"""L1: SQL formula evidence -- concept->formula units (spec D14b). A concept (e.g. 'fascia pediatrica') maps to a SQL formula (CASE WHEN ...) that derives it from physical columns. These are reusable, reviewable units: the gate surfaces a candidate formula, the reviewer approves or rejects it (recorded via concept_formula_approved / concept_formula_rejected), and approved formulas are part of the schema-linking artifact. The store is frontmatter-YAML + SQL body. """ from datetime import UTC, datetime from tht.decisions import DecisionRecord from tht.evidence import formula_store from tht.evidence.formula_store import ConceptFormula, retrieve_formula, save_formula from tht.evidence.session import project_session from tht.session.models import SchemaLinking def test_formula_retrieval_by_concept(tmp_path): f = ConceptFormula( concept="fascia pediatrica", columns=["data_nascita"], sql="CASE WHEN age < 18 THEN 'pediatrica' ELSE 'adulto' END", status="reviewed", sources=["src1"], ) save_formula(tmp_path, f) results = retrieve_formula(tmp_path, "fascia pediatrica") assert len(results) == 1 assert results[0].sql.startswith("CASE WHEN") assert results[0].concept == "fascia pediatrica" assert results[0].status == "reviewed" assert "data_nascita" in results[0].columns def test_save_and_reload_roundtrip_preserves_sql_body(tmp_path): f = ConceptFormula( concept="fascia pediatrica", columns=["data_nascita"], sql="CASE\n WHEN x THEN 1\nEND", status="draft", sources=[], ) path = save_formula(tmp_path, f) assert path.exists() # the file is frontmatter YAML + SQL body text = path.read_text() assert text.startswith("---") assert "concept: fascia pediatrica" in text assert "CASE" in text # SQL body preserved def test_retrieve_multiple_formulas_for_concept(tmp_path): # two competing formulas for the same concept (different sources/status) save_formula(tmp_path, ConceptFormula(concept="ablazione", columns=["flag"], sql="SELECT 1", status="draft", sources=["a"])) save_formula(tmp_path, ConceptFormula(concept="ablazione", columns=["testo"], sql="SELECT 2", status="reviewed", sources=["b"])) results = retrieve_formula(tmp_path, "ablazione") assert len(results) == 2 statuses = {r.status for r in results} assert statuses == {"draft", "reviewed"} def test_retrieve_empty_when_no_match(tmp_path): save_formula(tmp_path, ConceptFormula(concept="altro", columns=["c"], sql="SELECT 1", status="reviewed", sources=[])) assert retrieve_formula(tmp_path, "inesistente") == [] def test_retrieve_empty_on_missing_dir(tmp_path): # no formulas dir at all -> empty list, not error assert retrieve_formula(tmp_path / "nope", "anything") == [] def test_concept_formula_decision_types_exist(): import typing from tht.decisions import DecisionType args = typing.get_args(DecisionType) assert "concept_formula_approved" in args assert "concept_formula_rejected" in args def test_concept_formula_default_status(tmp_path): # status has a sensible default so an author can write a draft quickly f = ConceptFormula(concept="x", columns=["c"], sql="SELECT 1") assert f.status == "draft" # not yet reviewed assert f.sources == [] def test_reviewed_legacy_formula_becomes_curated_formula_with_stable_provenance(): formula = ConceptFormula( concept="fascia pediatrica", columns=["clinical.patient.birth_date"], sql="CASE WHEN age < 18 THEN 'pediatrica' ELSE 'adulta' END", status="reviewed", sources=["Regola clinica approvata dal gruppo pediatrico."], ) migrated = formula_store.legacy_formula_to_curated( formula, legacy_path="formulas/fascia-pediatrica-1.sql.md", source_content=formula.dump(), ) assert migrated is not None assert migrated.id.startswith("evidence:fascia-pediatrica-") assert migrated.title == "Fascia pediatrica" assert migrated.kind == "formula" assert migrated.payload.concept == formula.concept assert migrated.payload.columns == ("clinical.patient.birth_date",) assert migrated.payload.sql == formula.sql assert migrated.provenance.source_file == "source/formulas/fascia-pediatrica-1.sql.md" assert migrated.provenance.supporting_excerpts == tuple(formula.sources) assert migrated.review_items == () assert formula_store.legacy_formula_to_curated( formula, legacy_path="formulas/fascia-pediatrica-1.sql.md", source_content=formula.dump(), ).id == migrated.id def test_reviewed_legacy_formulas_with_the_same_concept_keep_distinct_path_identities(): first = ConceptFormula( concept="fascia pediatrica", columns=["clinical.patient.birth_date"], sql="CASE WHEN age < 18 THEN 'pediatrica' ELSE 'adulta' END", status="reviewed", sources=["Regola legacy revisionata."], ) second = ConceptFormula( concept="fascia pediatrica", columns=["clinical.patient.birth_date"], sql="CASE WHEN age < 16 THEN 'pediatrica' ELSE 'adulta' END", status="reviewed", sources=["Regola legacy revisionata."], ) first_migration = formula_store.legacy_formula_to_curated( first, legacy_path="formulas/fascia-pediatrica-1.sql.md", source_content=first.dump(), ) second_migration = formula_store.legacy_formula_to_curated( second, legacy_path="formulas/fascia-pediatrica-2.sql.md", source_content=second.dump(), ) assert first_migration is not None assert second_migration is not None assert first_migration.id != second_migration.id assert first_migration.id.startswith("evidence:fascia-pediatrica-") assert second_migration.id.startswith("evidence:fascia-pediatrica-") def test_session_formula_proposal_is_versioned_and_is_not_published_evidence(tmp_path): linking = SchemaLinking( question="Conta i pazienti pediatrici", concept_formulas=[{ "concept": "fascia pediatrica", "columns": ["clinical.patient.birth_date"], "sql": "CASE WHEN age < 18 THEN 'pediatrica' ELSE 'adulta' END", "sources": ["Sintetizzata nella sessione"], }], ) decisions = [DecisionRecord( seq=9, ts=datetime(2026, 8, 25, tzinfo=UTC), type="concept_formula_approved", subject="phase:4", detail="fascia pediatrica", )] projected = project_session(decisions, linking, tmp_path / "evidence") assert projected == [{ "schema_version": 1, "kind": "formula_proposal", "publication": "session_only", "concept": "fascia pediatrica", "columns": ["clinical.patient.birth_date"], "sql": "CASE WHEN age < 18 THEN 'pediatrica' ELSE 'adulta' END", "sources": ["Sintetizzata nella sessione"], "decision_seq": 9, }] def test_session_formula_proposals_require_an_unretracted_positive_f4_decision(tmp_path): linking = SchemaLinking( question="Conta i pazienti pediatrici", concept_formulas=[ {"concept": "approvata", "columns": [], "sql": "1", "sources": []}, {"concept": "rifiutata", "columns": [], "sql": "2", "sources": []}, {"concept": "indecisa", "columns": [], "sql": "3", "sources": []}, ], ) decisions = [ DecisionRecord( seq=3, ts=datetime(2026, 8, 25, tzinfo=UTC), type="concept_formula_approved", subject="phase:4", detail="approvata", ), DecisionRecord( seq=4, ts=datetime(2026, 8, 25, tzinfo=UTC), type="concept_formula_rejected", subject="phase:4", detail="rifiutata", ), ] projected = project_session(decisions, linking, tmp_path / "evidence") assert [(proposal["concept"], proposal["decision_seq"]) for proposal in projected] == [ ("approvata", 3), ]