Rename DataAdapter -> Adapter

bayesflow/__init__.py

@@ -1,7 +1,7 @@
 from . import (
     approximators,
     benchmarks,
-    data_adapters,
+    adapters,
     datasets,
     diagnostics,
     distributions,
@@ -11,7 +11,7 @@
 )
 
 from .approximators import ContinuousApproximator
-from .data_adapters import DataAdapter
+from .adapters import Adapter
 from .datasets import OfflineDataset, OnlineDataset, DiskDataset
 from .simulators import make_simulator
 

bayesflow/adapters/__init__.py

@@ -0,0 +1,2 @@
+from . import transforms
+from .adapter import Adapter

bayesflow/data_adapters/data_adapter.py → bayesflow/adapters/adapter.py

@@ -26,16 +26,16 @@
 )
 
 
-@serializable(package="bayesflow.data_adapters")
-class DataAdapter:
+@serializable(package="bayesflow.adapters")
+class Adapter:
     def __init__(self, transforms: Sequence[Transform] | None = None):
         if transforms is None:
             transforms = []
 
         self.transforms = transforms
 
     @classmethod
-    def from_config(cls, config: dict, custom_objects=None) -> "DataAdapter":
+    def from_config(cls, config: dict, custom_objects=None) -> "Adapter":
         return cls(transforms=deserialize(config["transforms"], custom_objects))
 
     def get_config(self) -> dict:
@@ -64,7 +64,7 @@ def __call__(self, data: dict[str, any], *, inverse: bool = False, **kwargs) ->
         return self.forward(data, **kwargs)
 
     def __repr__(self):
-        return f"DataAdapter({' -> '.join(map(repr, self.transforms))})"
+        return f"Adapter([{' -> '.join(map(repr, self.transforms))}])"
 
     def add_transform(self, transform: Transform):
         self.transforms.append(transform)

bayesflow/data_adapters/transforms/broadcast.py → bayesflow/adapters/transforms/broadcast.py

@@ -24,7 +24,7 @@ class Broadcast(ElementwiseTransform):
         >>> bc(np.array(5), batch_size=3).shape
         (3,)
 
-    It is recommended to precede this transform with a :class:`bayesflow.data_adapters.transforms.ToArray` transform.
+    It is recommended to precede this transform with a :class:`bayesflow.adapters.transforms.ToArray` transform.
     """
 
     def __init__(self, *, expand_scalars: bool = True):

bayesflow/data_adapters/transforms/concatenate.py → bayesflow/adapters/transforms/concatenate.py

@@ -10,7 +10,7 @@
 from .transform import Transform
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class Concatenate(Transform):
     """Concatenate multiple arrays into a new key."""
 

bayesflow/data_adapters/transforms/constrain.py → bayesflow/adapters/transforms/constrain.py

@@ -13,7 +13,7 @@
 from .elementwise_transform import ElementwiseTransform
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class Constrain(ElementwiseTransform):
     def __init__(
         self, *, lower: int | float | np.ndarray = None, upper: int | float | np.ndarray = None, method: str = "default"

bayesflow/data_adapters/transforms/convert_dtype.py → bayesflow/adapters/transforms/convert_dtype.py

@@ -8,7 +8,7 @@
 from .elementwise_transform import ElementwiseTransform
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class ConvertDType(ElementwiseTransform):
     def __init__(self, from_dtype: str, to_dtype: str):
         super().__init__()

bayesflow/data_adapters/transforms/drop.py → bayesflow/adapters/transforms/drop.py

@@ -9,7 +9,7 @@
 from .transform import Transform
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class Drop(Transform):
     def __init__(self, keys: Sequence[str]):
         self.keys = keys

bayesflow/data_adapters/transforms/elementwise_transform.py → bayesflow/adapters/transforms/elementwise_transform.py

@@ -2,7 +2,7 @@
 import numpy as np
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class ElementwiseTransform:
     def __call__(self, data: np.ndarray, inverse: bool = False, **kwargs) -> np.ndarray:
         if inverse:

bayesflow/data_adapters/transforms/filter_transform.py → bayesflow/adapters/transforms/filter_transform.py

@@ -17,7 +17,7 @@ def __call__(self, key: str, value: np.ndarray, inverse: bool) -> bool:
         raise NotImplementedError
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class FilterTransform(Transform):
     def __init__(
         self,

bayesflow/data_adapters/transforms/keep.py → bayesflow/adapters/transforms/keep.py

@@ -9,7 +9,7 @@
 from .transform import Transform
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class Keep(Transform):
     def __init__(self, keys: Sequence[str]):
         self.keys = keys

bayesflow/data_adapters/transforms/lambda_transform.py → bayesflow/adapters/transforms/lambda_transform.py

@@ -7,7 +7,7 @@
 from .elementwise_transform import ElementwiseTransform
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class LambdaTransform(ElementwiseTransform):
     """
     Transforms a parameter using a pair of forward and inverse functions.

bayesflow/data_adapters/transforms/map_transform.py → bayesflow/adapters/transforms/map_transform.py

@@ -9,7 +9,7 @@
 from .transform import Transform
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class MapTransform(Transform):
     """
     Implements a transform that applies a set of elementwise transforms

bayesflow/data_adapters/transforms/one_hot.py → bayesflow/adapters/transforms/one_hot.py

@@ -7,7 +7,7 @@
 from .elementwise_transform import ElementwiseTransform
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class OneHot(ElementwiseTransform):
     def __init__(self, num_classes: int):
         super().__init__()

bayesflow/data_adapters/transforms/rename.py → bayesflow/adapters/transforms/rename.py

@@ -5,7 +5,7 @@
 from .transform import Transform
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class Rename(Transform):
     def __init__(self, from_key: str, to_key: str):
         super().__init__()

bayesflow/data_adapters/transforms/standardize.py → bayesflow/adapters/transforms/standardize.py

@@ -8,7 +8,7 @@
 from .elementwise_transform import ElementwiseTransform
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class Standardize(ElementwiseTransform):
     def __init__(self, mean: int | float | np.ndarray = None, std: int | float | np.ndarray = None, axis: int = None):
         super().__init__()

bayesflow/data_adapters/transforms/to_array.py → bayesflow/adapters/transforms/to_array.py

@@ -9,7 +9,7 @@
 from .elementwise_transform import ElementwiseTransform
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class ToArray(ElementwiseTransform):
     def __init__(self):
         super().__init__()

bayesflow/data_adapters/transforms/transform.py → bayesflow/adapters/transforms/transform.py

@@ -2,7 +2,7 @@
 import numpy as np
 
 
-@serializable(package="bayesflow.data_adapters")
+@serializable(package="bayesflow.adapters")
 class Transform:
     def __call__(self, data: dict[str, np.ndarray], *, inverse: bool = False, **kwargs) -> dict[str, np.ndarray]:
         if inverse:

bayesflow/approximators/approximator.py

@@ -1,7 +1,7 @@
 import keras
 import multiprocessing as mp
 
-from bayesflow.data_adapters import DataAdapter
+from bayesflow.adapters import Adapter
 from bayesflow.datasets import OnlineDataset
 from bayesflow.simulators import Simulator
 from bayesflow.utils import find_batch_size, filter_kwargs, logging
@@ -15,7 +15,7 @@ def build(self, data_shapes: any) -> None:
         self.build_from_data(mock_data)
 
     @classmethod
-    def build_data_adapter(cls, **kwargs) -> DataAdapter:
+    def build_adapter(cls, **kwargs) -> Adapter:
         # implemented by each respective architecture
         raise NotImplementedError
 
@@ -29,7 +29,7 @@ def build_dataset(
         *,
         batch_size: int = "auto",
         num_batches: int,
-        data_adapter: DataAdapter = "auto",
+        adapter: Adapter = "auto",
         memory_budget: str | int = "auto",
         simulator: Simulator,
         workers: int = "auto",
@@ -41,8 +41,8 @@ def build_dataset(
             batch_size = find_batch_size(memory_budget=memory_budget, sample=simulator.sample((1,)))
             logging.info(f"Using a batch size of {batch_size}.")
 
-        if data_adapter == "auto":
-            data_adapter = cls.build_data_adapter(**filter_kwargs(kwargs, cls.build_data_adapter))
+        if adapter == "auto":
+            adapter = cls.build_adapter(**filter_kwargs(kwargs, cls.build_adapter))
 
         if workers == "auto":
             workers = mp.cpu_count()
@@ -54,7 +54,7 @@ def build_dataset(
             simulator=simulator,
             batch_size=batch_size,
             num_batches=num_batches,
-            data_adapter=data_adapter,
+            adapter=adapter,
             workers=workers,
             use_multiprocessing=use_multiprocessing,
             max_queue_size=max_queue_size,

bayesflow/approximators/continuous_approximator.py

@@ -8,7 +8,7 @@
     serialize_keras_object as serialize,
 )
 
-from bayesflow.data_adapters import DataAdapter
+from bayesflow.adapters import Adapter
 from bayesflow.networks import InferenceNetwork, SummaryNetwork
 from bayesflow.types import Tensor
 from bayesflow.utils import logging, expand_left_to
@@ -25,43 +25,39 @@ class ContinuousApproximator(Approximator):
     def __init__(
         self,
         *,
-        data_adapter: DataAdapter,
+        adapter: Adapter,
         inference_network: InferenceNetwork,
         summary_network: SummaryNetwork = None,
         **kwargs,
     ):
         super().__init__(**kwargs)
-        self.data_adapter = data_adapter
+        self.adapter = adapter
         self.inference_network = inference_network
         self.summary_network = summary_network
 
     @classmethod
-    def build_data_adapter(
+    def build_adapter(
         cls,
         inference_variables: Sequence[str],
         inference_conditions: Sequence[str] = None,
         summary_variables: Sequence[str] = None,
-    ) -> DataAdapter:
-        data_adapter = (
-            DataAdapter()
+    ) -> Adapter:
+        adapter = (
+            Adapter()
             .to_array()
             .convert_dtype("float64", "float32")
             .concatenate(inference_variables, into="inference_variables")
         )
 
         if inference_conditions is not None:
-            data_adapter = data_adapter.concatenate(inference_conditions, into="inference_conditions")
+            adapter = adapter.concatenate(inference_conditions, into="inference_conditions")
 
         if summary_variables is not None:
-            data_adapter = data_adapter.as_set(summary_variables).concatenate(
-                summary_variables, into="summary_variables"
-            )
+            adapter = adapter.as_set(summary_variables).concatenate(summary_variables, into="summary_variables")
 
-        data_adapter = data_adapter.keep(
-            ["inference_variables", "inference_conditions", "summary_variables"]
-        ).standardize()
+        adapter = adapter.keep(["inference_variables", "inference_conditions", "summary_variables"]).standardize()
 
-        return data_adapter
+        return adapter
 
     def compile(
         self,
@@ -120,11 +116,11 @@ def compute_metrics(
         return metrics
 
     def fit(self, *args, **kwargs):
-        return super().fit(*args, **kwargs, data_adapter=self.data_adapter)
+        return super().fit(*args, **kwargs, adapter=self.adapter)
 
     @classmethod
     def from_config(cls, config, custom_objects=None):
-        config["data_adapter"] = deserialize(config["data_adapter"], custom_objects=custom_objects)
+        config["adapter"] = deserialize(config["adapter"], custom_objects=custom_objects)
         config["inference_network"] = deserialize(config["inference_network"], custom_objects=custom_objects)
         config["summary_network"] = deserialize(config["summary_network"], custom_objects=custom_objects)
 
@@ -133,7 +129,7 @@ def from_config(cls, config, custom_objects=None):
     def get_config(self):
         base_config = super().get_config()
         config = {
-            "data_adapter": serialize(self.data_adapter),
+            "adapter": serialize(self.adapter),
             "inference_network": serialize(self.inference_network),
             "summary_network": serialize(self.summary_network),
         }
@@ -148,11 +144,11 @@ def sample(
         conditions: dict[str, np.ndarray],
         **kwargs,
     ) -> dict[str, np.ndarray]:
-        conditions = self.data_adapter(conditions, strict=False, batch_size=batch_size, **kwargs)
+        conditions = self.adapter(conditions, strict=False, batch_size=batch_size, **kwargs)
         conditions = keras.tree.map_structure(keras.ops.convert_to_tensor, conditions)
         conditions = {"inference_variables": self._sample(num_samples=num_samples, batch_size=batch_size, **conditions)}
         conditions = keras.tree.map_structure(keras.ops.convert_to_numpy, conditions)
-        conditions = self.data_adapter(conditions, inverse=True, strict=False, **kwargs)
+        conditions = self.adapter(conditions, inverse=True, strict=False, **kwargs)
 
         return conditions
 
@@ -191,7 +187,7 @@ def _sample(
         return self.inference_network.sample(batch_shape, conditions=inference_conditions)
 
     def log_prob(self, data: dict[str, np.ndarray], *, batch_size: int) -> np.ndarray:
-        data = self.data_adapter(data, strict=False, batch_size=batch_size)
+        data = self.adapter(data, strict=False, batch_size=batch_size)
         data = keras.tree.map_structure(keras.ops.convert_to_tensor, data)
         log_prob = self._log_prob(**data)
         log_prob = keras.ops.convert_to_numpy(log_prob)