Production Readiness and Professional Standards Questions

def validate_feature_vector(fv: dict) -> (bool, list):
    """
    Validate a feature vector dict.
    Returns (is_valid, error_codes_list).
    Error codes: MISSING_<KEY>, TYPE_<KEY>, RANGE_<KEY>, MISSING_HEADER, HEADER_NOT_TRUSTED
    """
    errors = []

    # required keys and expected types
    required = {
        "user_id": str,
        "timestamp": (int, float),
        "features": dict,           # nested map of numeric features
        "model_version": str,
    }

    # numeric ranges for particular features (min, max)
    feature_ranges = {
        "click_rate": (0.0, 1.0),
        "session_length": (0.0, 36000.0),  # seconds
        "age": (0, 120),
    }

    # header flag indicating trusted ingestion
    header = fv.get("headers", {})
    if not header:
        errors.append("MISSING_HEADER")
    elif header.get("trusted") is not True:
        errors.append("HEADER_NOT_TRUSTED")

    # top-level required & types
    for k, expected in required.items():
        if k not in fv:
            errors.append(f"MISSING_{k.upper()}")
            continue
        if not isinstance(fv[k], expected):
            errors.append(f"TYPE_{k.upper()}")

    # features presence/type/range checks
    feats = fv.get("features")
    if isinstance(feats, dict):
        for fname, (lo, hi) in feature_ranges.items():
            if fname not in feats:
                errors.append(f"MISSING_FEATURE_{fname.upper()}")
                continue
            val = feats[fname]
            if not isinstance(val, (int, float)):
                errors.append(f"TYPE_FEATURE_{fname.upper()}")
                continue
            if not (lo <= float(val) <= hi):
                errors.append(f"RANGE_FEATURE_{fname.upper()}")
    else:
        # if features missing/type already flagged, avoid duplicate
        if "TYPE_FEATURES" not in errors:
            errors.append("TYPE_FEATURES")

    return (len(errors) == 0, errors)