refactor(unit-tests): exercise dict + csv + sql fixture formats across the 3 cute-dbt dogfooding tests

[cmbays]

Summary

Follow-up to #290. The three unit_tests authored in #290 all used format: sql for given to avoid the upstream-introspection prerequisite that dict/csv require. Visual review of the rendered cute-dbt report surfaced two problems:

1. cute-dbt's renderer shows "Fixture uses format: sql — not yet visualized" for all three tests (tracked at cute-dbt#66 — the renderer fix lands in the next cute-dbt PR).
2. Even after cute-dbt#66 ships, having every test on the same format defeats the format-coverage dogfooding.

This PR re-authors the three tests so each demonstrates a different fixture format end-to-end.

Changes

Test	Given format	Expect format	Demonstrates
test_dim_payers_injects_unknown_sentinel	dict	dict	Compact key-value mock for ~22 columns
test_mart_dq_summary_combines_encounter_and_medication_metrics	csv	csv	Tabular boolean grid (compact for many dq-flag columns)
test_mart_dq_summary_zero_quarantined_when_all_valid	sql	dict	Inline SELECT with explicit casts

Plus dbt_project/README.md gains a new ## Unit tests (cute-dbt format coverage) section documenting the persistent-DuckDB target prerequisite (dict + csv given require the upstream relation to exist in the warehouse — easiest via a file-backed DuckDB target + run-operation load_synthea_sources).

Related Issues

Related to:

• feat(unit-tests): add cute-dbt dogfooding unit_tests on dim_payers + mart_dq_summary #290 (the original cute-dbt dogfooding PR)
• breezy-bays-labs/cute-dbt#39 (closed; this PR provides better dogfooding source)
• breezy-bays-labs/cute-dbt#66 (cute-dbt's matching renderer fix for csv + sql formats)

Type of Change

• Refactoring (no functional changes) — same 3 tests, same assertions, different mock-data shape

Testing

All 3 tests pass via dbt-core 1.11.11 against a persistent DuckDB target:

$ dbt run-operation load_synthea_sources --target unit_test
[All Synthea sources loaded successfully!]

$ dbt build --empty --select "+mart_dq_summary +dim_payers" --target unit_test
[Completed successfully]

$ dbt test --select "test_type:unit" --target unit_test
1 of 3 PASS dim_payers::test_dim_payers_injects_unknown_sentinel
2 of 3 PASS mart_dq_summary::test_mart_dq_summary_combines_encounter_and_medication_metrics
3 of 3 PASS mart_dq_summary::test_mart_dq_summary_zero_quarantined_when_all_valid
Completed successfully
Done. PASS=3 WARN=0 ERROR=0 SKIP=0 NO-OP=0 TOTAL=3

Manual Testing

UI-specific test items (Chrome/Firefox/Safari/localStorage) are N/A — this PR only modifies dbt YAML + README.

Specific Test Cases

1. dim_payers dict given/expect: dim_payers UNION-ALL with unknown sentinel preserves payer_key=-1 + sequenced surrogate keys starting at 1.
2. mart_dq_summary csv given/expect: 2 encounter rows (1 invalid) + 2 medication rows (both invalid) → encounters quarantined_count=1 (50% rate), medications quarantined_count=2 (100% rate).
3. mart_dq_summary sql given + dict expect: all-valid rows produce zero quarantined_count + 0.00 rate per entity. dict expect skips the non-deterministic _generated_at column.

🤖 Generated with Claude Code