Add __all__ for core part

ignite/engine/__init__.py

@@ -3,6 +3,13 @@
 from ignite.engine.engine import Engine, State, Events
 from ignite.utils import convert_tensor
 
+__all__ = [
+    'create_supervised_trainer',
+    'create_supervised_evaluator',
+    'Engine',
+    'Events'
+]
+
 
 def _prepare_batch(batch, device=None, non_blocking=False):
     """Prepare batch for training: pass to a device with options.

ignite/engine/engine.py

@@ -13,6 +13,12 @@
 
 from ignite._utils import _to_hours_mins_secs
 
+__all__ = [
+    'Engine',
+    'Events',
+    'State'
+]
+
 
 class EventWithFilter:
 

ignite/exceptions.py

@@ -1,3 +1,8 @@
+__all__ = [
+    'NotComputableError'
+]
+
+
 class NotComputableError(RuntimeError):
     """
     Exception class to raise if Metric cannot be computed.

ignite/handlers/__init__.py

@@ -3,6 +3,16 @@
 from ignite.handlers.early_stopping import EarlyStopping
 from ignite.handlers.terminate_on_nan import TerminateOnNan
 
+__all__ = [
+    'ModelCheckpoint',
+    'Checkpoint',
+    'DiskSaver',
+    'Timer',
+    'EarlyStopping',
+    'TerminateOnNan',
+    'global_step_from_engine'
+]
+
 
 def global_step_from_engine(engine):
     """Helper method to setup `global_step_transform` function using another engine.

ignite/handlers/checkpoint.py

@@ -9,6 +9,11 @@
 
 from ignite.engine import Events
 
+__all__ = [
+    'Checkpoint',
+    'ModelCheckpoint'
+]
+
 
 class Checkpoint:
     """Checkpoint handler can be used to periodically save and load objects which have attribute

ignite/handlers/early_stopping.py

@@ -2,6 +2,10 @@
 
 from ignite.engine import Engine
 
+__all__ = [
+    'EarlyStopping'
+]
+
 
 class EarlyStopping:
     """EarlyStopping handler can be used to stop the training if no improvement after a given number of events.

ignite/handlers/terminate_on_nan.py

@@ -5,6 +5,10 @@
 
 from ignite.utils import apply_to_type
 
+__all__ = [
+    'TerminateOnNan'
+]
+
 
 class TerminateOnNan:
     """TerminateOnNan handler can be used to stop the training if the `process_function`'s output

ignite/handlers/timing.py

@@ -2,6 +2,10 @@
 
 from ignite.engine import Events
 
+__all__ = [
+    'Timer'
+]
+
 
 class Timer:
     """ Timer object can be used to measure (average) time between events.

ignite/metrics/__init__.py

@@ -14,3 +14,27 @@
 from ignite.metrics.confusion_matrix import ConfusionMatrix, IoU, mIoU, DiceCoefficient
 from ignite.metrics.accumulation import VariableAccumulation, Average, GeometricAverage
 from ignite.metrics.fbeta import Fbeta
+
+__all__ = [
+    'Metric',
+    'Accuracy',
+    'Loss',
+    'MetricsLambda',
+    'MeanAbsoluteError',
+    'MeanPairwiseDistance',
+    'MeanSquaredError',
+    'ConfusionMatrix',
+    'TopKCategoricalAccuracy',
+    'Average',
+    'DiceCoefficient',
+    'EpochMetric',
+    'Fbeta',
+    'GeometricAverage',
+    'IoU',
+    'mIoU',
+    'Precision',
+    'Recall',
+    'RootMeanSquaredError',
+    'RunningAverage',
+    'VariableAccumulation'
+]

ignite/metrics/accumulation.py

@@ -6,6 +6,12 @@
 
 import torch
 
+__all__ = [
+    'VariableAccumulation',
+    'GeometricAverage',
+    'Average'
+]
+
 
 class VariableAccumulation(Metric):
     """Single variable accumulator helper to compute (arithmetic, geometric, harmonic) average of a single variable.

ignite/metrics/accuracy.py

@@ -4,6 +4,10 @@
 
 import torch
 
+__all__ = [
+    'Accuracy'
+]
+
 
 class _BaseClassification(Metric):
 

ignite/metrics/confusion_matrix.py

@@ -6,6 +6,16 @@
 from ignite.exceptions import NotComputableError
 from ignite.metrics.metric import sync_all_reduce, reinit__is_reduced
 
+__all__ = [
+    'ConfusionMatrix',
+    'mIoU',
+    'IoU',
+    'DiceCoefficient',
+    'cmAccuracy',
+    'cmPrecision',
+    'cmRecall'
+]
+
 
 class ConfusionMatrix(Metric):
     """Calculates confusion matrix for multi-class data.

ignite/metrics/epoch_metric.py

@@ -4,6 +4,10 @@
 
 from ignite.metrics.metric import Metric
 
+__all__ = [
+    'EpochMetric'
+]
+
 
 class EpochMetric(Metric):
     """Class for metrics that should be computed on the entire output history of a model.

ignite/metrics/fbeta.py

@@ -1,5 +1,9 @@
 from ignite.metrics import Precision, Recall
 
+__all__ = [
+    'Fbeta'
+]
+
 
 def Fbeta(beta, average=True, precision=None, recall=None, output_transform=None, device=None):
     """Calculates F-beta score

ignite/metrics/loss.py

@@ -2,6 +2,10 @@
 from ignite.metrics import Metric
 from ignite.metrics.metric import sync_all_reduce, reinit__is_reduced
 
+__all__ = [
+    'Loss'
+]
+
 
 class Loss(Metric):
     """

ignite/metrics/mean_absolute_error.py

@@ -4,6 +4,10 @@
 from ignite.metrics.metric import Metric
 from ignite.metrics.metric import sync_all_reduce, reinit__is_reduced
 
+__all__ = [
+    'MeanAbsoluteError'
+]
+
 
 class MeanAbsoluteError(Metric):
     """

ignite/metrics/mean_pairwise_distance.py

@@ -5,6 +5,10 @@
 from ignite.metrics.metric import Metric
 from ignite.metrics.metric import sync_all_reduce, reinit__is_reduced
 
+__all__ = [
+    'MeanPairwiseDistance'
+]
+
 
 class MeanPairwiseDistance(Metric):
     """

ignite/metrics/mean_squared_error.py

@@ -4,6 +4,10 @@
 from ignite.metrics.metric import Metric
 from ignite.metrics.metric import sync_all_reduce, reinit__is_reduced
 
+__all__ = [
+    'MeanSquaredError'
+]
+
 
 class MeanSquaredError(Metric):
     """

ignite/metrics/metric.py

@@ -9,6 +9,10 @@
 
 from ignite.engine import Events
 
+__all__ = [
+    'Metric'
+]
+
 
 class Metric(metaclass=ABCMeta):
     """

ignite/metrics/metrics_lambda.py

@@ -3,6 +3,10 @@
 from ignite.metrics.metric import Metric, reinit__is_reduced
 from ignite.engine import Events
 
+__all__ = [
+    'MetricsLambda'
+]
+
 
 class MetricsLambda(Metric):
     """

ignite/metrics/precision.py

@@ -7,6 +7,10 @@
 from ignite.utils import to_onehot
 from ignite.metrics.metric import reinit__is_reduced
 
+__all__ = [
+    'Precision'
+]
+
 
 class _BasePrecisionRecall(_BaseClassification):
 

ignite/metrics/recall.py

@@ -4,6 +4,10 @@
 from ignite.utils import to_onehot
 from ignite.metrics.metric import reinit__is_reduced
 
+__all__ = [
+    'Recall'
+]
+
 
 class Recall(_BasePrecisionRecall):
     """

ignite/metrics/root_mean_squared_error.py

@@ -2,6 +2,10 @@
 
 from ignite.metrics.mean_squared_error import MeanSquaredError
 
+__all__ = [
+    'RootMeanSquaredError'
+]
+
 
 class RootMeanSquaredError(MeanSquaredError):
     """

ignite/metrics/running_average.py

@@ -2,6 +2,10 @@
 from ignite.metrics import Metric
 from ignite.metrics.metric import reinit__is_reduced, sync_all_reduce
 
+__all__ = [
+    'RunningAverage'
+]
+
 
 class RunningAverage(Metric):
     """Compute running average of a metric or the output of process function.

ignite/metrics/top_k_categorical_accuracy.py

@@ -4,6 +4,10 @@
 from ignite.exceptions import NotComputableError
 from ignite.metrics.metric import sync_all_reduce, reinit__is_reduced
 
+__all__ = [
+    'TopKCategoricalAccuracy'
+]
+
 
 class TopKCategoricalAccuracy(Metric):
     """

ignite/utils.py

@@ -3,6 +3,14 @@
 
 import torch
 
+__all__ = [
+    'convert_tensor',
+    'apply_to_tensor',
+    'apply_to_type',
+    'to_onehot',
+    'setup_logger'
+]
+
 
 def convert_tensor(input_, device=None, non_blocking=False):
     """Move tensors to relevant device."""