fd-knowledge-graph: add query-history domain selectors to analyze and curate

[ronsse]

Summary

The runbook calls for domain-by-domain query-history analyze/curate runs, but only scout query-history had a --domain selector. Analyze and curate needed the same configured-domain filter so staged runs can target amplitude, snowplow, or customer360 without editing config files.

Stack

• Branch: stack/query-history-domain-selectors
• Base: stack/enriched-query-history-safety-config
• Parent PR: fanduel/fd-knowledge-graph#12

Expected behavior

• analyze query-history accepts repeated --domain values and passes only selected configured domains downstream.
• curate query-history accepts repeated --domain values and passes only selected configured domains downstream.
• Unknown domain names reuse the existing query-history domain validation path.
• Table scope still comes from the config; --domain narrows enrichment domains and optional domain predicates.

Safety

No live runs or graph writes are introduced. This is CLI plumbing, tests, and runbook text only.

Verification

• ruff check .
• pytest -q
• git diff --check