NegBioDB / tests /test_etl_outcomes.py
jang1563's picture
NegBioDB final: 4 domains, fully audited
6d1bbc7
"""Tests for outcome enrichment pipeline."""
import sqlite3
from pathlib import Path
import pandas as pd
import pytest
from negbiodb_ct.ct_db import create_ct_database, get_connection
from negbiodb_ct.etl_outcomes import (
enrich_results_with_aact,
enrich_results_with_shi_du,
load_shi_du_efficacy,
load_shi_du_safety,
upgrade_confidence_tiers,
)
MIGRATIONS_DIR = Path(__file__).resolve().parent.parent / "migrations_ct"
@pytest.fixture
def ct_db(tmp_path):
"""Create a fresh CT database with all migrations applied."""
db_path = tmp_path / "test_ct.db"
create_ct_database(db_path, MIGRATIONS_DIR)
return db_path
@pytest.fixture
def enrichment_db(ct_db):
"""CT database with sample data for enrichment testing."""
conn = get_connection(ct_db)
try:
# Interventions
conn.execute(
"INSERT INTO interventions (intervention_type, intervention_name) "
"VALUES ('drug', 'DrugA')"
)
# Conditions
conn.execute("INSERT INTO conditions (condition_name) VALUES ('DiseaseX')")
# Trials
conn.execute(
"INSERT INTO clinical_trials "
"(source_db, source_trial_id, overall_status, trial_phase, has_results) "
"VALUES ('clinicaltrials_gov', 'NCT001', 'Completed', 'phase_3', 1)"
)
conn.execute(
"INSERT INTO clinical_trials "
"(source_db, source_trial_id, overall_status, trial_phase, has_results) "
"VALUES ('clinicaltrials_gov', 'NCT002', 'Terminated', 'phase_2', 0)"
)
# Junction tables
conn.execute(
"INSERT INTO trial_interventions (trial_id, intervention_id) VALUES (1, 1)"
)
conn.execute(
"INSERT INTO trial_conditions (trial_id, condition_id) VALUES (1, 1)"
)
conn.execute(
"INSERT INTO trial_interventions (trial_id, intervention_id) VALUES (2, 1)"
)
conn.execute(
"INSERT INTO trial_conditions (trial_id, condition_id) VALUES (2, 1)"
)
# Publications (for tier upgrade test)
conn.execute(
"INSERT INTO trial_publications (trial_id, pubmed_id) VALUES (1, 12345678)"
)
# Failure results — one bronze, one silver
conn.execute(
"INSERT INTO trial_failure_results "
"(intervention_id, condition_id, trial_id, "
" failure_category, confidence_tier, highest_phase_reached, "
" source_db, source_record_id, extraction_method) "
"VALUES (1, 1, 1, 'efficacy', 'bronze', 'phase_3', "
" 'clinicaltrials_gov', 'terminated:NCT001', 'nlp_classified')"
)
conn.execute(
"INSERT INTO trial_failure_results "
"(intervention_id, condition_id, trial_id, "
" failure_category, confidence_tier, highest_phase_reached, "
" source_db, source_record_id, extraction_method) "
"VALUES (1, 1, 2, 'efficacy', 'silver', 'phase_2', "
" 'clinicaltrials_gov', 'terminated:NCT002', 'nlp_classified')"
)
conn.commit()
finally:
conn.close()
return ct_db
# ============================================================
# AACT ENRICHMENT TESTS
# ============================================================
class TestEnrichResultsWithAact:
def test_updates_p_value(self, enrichment_db):
conn = get_connection(enrichment_db)
try:
aact_df = pd.DataFrame({
"nct_id": ["NCT001"],
"p_value": [0.073],
"ci_lower_limit": [-0.5],
"ci_upper_limit": [0.1],
"param_value": [-0.2],
"param_type": ["Mean Difference"],
"method": ["ANCOVA"],
})
n = enrich_results_with_aact(conn, aact_df)
assert n >= 1
row = conn.execute(
"SELECT p_value_primary, primary_endpoint_met "
"FROM trial_failure_results WHERE trial_id = 1"
).fetchone()
assert abs(row[0] - 0.073) < 1e-6
assert row[1] == 0 # p > 0.05
finally:
conn.close()
def test_endpoint_met_when_p_low(self, enrichment_db):
conn = get_connection(enrichment_db)
try:
aact_df = pd.DataFrame({
"nct_id": ["NCT002"],
"p_value": [0.003],
"ci_lower_limit": [None],
"ci_upper_limit": [None],
"param_value": [None],
"param_type": [None],
"method": ["t-test"],
})
enrich_results_with_aact(conn, aact_df)
row = conn.execute(
"SELECT primary_endpoint_met FROM trial_failure_results "
"WHERE trial_id = 2"
).fetchone()
assert row[0] == 1 # p <= 0.05
finally:
conn.close()
def test_empty_df(self, enrichment_db):
conn = get_connection(enrichment_db)
try:
n = enrich_results_with_aact(conn, pd.DataFrame())
assert n == 0
finally:
conn.close()
# ============================================================
# SHI & DU LOAD TESTS
# ============================================================
class TestLoadShiDuEfficacy:
def test_basic_load(self, tmp_path):
csv = tmp_path / "efficacy.csv"
csv.write_text("nct_id,p_value,effect_size\nNCT001,0.05,1.2\nNCT002,0.8,-0.3\n")
result = load_shi_du_efficacy(csv)
assert len(result) == 2
assert "p_value" in result.columns
def test_missing_file(self, tmp_path):
result = load_shi_du_efficacy(tmp_path / "nonexistent.csv")
assert result.empty
class TestLoadShiDuSafety:
def test_basic_load(self, tmp_path):
csv = tmp_path / "safety.csv"
csv.write_text(
"NCT_ID,serious/other,affected,at_risk\n"
"NCT001,Serious,5,100\n"
"NCT001,Other,10,100\n"
"NCT002,Serious,3,50\n"
)
result = load_shi_du_safety(csv)
# NCT001 has 5 serious affected, NCT002 has 3
assert len(result) == 2
assert "sae_total" in result.columns
nct1 = result[result["nct_id"] == "NCT001"]["sae_total"].iloc[0]
assert nct1 == 5 # Only "Serious" rows counted
def test_missing_file(self, tmp_path):
result = load_shi_du_safety(tmp_path / "nonexistent.csv")
assert result.empty
# ============================================================
# SHI & DU ENRICHMENT TESTS
# ============================================================
class TestEnrichResultsWithShiDu:
def test_updates_safety_data(self, enrichment_db):
conn = get_connection(enrichment_db)
try:
safety_df = pd.DataFrame({
"nct_id": ["NCT001"],
"sae_total": [10],
})
n = enrich_results_with_shi_du(conn, safety_df)
assert n >= 1
row = conn.execute(
"SELECT serious_adverse_events "
"FROM trial_failure_results WHERE trial_id = 1"
).fetchone()
assert row[0] == 10
finally:
conn.close()
def test_empty_df(self, enrichment_db):
conn = get_connection(enrichment_db)
try:
n = enrich_results_with_shi_du(conn, pd.DataFrame())
assert n == 0
finally:
conn.close()
# ============================================================
# TIER UPGRADE TESTS
# ============================================================
class TestUpgradeConfidenceTiers:
def test_bronze_to_silver(self, enrichment_db):
conn = get_connection(enrichment_db)
try:
# Use trial_id=2 (Phase 2, no PubMed) — won't cascade to gold
conn.execute(
"UPDATE trial_failure_results SET "
"confidence_tier = 'bronze', p_value_primary = 0.12 "
"WHERE trial_id = 2"
)
conn.commit()
stats = upgrade_confidence_tiers(conn)
assert stats["bronze_to_silver"] >= 1
row = conn.execute(
"SELECT confidence_tier FROM trial_failure_results WHERE trial_id = 2"
).fetchone()
assert row[0] == "silver"
finally:
conn.close()
def test_silver_to_gold(self, enrichment_db):
conn = get_connection(enrichment_db)
try:
# Set up silver Phase III with PubMed
# trial_id=1 is already Phase III with results and has a publication
conn.execute(
"UPDATE trial_failure_results SET "
"confidence_tier = 'silver', highest_phase_reached = 'phase_3' "
"WHERE trial_id = 1"
)
conn.commit()
stats = upgrade_confidence_tiers(conn)
assert stats["silver_to_gold"] == 1
row = conn.execute(
"SELECT confidence_tier FROM trial_failure_results WHERE trial_id = 1"
).fetchone()
assert row[0] == "gold"
finally:
conn.close()
def test_no_upgrade_without_evidence(self, enrichment_db):
conn = get_connection(enrichment_db)
try:
stats = upgrade_confidence_tiers(conn)
# Bronze result has no p-value → stays bronze
assert stats["bronze_to_silver"] == 0
finally:
conn.close()