feat(arrs): optimize blocklist precision with local database lookups

[drondeseries]

Summary

This PR optimizes blocklisting precision for Sonarr/Radarr integrations in Altmount by checking local database import records in addition to active download queues.

Detailed Changes

1. Scanner DB Lookup: Updates the Sonarr/Radarr scanner (internal/arrs) to query the local database database for mapped download IDs associated with target file paths, resolving edge cases where history records were otherwise lost/not found in the active client queues.

[drondeseries]

I have updated the PR to resolve the questions around import_history transience:

• Moved Tracking to file_health Table: Added a dedicated download_id column to the file_health table. This serves as the persistent database layout for tracking library files and mapping them to ARR download/NZB IDs independent of the transient import_history.
• Database Migration & Backfill: Added database migration 035_add_download_id_to_file_health.sql for both Postgres and SQLite. The migration automatically backfills existing file_health entries using matching paths from the import_history table (if history hasn't been cleared yet).
• Import & Webhook Integration: Updated AddFileToHealthCheckWithMetadata and BatchAddFileToHealthCheck to store download_id upon initial import and webhook triggers.
• Removed History Dependency in Sync: Cleaned up the sync safety check so it relies on the active library records rather than history.

[drondeseries]

I have pushed the updated clean commits branch on top of upstream main. Here is the full breakdown of the 6 commits included in this PR and what they address:

1. feat(arrs): optimize blocklist precision... (8dfacbe):

   • What it does: Optimizes blocklisting checks in the Sonarr/Radarr scanner (internal/arrs) by checking local database records in addition to active download queues.

2. chore: add local IDE and scanner... (bddfd4f):

   • What it does: Updates .gitignore to exclude assistant and semantic indexing directories (.antigravitycli, .codegraph, .medusa) from dirtying the git status.

3. feat(health): implement streaming failure masking... (7b69972):

   • What it does: Prevents minor or transient network/article errors from immediately triggering repair loops. It introduces a configurable threshold (failure masking) before marking a streaming failure as corrupted.

4. fix(quota): restore provider quota... (5266351):

   • What it does: Corrects NNTP provider quota display issues, ensuring automatic quota resets and database persistence operate cleanly.

5. fix(importer): normalize multi-volume archives... (56322a6):

   • What it does: Handles imports of obfuscated multi-volume archive sets without PAR2 metadata by correctly resolving file numbering sequences.

6. feat(health): persist download_id in file_health... (e3eb8e6):

   • What it does: Decouples the health and repair system from the transient import_history table. It introduces a persistent download_id column to the file_health table, allowing users to prune their import history (e.g., via HistoryRetentionDays) without breaking ARR automatic repairs. Includes automatic database migration backfills from existing history records.