Froidefond Julien
885ef7b5b2
- Nouvelle table available_downloads (library_id, series_name) unique comme source de vérité pour les téléchargements disponibles - Les jobs de détection font UPSERT (ajout/mise à jour) et DELETE (séries complètes ou sans résultat) - Après import, mise à jour ciblée : retire les volumes importés des releases, supprime l'entrée si plus de releases - Migration avec import des données existantes depuis detection_results - Endpoint latest-found simplifié : une seule query sur la table Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
21 lines
944 B
SQL
21 lines
944 B
SQL
CREATE TABLE available_downloads (
|
|
id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
|
|
library_id UUID NOT NULL REFERENCES libraries(id) ON DELETE CASCADE,
|
|
series_name TEXT NOT NULL,
|
|
missing_count INTEGER NOT NULL DEFAULT 0,
|
|
available_releases JSONB,
|
|
updated_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
|
|
UNIQUE(library_id, series_name)
|
|
);
|
|
|
|
CREATE INDEX idx_available_downloads_library ON available_downloads(library_id);
|
|
|
|
-- Migrate existing detection results into the new table
|
|
INSERT INTO available_downloads (library_id, series_name, missing_count, available_releases, updated_at)
|
|
SELECT DISTINCT ON (library_id, series_name)
|
|
library_id, series_name, missing_count, available_releases, created_at
|
|
FROM download_detection_results
|
|
WHERE status = 'found' AND available_releases IS NOT NULL
|
|
ORDER BY library_id, series_name, created_at DESC
|
|
ON CONFLICT DO NOTHING;
|