Building 70 Projects ranging from beginner to advanced so anyone can — learn from, build upon, use as a reference, or even copy directly. Owned by my cybersecurity company 👇
Go aicertificationcloud-engineeringcybercybersecuritycybersecurity-analystcybersecurity-certificationcybersecurity-certificationscybersecurity-educationcybersecurity-jobscybersecurity-portfoliocybersecurity-projectscybersecurity-toolsgrchackinglearningpentestingproject-repositoryprojetcspython
test: add end-to-end training integration test
CarterPerez-dev37738b4d0b2e
diff --git a/PROJECTS/advanced/ai-threat-detection/backend/tests/test_training_e2e.py b/PROJECTS/advanced/ai-threat-detection/backend/tests/test_training_e2e.pynew file mode 100644index 00000000..fcdd467d--- /dev/null+++ b/PROJECTS/advanced/ai-threat-detection/backend/tests/test_training_e2e.py@@ -0,0 +1,97 @@+"""+©AngelaMos | 2026+test_training_e2e.py+"""++from pathlib import Path+++from app.core.detection.ensemble import (+ blend_scores,+ fuse_scores,+ normalize_ae_score,+ normalize_if_score,+)+from app.core.detection.inference import InferenceEngine+from ml.orchestrator import TrainingOrchestrator+from ml.synthetic import generate_mixed_dataset++N_NORMAL = 500+N_ATTACK = 200+N_FEATURES = 35+ENSEMBLE_WEIGHTS = {"ae": 0.4, "rf": 0.4, "if": 0.2}+++class TestTrainingE2E:+ """+ End-to-end training integration test+ """++ def test_full_training_produces_loadable_models(+ self, tmp_path: Path+ ) -> None:+ """+ Full pipeline produces models that load and predict+ """+ X, y = generate_mixed_dataset(+ N_NORMAL, N_ATTACK+ )+ assert X.shape == (+ N_NORMAL + N_ATTACK,+ N_FEATURES,+ )++ model_dir = tmp_path / "models"+ orch = TrainingOrchestrator(+ output_dir=model_dir, epochs=3+ )+ result = orch.run(X, y)++ expected_files = [+ "ae.onnx",+ "rf.onnx",+ "if.onnx",+ "scaler.json",+ "threshold.json",+ ]+ for filename in expected_files:+ assert (+ model_dir / filename+ ).exists(), f"Missing {filename}"++ engine = InferenceEngine(str(model_dir))+ assert engine.is_loaded++ sample = X[:5]+ predictions = engine.predict(sample)+ assert predictions is not None+ assert "ae" in predictions+ assert "rf" in predictions+ assert "if" in predictions+ assert len(predictions["ae"]) == 5++ threshold = engine.threshold+ for i in range(5):+ ae_score = normalize_ae_score(+ predictions["ae"][i], threshold+ )+ if_score = normalize_if_score(+ predictions["if"][i]+ )+ rf_score = predictions["rf"][i]++ scores = {+ "ae": ae_score,+ "rf": rf_score,+ "if": if_score,+ }+ fused = fuse_scores(+ scores, ENSEMBLE_WEIGHTS+ )+ assert 0.0 <= fused <= 1.0++ blended = blend_scores(fused, 0.0)+ assert 0.0 <= blended <= 1.0++ assert result.passed_gates is not None+ assert isinstance(result.passed_gates, bool)