feat(backup): add option to skip media download for some chats

[Farzadd]

Summary

For users with certain chats or groups that generate a lot of media, this option adds the ability to skip media download for those chats only

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that causes existing functionality to change)
• Documentation update
• Infrastructure/CI change

Database Changes

• Schema changes (Alembic migration required)
• Data migration script added in scripts/
• No database changes

Data Consistency Checklist

• All chat_id values use marked format (via _get_marked_id())
• All datetime values pass through _strip_tz() before DB operations
• INSERT and UPDATE operations handle the same fields identically

Testing

• Tests pass locally (python -m pytest tests/ -v)
• Linting passes (ruff check .)
• Formatting passes (ruff format --check .)
• Manually tested in development environment

Security Checklist

• No secrets or credentials committed
• User input properly validated/sanitized
• Authentication/authorization properly checked

Deployment Notes

• Includes an update to the precommit hooks configuration as previous config did not support py314 as the target version for ruff

[GeiserX]

Thanks for the PR! The skip logic looks solid, but there's a missing piece:

When SKIP_MEDIA_CHAT_IDS is enabled for a chat that already has downloaded media, the existing media files on disk (and their database records) are left untouched. This means you end up in an inconsistent state — old media is kept but new media is never downloaded.

Requested change: When a chat is in SKIP_MEDIA_CHAT_IDS, the backup process should also delete existing media files from disk and remove the corresponding media records from the database for those chats. This way, enabling the skip actually reclaims the storage those files were using, which is likely the whole reason someone would want to skip media for a chat in the first place.

This could run once at startup or at the beginning of each backup cycle — detect chats in the skip list that still have media records, delete the files, and clean up the DB rows.

[Farzadd]

Thanks for the PR! The skip logic looks solid, but there's a missing piece:

When SKIP_MEDIA_CHAT_IDS is enabled for a chat that already has downloaded media, the existing media files on disk (and their database records) are left untouched. This means you end up in an inconsistent state — old media is kept but new media is never downloaded.

Requested change: When a chat is in SKIP_MEDIA_CHAT_IDS, the backup process should also delete existing media files from disk and remove the corresponding media records from the database for those chats. This way, enabling the skip actually reclaims the storage those files were using, which is likely the whole reason someone would want to skip media for a chat in the first place.

This could run once at startup or at the beginning of each backup cycle — detect chats in the skip list that still have media records, delete the files, and clean up the DB rows.

Thanks for the review! That's a great idea, any objections to making that an additional configuration (defaulting to TRUE)?

Otherwise, I'll update the PR with that added.

[GeiserX]

Merged! Thanks for the contribution @Farzadd, great feature and clean implementation. I'll push a follow-up commit to handle an edge case with media deduplication and then cut a v6.3.0 release including this.