Python binding

n4m is the full current ctypes binding over libn4m (ABI 2.5). It exposes ABI introspection, context/config lifecycles, and the method surface through role subpackages that mirror the n4m.<role> namespace. See bindings/python/README.md for installation and loader rules, and the ABI 2.0 migration guide for the old→new mapping.

Hello-version

import n4m

print(n4m.abi_version())          # (2, 5, 0)

with n4m.Context() as ctx:
    ctx.seed = 42
    assert ctx.seed == 42

Role packages (ABI 2.x)

n4m.__init__ exposes only metadata/helpers and the role subpackages. Public classes use plain role names (not Native*):

from n4m.estimators.regression.regularized import Ridge, RidgePLS
from n4m.estimators.regression.latent import PLS, PCR
from n4m.transform.scatter import SNV, MSC
from n4m.transform.smoothing import SavitzkyGolay
from n4m.feature_selection.wrapper import CARS
from n4m.model_selection.splitters import KennardStone
from n4m.domain_adaptation.orthogonalization import EPO
from n4m.augmentation.noise import GaussianAdditiveNoise
from n4m.ensemble import AOMRidgeBlender
from n4m.compose.aom_superblock import AOMRidgePLSSuperblock
from n4m.decomposition import FlexiblePCA

The estimators/transformers are sklearn-compatible with zero-copy NumPy n4m_matrix_view_t round-trips. The separately shipped pls4all package remains an active slim PLS-focused compatibility subset over the same ABI-2 library. Method pages use n4m imports only when their exact current public role export was verified from source; they do not substitute a pls4all route for a full n4m method.

For verified migration tooling, pls4all.export_linear_predictor_n4mm(...) constructs a PREDICT-only N4MM from finite affine coefficients and intercepts. It does not deserialize a foreign pickle/joblib model or retrain; the caller must first attest that the source predictor is exactly affine. The imported model intentionally does not expose a latent PLS transform.