Spaces:
Sleeping
Sleeping
File size: 7,217 Bytes
9ff494f 4e13ce7 c7ee404 9ff494f c7ee404 9ff494f 4e13ce7 9ff494f c7ee404 4e13ce7 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 | from fastapi.testclient import TestClient
from app import main as api
from app.schemas import CreditApplication, MODEL_FEATURES, TOP_SHAP_FEATURES
import numpy as np
import pandas as pd
import logging
from app.drift_monitor import DriftMonitor
AUTH_HEADER = {"Authorization": "Bearer test-token"}
class FakeModel:
def __init__(self):
self.last_frame = None
def predict_proba(self, model_frame):
self.last_frame = model_frame
return [[0.2, 0.8]]
def build_test_monitor(batch_size=2):
test_monitor = DriftMonitor.__new__(DriftMonitor)
test_monitor.batch_size = batch_size
test_monitor.drift_threshold = 0.05
test_monitor.prediction_buffer = []
test_monitor.logger = logging.getLogger("drift_monitor")
test_monitor.reference_data = pd.DataFrame([
{
"EXT_SOURCE_MEAN": 0.5,
"AMT_CREDIT": 200000,
"AMT_INCOME_TOTAL": 150000,
"DAYS_BIRTH": -15000,
"DAYS_EMPLOYED": -3000,
}
])
return test_monitor
def test_health_reports_loaded_model(monkeypatch):
fake_model = FakeModel()
monkeypatch.setattr(api, "load_model", lambda: fake_model)
monkeypatch.setattr(api, "get_model", lambda: fake_model)
with TestClient(api.app) as client:
response = client.get("/health")
assert response.status_code == 200
assert response.json() == {"status": "healthy", "model_loaded": True}
def test_features_endpoint_lists_exposed_top_shap_features(monkeypatch):
fake_model = FakeModel()
monkeypatch.setattr(api, "load_model", lambda: fake_model)
with TestClient(api.app) as client:
response = client.get("/features")
assert response.status_code == 200
assert response.json() == list(TOP_SHAP_FEATURES)
def test_predict_uses_top_features_and_imputes_hidden_features(monkeypatch):
fake_model = FakeModel()
monkeypatch.setenv("API_TOKEN", "test-token")
monkeypatch.setattr(api, "load_model", lambda: fake_model)
monkeypatch.setattr(api, "get_model", lambda: fake_model)
with TestClient(api.app) as client:
response = client.post("/predict", json={}, headers=AUTH_HEADER)
assert response.status_code == 200
assert response.json() == {
"prediction": 1,
"probability_of_default": 0.8,
"risk_category": "High",
}
expected_input = CreditApplication().to_model_input()
expected_values = [expected_input[feature] for feature in MODEL_FEATURES]
assert fake_model.last_frame.shape == (1, len(MODEL_FEATURES))
np.testing.assert_allclose(fake_model.last_frame[0], expected_values)
def test_predict_rejects_hidden_model_features(monkeypatch):
fake_model = FakeModel()
monkeypatch.setenv("API_TOKEN", "test-token")
monkeypatch.setattr(api, "load_model", lambda: fake_model)
monkeypatch.setattr(api, "get_model", lambda: fake_model)
with TestClient(api.app) as client:
response = client.post(
"/predict",
json={"CNT_CHILDREN": 2},
headers=AUTH_HEADER,
)
assert response.status_code == 422
def test_predict_returns_503_when_model_is_not_loaded(monkeypatch):
fake_model = FakeModel()
monkeypatch.setenv("API_TOKEN", "test-token")
monkeypatch.setattr(api, "load_model", lambda: fake_model)
monkeypatch.setattr(
api,
"get_model",
lambda: (_ for _ in ()).throw(RuntimeError("Model not loaded")),
)
with TestClient(api.app) as client:
response = client.post("/predict", json={}, headers=AUTH_HEADER)
assert response.status_code == 503
assert response.json()["detail"] == "Model not loaded"
def test_predict_requires_bearer_token(monkeypatch):
fake_model = FakeModel()
monkeypatch.setenv("API_TOKEN", "test-token")
monkeypatch.setattr(api, "load_model", lambda: fake_model)
monkeypatch.setattr(api, "get_model", lambda: fake_model)
with TestClient(api.app) as client:
response = client.post("/predict", json={})
assert response.status_code == 401
assert response.json()["detail"] == "Missing bearer token"
def test_predict_rejects_invalid_bearer_token(monkeypatch):
fake_model = FakeModel()
monkeypatch.setenv("API_TOKEN", "test-token")
monkeypatch.setattr(api, "load_model", lambda: fake_model)
monkeypatch.setattr(api, "get_model", lambda: fake_model)
with TestClient(api.app) as client:
response = client.post(
"/predict",
json={},
headers={"Authorization": "Bearer wrong-token"},
)
assert response.status_code == 401
assert response.json()["detail"] == "Invalid bearer token"
def test_predict_returns_503_when_api_token_is_not_configured(monkeypatch):
fake_model = FakeModel()
monkeypatch.delenv("API_TOKEN", raising=False)
monkeypatch.setattr(api, "load_model", lambda: fake_model)
monkeypatch.setattr(api, "get_model", lambda: fake_model)
with TestClient(api.app) as client:
response = client.post("/predict", json={}, headers=AUTH_HEADER)
assert response.status_code == 503
assert response.json()["detail"] == "API_TOKEN is not configured"
def test_drift_monitor_triggers(monkeypatch):
fake_model = FakeModel()
monkeypatch.setenv("API_TOKEN", "test-token")
monkeypatch.setattr(api, "load_model", lambda: fake_model)
monkeypatch.setattr(api, "get_model", lambda: fake_model)
# 🔥 Replace drift monitor with small batch size
# from app.drift_monitor import DriftMonitor
test_monitor = build_test_monitor(batch_size=3)
monkeypatch.setattr(api, "drift_monitor", test_monitor)
with TestClient(api.app) as client:
for _ in range(3): # trigger batch
response = client.post("/predict", json={}, headers=AUTH_HEADER)
assert response.status_code == 200
def test_drift_monitor_logs(monkeypatch, caplog):
fake_model = FakeModel()
monkeypatch.setenv("API_TOKEN", "test-token")
monkeypatch.setattr(api, "load_model", lambda: fake_model)
monkeypatch.setattr(api, "get_model", lambda: fake_model)
# from app.drift_monitor import DriftMonitor
test_monitor = build_test_monitor(batch_size=2)
# Fake reference data
test_monitor.reference_data = pd.DataFrame([
{
"EXT_SOURCE_MEAN": 0.5,
"AMT_CREDIT": 200000,
"AMT_INCOME_TOTAL": 150000,
"DAYS_BIRTH": -15000,
"DAYS_EMPLOYED": -3000,
}
])
monkeypatch.setattr(api, "drift_monitor", test_monitor)
caplog.set_level(logging.INFO)
with TestClient(api.app) as client:
for _ in range(2):
client.post("/predict", json={}, headers=AUTH_HEADER)
# ✅ Check drift analysis triggered
drift_logs = [
record for record in caplog.records
if "drift_analysis_completed" in record.getMessage()
]
assert len(drift_logs) > 0
# ✅ Check feature-level drift (OPTIONAL)
feature_logs = [
record for record in caplog.records
if "feature_drift_detected" in record.getMessage()
]
# You can assert or just inspect
assert isinstance(feature_logs, list)
|