What this changes

This PR fixes Plex IMDb rules so they behave more reliably when you test or run rules.

In short:

• Maintainerr now asks Plex for the extra metadata needed to read IMDb ratings
• it does that for the item itself, and also for parent or grandparent items when needed
• rule comparisons are stricter when the main value is missing, so we do not get false matches from null values
• at the same time, it keeps the older behavior that avoided a separate set of regressions in 3.2.0

Why this was needed

There were really two problems mixed together:

1. Plex was sometimes not returning the IMDb rating in the metadata we were using for rule evaluation.
2. A null-handling fix stopped one bad result, but the broader version of that fix also changed other rule behavior compared with 3.1.0.

This PR fixes both without adding extra load to the users Plex server.

What changed in practice

• Plex metadata requests used during rule evaluation now ask for external metadata, which is where the IMDb rating comes from.
• Those requests were kept intentionally narrow so we do not trigger extra Plex refresh work during a normal rule check.
• If the main value for a comparison is missing, the rule now fails safely instead of comparing through null.
• If the secondary comparison value is missing, we keep the older behavior so existing rule results do not unexpectedly change.
• This also brings back useful Test Media results for the missing-value cases we were fixing, instead of leaving those checks confusing or misleading.

Tests run

• yarn workspace @maintainerr/server test --runTestsByPath src/modules/api/plex-api/plex-api.service.spec.ts src/modules/rules/getter/plex-getter.service.spec.ts src/modules/rules/tests/rule.comparator.service.executeRulesWithData.spec.ts

Should help with #2441