# Author: LKouadio <etanoyau@gmail.com>
# License: LGPL-3.0
"""Artificial intelligence and machine learning for EM processing and inversion.
Phase 2 additions
-----------------
* :class:`~pycsamt.ai.inversion.inv1d.EMInverter1D` — full 1-D inversion
pipeline (data loading → normalisation → training → prediction)
* :mod:`~pycsamt.ai.nets` — CNN1D, ResNet1D, FCN1D architectures
* :mod:`~pycsamt.ai.training` — EMDataset, EMTrainer, metrics
* :mod:`~pycsamt.ai.plot` — EMStyle, plot_compare, plot_convergence
Quick start (requires PyTorch)
------------------------------
>>> from pycsamt.forward.batch import generate_dataset
>>> from pycsamt.ai.inversion import EMInverter1D
>>> ds = generate_dataset(n_samples=2_000, seed=0, n_layers=5)
>>> inv = EMInverter1D(arch="resnet", n_layers=5)
>>> inv.fit(ds, epochs=30)
>>> # Predict on new Z objects or ForwardResponse
>>> y_pred = inv.predict(X_test)
"""
from ._base import BaseEMNet, BaseEMProcessor, EMCheckpoint
from ._zoo import (
download_checkpoint,
get_pretrained_info,
list_pretrained,
)
from .inversion import (
ConformalPredictor,
EMInverter1D,
EMInverter2D,
EnsembleInverter,
GCNInverter3D,
JointInverter,
PosteriorCalibrator,
)
from .nets import (
CNN1DNet,
DRCNNNet,
FCN1DNet,
GCNNet,
ResNet1DNet,
UNet2DNet,
build_adjacency,
)
from .plot import (
EM_CMAPS,
EM_COLORS,
EM_FIGSIZE,
EMStyle,
add_colorbar,
em_context,
plot_compare,
plot_confusion_matrix,
plot_convergence,
plot_feature_importance,
plot_layer_errors,
plot_lr_schedule,
plot_profile_pair,
plot_pseudo_section,
plot_residuals,
plot_section,
plot_section_pair,
plot_uncertainty_bands,
)
from .processing import (
AnomalyDetector,
DimensionalityClassifier,
EMDenoiser,
EMQCScorer,
prepare_z_features,
)
from .training import (
AugmentFreqDrop,
AugmentMixup,
AugmentNoise,
AugmentStaticShift,
Compose,
EMDataset,
EMTrainer,
Normalizer,
RandomApply,
depth_rmse,
layer_rmse,
mae,
masked_mse_loss,
r2,
relative_rmse,
rmse,
summarise,
)
__all__ = [
# base
"BaseEMNet",
"BaseEMProcessor",
"EMCheckpoint",
# inversion
"EMInverter1D",
"EMInverter2D",
"GCNInverter3D",
"JointInverter",
"EnsembleInverter",
# calibrated UQ
"ConformalPredictor",
"PosteriorCalibrator",
# model zoo
"list_pretrained",
"get_pretrained_info",
"download_checkpoint",
# nets
"CNN1DNet",
"ResNet1DNet",
"FCN1DNet",
"UNet2DNet",
"DRCNNNet",
"GCNNet",
"build_adjacency",
# training
"Normalizer",
"EMDataset",
"EMTrainer",
"rmse",
"mae",
"r2",
"relative_rmse",
"depth_rmse",
"layer_rmse",
"masked_mse_loss",
"summarise",
# augmentation (Phase 5)
"AugmentNoise",
"AugmentStaticShift",
"AugmentFreqDrop",
"AugmentMixup",
"Compose",
"RandomApply",
# plot
"EMStyle",
"EM_COLORS",
"EM_CMAPS",
"EM_FIGSIZE",
"em_context",
"add_colorbar",
"plot_compare",
"plot_profile_pair",
"plot_convergence",
"plot_lr_schedule",
# section & diagnostics (Phase 4)
"plot_section",
"plot_section_pair",
"plot_pseudo_section",
"plot_confusion_matrix",
"plot_residuals",
"plot_layer_errors",
"plot_uncertainty_bands",
"plot_feature_importance",
# processing (Phase 3)
"EMDenoiser",
"prepare_z_features",
"EMQCScorer",
"AnomalyDetector",
"DimensionalityClassifier",
]
