feat(graphile-llm): add onQuotaExceeded option to control embedding failure behavior

[pyramation]

Summary

Adds onQuotaExceeded: 'degrade' | 'throw' option to createLlmTextSearchPlugin() and GraphileLlmPreset().

When the embedder returns null (billing quota exceeded, provider unavailable), the current behavior silently falls back to text-only search. This is fine for most cases, but some apps need semantic search to be critical — keyword-only results are unacceptable.

// Default: graceful degradation (skip vector, text adapters still work)
GraphileLlmPreset({ onQuotaExceeded: 'degrade' })

// Strict: throw error even if text adapters could handle it
GraphileLlmPreset({ onQuotaExceeded: 'throw' })

When no text adapters exist on the table, throws regardless of this setting (no fallback possible).

The option flows through: GraphileLlmOptions.onQuotaExceeded → createLlmTextSearchPlugin({ onQuotaExceeded }) → embedTextInWhere(where, embedder, hasTextAdapters, onQuotaExceeded).

6 new unit tests covering both modes + edge cases (no text adapters, successful embed in throw mode).

Link to Devin session: https://app.devin.ai/sessions/4a9f098c74fb4cb6a9b6868fcff321db
Requested by: @pyramation

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment, CI, and merge conflict monitoring