Fix dbt Core external seed bootstrap

[seamus-mckinsey]

Summary

• add one all-null bootstrap record to every externally loaded package seed so dbt Core creates its typed relation before the post-hook runs
• retain existing seed column types and external-load macros; the hook still truncates the bootstrap row before loading the versioned external asset
• add a dependency-free validator and CI coverage for the placeholder contract

Why

Header-only seed files are skipped by dbt Core, so the external-load post-hook attempts to inspect or truncate a relation that does not exist. This produces BigQuery 404 errors before GCS loading begins.

Validation

• python3 scripts/check_external_seed_placeholders.py
• python3 -m unittest scripts/test_check_external_seed_placeholders.py
• dbt seed --full-refresh --select chronic_conditions__tuva_chronic_conditions_hierarchy with dbt Core + BigQuery: passed; truncate and external GCS load completed, final relation contained 1,583 rows
• dbt build --full-refresh with dbt Core + DuckDB integration project: PASS=1802 WARN=45 ERROR=0
• git diff --check

[netlify]

✅ Deploy Preview for thetuvaproject canceled.

Name	Link
🔨 Latest commit	da98798
🔍 Latest deploy log	https://app.netlify.com/projects/thetuvaproject/deploys/6a550acbe1a14700082228b0