[codex] Make jaffle shop Fusion compatible

[dataders]

Summary

• migrate deprecated dbt YAML shapes for Fusion, including generic test arguments and source freshness config
• migrate legacy Semantic Layer YAML into model-owned Fusion-compatible metadata
• refresh package-lock.yml and ignore dbt_internal_packages/
• make raw seed/source schema configurable via raw_schema, defaulting to raw; local builds can override it to a writable schema

Root cause

Fusion rejected the project before execution because the baseline YAML used deprecated generic-test argument placement, source loaded_at_field at the old level, and legacy semantic model/metric YAML. Local Snowflake builds also attempted to seed into the global raw schema, where the tester role lacks CREATE TABLE privileges.

Validation

• ~/.local/bin/dbt --version -> dbt-fusion 2.0.0-preview.190
• ~/.local/bin/dbt build --vars '{"load_source_data": true, "raw_schema": "webbeard"}' -> 49 total, 49 success\n- DBT_PROFILES_DIR=$HOME/.dbt uvx --python 3.12 --from 'dbt-metricflow[dbt-snowflake]' mf validate-configs --skip-dw --show-all --verbose-issues -> ERRORS: 0, FUTURE_ERRORS: 0, WARNINGS: 0\n\nNote: dbt sl validate is still blocked in this local environment because dbt_cloud.yml does not contain the dbt-cloud project-id configured by this repo. The MetricFlow fallback is the docs-recommended local semantic validation path when not connected to dbt platform.

[b-per]

This has been discussed a few times already.

We can't merge this before 1.12 is out otherwise dbt-core v1 users won't be able to use jaffle-shop.