سریال‌سازی مدل همراه با متادیتای ردیابی

import hashlib
import json
from datetime import datetime, timezone

import joblib
import sklearn

donnees_hash = hashlib.sha256(
    X_train.to_csv(index=False).encode()
).hexdigest()[:16]

bundle = {
    "pipeline": pipe,
    "meta": {
        "version": "2.4.0",
        "entraine_le": datetime.now(timezone.utc).isoformat(),
        "sklearn": sklearn.__version__,
        "hash_donnees": donnees_hash,
        "n_train": len(X_train),
        "colonnes": list(X_train.columns),
        "metriques": {"auc_cv": 0.812, "ap_cv": 0.341},
    },
}
joblib.dump(bundle, "modele_v2.4.0.joblib", compress=3)

charge = joblib.load("modele_v2.4.0.joblib")
print(json.dumps(charge["meta"], indent=2, default=str))