Tests update

README.md

@@ -11,6 +11,8 @@ license: mit
 
 # Under Construction
 
+Build a production-ready ML inference service with post-deployment drift detection, governance, and alerting, demonstrating real MLOps practices rather than offline modeling.
+
 py -3.9 -m venv .venv
 
 .venv\Scripts\activate

requirements-dev.txt

@@ -11,3 +11,5 @@ scipy>=1.10.0,<2.0.0
 python-multipart>=0.0.6
 typing-extensions>=4.0.0
 jinja2
+pytest
+httpx

tests/conftest.py

@@ -0,0 +1,9 @@
+# tests/conftest.py
+
+import sys
+from pathlib import Path
+
+# Add project root to PYTHONPATH so `import app` works
+PROJECT_ROOT = Path(__file__).parent.parent.resolve()
+if str(PROJECT_ROOT) not in sys.path:
+    sys.path.insert(0, str(PROJECT_ROOT))

tests/integration/test_api.py

@@ -0,0 +1,74 @@
+# tests/test_api.py
+
+import io
+import pandas as pd
+from fastapi.testclient import TestClient
+from app.main import app
+
+client = TestClient(app)
+
+
+def test_health_endpoint():
+    response = client.get("/health")
+    assert response.status_code == 200
+    assert response.json() == {"status": "ok"}
+
+
+def test_predict_endpoint_valid_csv():
+    df = pd.DataFrame([
+        {
+            "credit_limit": 50000.0,
+            "age": 35,
+            "pay_delay_sep": 0,
+            "pay_delay_aug": -1,
+            "bill_amt_sep": 12000.0,
+            "bill_amt_aug": 11000.0,
+            "pay_amt_sep": 3000.0,
+            "pay_amt_aug": 2500.0
+        },
+        {
+            "credit_limit": 200000.0,
+            "age": 42,
+            "pay_delay_sep": 2,
+            "pay_delay_aug": 0,
+            "bill_amt_sep": 60000.0,
+            "bill_amt_aug": 58000.0,
+            "pay_amt_sep": 10000.0,
+            "pay_amt_aug": 9000.0
+        }
+    ])
+
+    csv_bytes = df.to_csv(index=False).encode("utf-8")
+    file = io.BytesIO(csv_bytes)
+
+    response = client.post(
+        "/predict",
+        files={"file": ("test.csv", file, "text/csv")}
+    )
+
+    assert response.status_code == 200
+
+    body = response.json()
+    assert "results" in body
+    assert "drift" in body
+    assert body["n_rows"] == 2
+
+
+def test_predict_endpoint_missing_columns():
+    df = pd.DataFrame([
+        {
+            "credit_limit": 50000.0,
+            "age": 35
+        }
+    ])
+
+    csv_bytes = df.to_csv(index=False).encode("utf-8")
+    file = io.BytesIO(csv_bytes)
+
+    response = client.post(
+        "/predict",
+        files={"file": ("bad.csv", file, "text/csv")}
+    )
+
+    assert response.status_code == 400
+    assert "Invalid schema" in response.json()["detail"]

tests/integration/test_governance.py

@@ -0,0 +1,14 @@
+# tests/test_governance.py
+
+import json
+from app.monitoring.governance import run_governance_checks
+
+
+def test_governance_detects_alerts():
+    with open("tests/integration/test_governance.json", "r") as f:
+        report = json.load(f)
+
+    alerts = run_governance_checks(report, model_version="v1")
+
+    assert isinstance(alerts, list)
+    assert len(alerts) >= 0

tests/integration/test_run_drift.py

@@ -0,0 +1,25 @@
+import os
+from pathlib import Path
+import pandas as pd
+
+from app.monitoring.drift import run_drift_check
+
+def test_run_drift_check_outputs_metrics():
+    repo_root = Path(__file__).resolve().parents[2]
+
+    current_path = repo_root / "data" / "processed" / "current_data.csv"
+    reference_path = repo_root / "models" / "v1" / "reference_data.csv"
+
+    assert current_path.exists()
+    assert reference_path.exists()
+
+    current_df = pd.read_csv(current_path)
+    reference_df = pd.read_csv(reference_path)
+
+    report = run_drift_check(
+        current_df,
+        reference_df,
+        model_version="v1"
+    )
+
+    assert report is not None

tests/test_governance.py

@@ -1,15 +0,0 @@
-import sys
-from pathlib import Path
-import json
-
-repo_root = Path(__file__).parent.parent.resolve()
-sys.path.insert(0, str(repo_root))
-
-from app.monitoring.governance import run_governance_checks
-
-# Load a sample report JSON (create this for testing)
-with open('tests/test_governance.json', 'r') as f:
-    report = json.load(f)
-
-alerts = run_governance_checks(report, model_version="v1")
-print("Governance alerts:", alerts)

tests/test_run_drift.py

@@ -1,35 +0,0 @@
-import sys
-import os
-import pandas as pd
-
-# Ensure project root is in sys.path
-sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
-
-from app.monitoring.drift import run_drift_check
-from app.monitoring.governance import run_governance_checks
-
-def main():
-    root_dir = os.path.dirname(os.path.dirname(__file__))
-    
-    # Load current and reference data
-    current_path = os.path.join(root_dir, "data", "processed", "current_data.csv")
-    reference_path = os.path.join(root_dir, "models", "v1", "reference_data.csv")
-
-    if not os.path.exists(current_path):
-        raise FileNotFoundError(f"{current_path} does not exist.")
-    if not os.path.exists(reference_path):
-        raise FileNotFoundError(f"{reference_path} does not exist.")
-
-    current_df = pd.read_csv(current_path)
-    reference_df = pd.read_csv(reference_path)
-
-    # Run drift check
-    report = run_drift_check(current_df, reference_df, model_version="v1")
-
-    # Run drift check
-    print("Metrics from Evidently report:", report)
-
-
-
-if __name__ == "__main__":
-    main()

tests/unit/test_schemas.py

@@ -0,0 +1,59 @@
+# tests/test_schemas.py
+
+import pytest
+from pydantic import ValidationError
+from app.api.schemas import PredictionRequest, PredictionResponse
+
+
+def test_prediction_request_valid():
+    payload = {
+        "credit_limit": 50000.0,
+        "age": 35,
+        "pay_delay_sep": 0,
+        "pay_delay_aug": -1,
+        "bill_amt_sep": 12000.0,
+        "bill_amt_aug": 11000.0,
+        "pay_amt_sep": 3000.0,
+        "pay_amt_aug": 2500.0
+    }
+
+    req = PredictionRequest(**payload)
+    assert req.credit_limit == 50000.0
+    assert req.age == 35
+
+
+def test_prediction_request_missing_field():
+    payload = {
+        "credit_limit": 50000.0,
+        "age": 35
+    }
+
+    with pytest.raises(ValidationError):
+        PredictionRequest(**payload)
+
+
+def test_prediction_request_invalid_type():
+    payload = {
+        "credit_limit": "not-a-number",
+        "age": "thirty",
+        "pay_delay_sep": 0,
+        "pay_delay_aug": 0,
+        "bill_amt_sep": 1000.0,
+        "bill_amt_aug": 1000.0,
+        "pay_amt_sep": 100.0,
+        "pay_amt_aug": 100.0
+    }
+
+    with pytest.raises(ValidationError):
+        PredictionRequest(**payload)
+
+
+def test_prediction_response_valid():
+    payload = {
+        "prediction": 1,
+        "probability": 0.82
+    }
+
+    resp = PredictionResponse(**payload)
+    assert resp.prediction in (0, 1)
+    assert 0.0 <= resp.probability <= 1.0