

@@ -10,14 +10,13 @@ import (
 
 // Consumer wraps the PocketBase-backed Consumer for result write-back only.
 //
-// When using Asynq, the runner no longer polls for work — Asynq delivers
+// When using Asynq, the runner no longer polls for scrape/audio work — Asynq
-// tasks via the ServeMux handlers. The only Consumer operations the handlers
+// delivers those tasks via the ServeMux handlers. However translation tasks
-// need are:
+// live in PocketBase (not Redis), so ClaimNextTranslationTask and HeartbeatTask
-//   - FinishAudioTask / FinishScrapeTask — write result back to PocketBase
+// still delegate to the underlying PocketBase consumer.
-//   - FailTask — mark PocketBase record as failed
 //
-// ClaimNextAudioTask, ClaimNextScrapeTask, HeartbeatTask, and ReapStaleTasks
+// ClaimNextAudioTask, ClaimNextScrapeTask are no-ops here because Asynq owns
-// are all no-ops here because Asynq owns those responsibilities.
+// those responsibilities.
 type Consumer struct {
 	pb taskqueue.Consumer // underlying PocketBase consumer (for write-back)
 }
@@ -55,10 +54,18 @@ func (c *Consumer) ClaimNextAudioTask(_ context.Context, _ string) (domain.Audio
 	return domain.AudioTask{}, false, nil
 }
 
-func (c *Consumer) ClaimNextTranslationTask(_ context.Context, _ string) (domain.TranslationTask, bool, error) {
+// ClaimNextTranslationTask delegates to PocketBase because translation tasks
-	return domain.TranslationTask{}, false, nil
+// are stored in PocketBase (not Redis/Asynq) and must still be polled directly.
+func (c *Consumer) ClaimNextTranslationTask(ctx context.Context, workerID string) (domain.TranslationTask, bool, error) {
+	return c.pb.ClaimNextTranslationTask(ctx, workerID)
 }
 
-func (c *Consumer) HeartbeatTask(_ context.Context, _ string) error { return nil }
+func (c *Consumer) HeartbeatTask(ctx context.Context, id string) error {
+	return c.pb.HeartbeatTask(ctx, id)
+}
 
-func (c *Consumer) ReapStaleTasks(_ context.Context, _ time.Duration) (int, error) { return 0, nil }
+// ReapStaleTasks delegates to PocketBase so stale translation tasks are reset
+// to pending and can be reclaimed.
+func (c *Consumer) ReapStaleTasks(ctx context.Context, staleAfter time.Duration) (int, error) {
+	return c.pb.ReapStaleTasks(ctx, staleAfter)
+}

homelab/runner/docker-compose.yml

@@ -28,20 +28,13 @@ services:
     volumes:
       - libretranslate_models:/home/libretranslate/.local/share/argos-translate
       - libretranslate_db:/app/db
-    healthcheck:
-      test: ["CMD", "python3", "-c", "import urllib.request,sys; urllib.request.urlopen('http://localhost:5000/languages'); sys.exit(0)"]
-      interval: 30s
-      timeout: 10s
-      retries: 5
-      start_period: 120s
 
   runner:
     image: kalekber/libnovel-runner:latest
     restart: unless-stopped
     stop_grace_period: 135s
     depends_on:
-      libretranslate:
+      - libretranslate
-        condition: service_healthy
     environment:
       # ── PocketBase ──────────────────────────────────────────────────────────
       POCKETBASE_URL: "https://pb.libnovel.cc"