dataeval 0.71.0__tar.gz → 0.71.1__tar.gz

{dataeval-0.71.0 → dataeval-0.71.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: dataeval
-Version: 0.71.0
+Version: 0.71.1
 Summary: DataEval provides a simple interface to characterize image data and its impact on model performance across classification and object-detection tasks
 Home-page: https://dataeval.ai/
 License: MIT
@@ -29,8 +29,9 @@ Requires-Dist: nvidia-cudnn-cu11 (>=8.6.0.163) ; extra == "tensorflow" or extra
 Requires-Dist: pillow (>=10.3.0)
 Requires-Dist: scikit-learn (>=1.5.0)
 Requires-Dist: scipy (>=1.10)
-Requires-Dist: tensorflow (>=2.14.1,<2.16) ; extra == "tensorflow" or extra == "all"
-Requires-Dist: tensorflow_probability (>=0.22.1,<0.24) ; extra == "tensorflow" or extra == "all"
+Requires-Dist: tensorflow (>=2.16) ; extra == "tensorflow" or extra == "all"
+Requires-Dist: tensorflow_probability (>=0.24) ; extra == "tensorflow" or extra == "all"
+Requires-Dist: tf-keras (>2.16) ; extra == "tensorflow" or extra == "all"
 Requires-Dist: torch (>=2.2.0) ; extra == "torch" or extra == "all"
 Requires-Dist: torchvision (>=0.17.0) ; extra == "torch" or extra == "all"
 Requires-Dist: tqdm

{dataeval-0.71.0 → dataeval-0.71.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "dataeval"
-version = "0.71.0" # dynamic
+version = "0.71.1" # dynamic
 description = "DataEval provides a simple interface to characterize image data and its impact on model performance across classification and object-detection tasks"
 license = "MIT"
 readme = "README.md"
@@ -52,15 +52,16 @@ xxhash = {version = ">=3.3"}
 # optional
 matplotlib = {version = "*", optional = true}
 nvidia-cudnn-cu11 = {version = ">=8.6.0.163", optional = true}
-tensorflow = {version = ">=2.14.1, <2.16", optional = true}
-tensorflow_probability = {version = ">=0.22.1, <0.24", optional = true}
+tensorflow = {version = ">=2.16", optional = true}
+tensorflow_probability = {version = ">=0.24", optional = true}
+tf-keras = {version = ">2.16", optional = true}
 torch = {version = ">=2.2.0", source = "pytorch", optional = true}
 torchvision = {version = ">=0.17.0", source = "pytorch", optional = true}
 
 [tool.poetry.extras]
-tensorflow = ["tensorflow", "tensorflow_probability", "nvidia-cudnn-cu11"]
+tensorflow = ["tensorflow", "tensorflow_probability", "tf-keras", "nvidia-cudnn-cu11"]
 torch = ["torch", "torchvision", "matplotlib", "nvidia-cudnn-cu11"]
-all = ["matplotlib", "nvidia-cudnn-cu11", "tensorflow", "tensorflow_probability", "torch", "torchvision"]
+all = ["matplotlib", "nvidia-cudnn-cu11", "tensorflow", "tensorflow_probability", "tf-keras", "torch", "torchvision"]
 
 [tool.poetry.group.dev]
 optional = true
@@ -131,6 +132,7 @@ reportMissingImports = false
 
 [tool.pytest.ini_options]
 norecursedirs = ["prototype"]
+testpaths = ["tests"]
 addopts = ["--pythonwarnings=ignore::DeprecationWarning", "--verbose", "--durations=20", "--durations-min=1.0"]
 
 [tool.coverage.run]

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/__init__.py

@@ -1,8 +1,9 @@
-__version__ = "0.71.0"
+__version__ = "0.71.1"
 
 from importlib.util import find_spec
 
 _IS_TORCH_AVAILABLE = find_spec("torch") is not None
+_IS_TORCHVISION_AVAILABLE = find_spec("torchvision") is not None
 _IS_TENSORFLOW_AVAILABLE = find_spec("tensorflow") is not None and find_spec("tensorflow_probability") is not None
 
 del find_spec

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/detectors/ood/ae.py

@@ -10,9 +10,9 @@ from __future__ import annotations
 
 from typing import Callable
 
-import keras
 import numpy as np
 import tensorflow as tf
+import tf_keras as keras
 from numpy.typing import ArrayLike
 
 from dataeval._internal.detectors.ood.base import OODBase, OODScoreOutput

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/detectors/ood/aegmm.py

@@ -10,8 +10,8 @@ from __future__ import annotations
 
 from typing import Callable
 
-import keras
 import tensorflow as tf
+import tf_keras as keras
 from numpy.typing import ArrayLike
 
 from dataeval._internal.detectors.ood.base import OODGMMBase, OODScoreOutput

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/detectors/ood/base.py

@@ -12,9 +12,9 @@ from abc import ABC, abstractmethod
 from dataclasses import dataclass
 from typing import Callable, Literal, cast
 
-import keras
 import numpy as np
 import tensorflow as tf
+import tf_keras as keras
 from numpy.typing import ArrayLike, NDArray
 
 from dataeval._internal.interop import to_numpy

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/detectors/ood/llr.py

@@ -11,12 +11,12 @@ from __future__ import annotations
 from functools import partial
 from typing import Callable
 
-import keras
 import numpy as np
 import tensorflow as tf
-from keras.layers import Input
-from keras.models import Model
+import tf_keras as keras
 from numpy.typing import ArrayLike, NDArray
+from tf_keras.layers import Input
+from tf_keras.models import Model
 
 from dataeval._internal.detectors.ood.base import OODBase, OODScoreOutput
 from dataeval._internal.interop import to_numpy

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/detectors/ood/vae.py

@@ -10,9 +10,9 @@ from __future__ import annotations
 
 from typing import Callable
 
-import keras
 import numpy as np
 import tensorflow as tf
+import tf_keras as keras
 from numpy.typing import ArrayLike
 
 from dataeval._internal.detectors.ood.base import OODBase, OODScoreOutput

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/detectors/ood/vaegmm.py

@@ -10,9 +10,9 @@ from __future__ import annotations
 
 from typing import Callable
 
-import keras
 import numpy as np
 import tensorflow as tf
+import tf_keras as keras
 from numpy.typing import ArrayLike
 
 from dataeval._internal.detectors.ood.base import OODGMMBase, OODScoreOutput

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/detectors/outliers.py

@@ -147,7 +147,7 @@ class Outliers:
                 mask = _get_outlier_mask(values.astype(np.float64), self.outlier_method, self.outlier_threshold)
                 indices = np.flatnonzero(mask)
                 for i, value in zip(indices, values[mask]):
-                    flagged_images.setdefault(i, {}).update({stat: value})
+                    flagged_images.setdefault(int(i), {}).update({stat: value})
 
         return dict(sorted(flagged_images.items()))
 

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/metrics/stats/datasetstats.py

@@ -162,10 +162,10 @@ def channelstats(
 
     >>> stats = channelstats(images)
     >>> print(stats.visualstats.darkness)
-    [0.02124 0.1213  0.2212  0.1013  0.1076  0.11383 0.2013  0.2076  0.2139
-     0.3013  0.3076  0.3137  0.4014  0.4075  0.4138  0.5015  0.508   0.5137
-     0.6016  0.6074  0.614   0.701   0.7075  0.714   0.8013  0.8076  0.814
-     0.9014  0.9077  0.914  ]
+    [0.07495 0.1748  0.275   0.1047  0.11096 0.1172  0.2047  0.2109  0.2172
+     0.3047  0.311   0.3171  0.4048  0.411   0.4172  0.505   0.5107  0.517
+     0.6045  0.611   0.617   0.7046  0.711   0.7173  0.8047  0.811   0.8174
+     0.905   0.911   0.917  ]
     """
     outputs = run_stats(images, bboxes, True, [PixelStatsProcessor, VisualStatsProcessor])
     return ChannelStatsOutput(*outputs)  # type: ignore

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/metrics/stats/dimensionstats.py

@@ -57,16 +57,16 @@ class DimensionStatsProcessor(StatsProcessor[DimensionStatsOutput]):
     image_function_map = {
         "left": lambda x: x.box[0],
         "top": lambda x: x.box[1],
-        "width": lambda x: x.shape[-1],
-        "height": lambda x: x.shape[-2],
+        "width": lambda x: x.box[2] - x.box[0],
+        "height": lambda x: x.box[3] - x.box[1],
         "channels": lambda x: x.shape[-3],
-        "size": lambda x: np.prod(x.shape[-2:]),
-        "aspect_ratio": lambda x: x.shape[-1] / x.shape[-2],
+        "size": lambda x: (x.box[2] - x.box[0]) * (x.box[3] - x.box[1]),
+        "aspect_ratio": lambda x: (x.box[2] - x.box[0]) / (x.box[3] - x.box[1]),
         "depth": lambda x: get_bitdepth(x.image).depth,
         "center": lambda x: np.asarray([(x.box[0] + x.box[2]) / 2, (x.box[1] + x.box[3]) / 2]),
         "distance": lambda x: np.sqrt(
-            np.square(((x.box[0] + x.box[2]) / 2) - (x.width / 2))
-            + np.square(((x.box[1] + x.box[3]) / 2) - (x.height / 2))
+            np.square(((x.box[0] + x.box[2]) / 2) - (x.shape[-1] / 2))
+            + np.square(((x.box[1] + x.box[3]) / 2) - (x.shape[-2] / 2))
         ),
     }
 

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/metrics/stats/visualstats.py

@@ -22,14 +22,14 @@ class VisualStatsOutput(BaseStatsOutput):
     ----------
     brightness : NDArray[np.float16]
         Brightness of the images
-    sharpness : NDArray[np.float16]
-        Blurriness of the images
     contrast : NDArray[np.float16]
         Image contrast ratio
     darkness : NDArray[np.float16]
         Darkness of the images
     missing : NDArray[np.float16]
         Percentage of the images with missing pixels
+    sharpness : NDArray[np.float16]
+        Sharpness of the images
     zeros : NDArray[np.float16]
         Percentage of the images with zero value pixels
     percentiles : NDArray[np.float16]
@@ -37,10 +37,10 @@ class VisualStatsOutput(BaseStatsOutput):
     """
 
     brightness: NDArray[np.float16]
-    sharpness: NDArray[np.float16]
     contrast: NDArray[np.float16]
     darkness: NDArray[np.float16]
     missing: NDArray[np.float16]
+    sharpness: NDArray[np.float16]
     zeros: NDArray[np.float16]
     percentiles: NDArray[np.float16]
 
@@ -49,25 +49,25 @@ class VisualStatsProcessor(StatsProcessor[VisualStatsOutput]):
     output_class = VisualStatsOutput
     cache_keys = ["percentiles"]
     image_function_map = {
-        "brightness": lambda x: x.get("percentiles")[-2],
-        "sharpness": lambda x: np.std(edge_filter(np.mean(x.image, axis=0))),
+        "brightness": lambda x: x.get("percentiles")[1],
         "contrast": lambda x: np.nan_to_num(
             (np.max(x.get("percentiles")) - np.min(x.get("percentiles"))) / np.mean(x.get("percentiles"))
         ),
-        "darkness": lambda x: x.get("percentiles")[1],
+        "darkness": lambda x: x.get("percentiles")[-2],
         "missing": lambda x: np.count_nonzero(np.isnan(np.sum(x.image, axis=0))) / np.prod(x.shape[-2:]),
+        "sharpness": lambda x: np.std(edge_filter(np.mean(x.image, axis=0))),
         "zeros": lambda x: np.count_nonzero(np.sum(x.image, axis=0) == 0) / np.prod(x.shape[-2:]),
         "percentiles": lambda x: np.nanpercentile(x.scaled, q=QUARTILES),
     }
     channel_function_map = {
-        "brightness": lambda x: x.get("percentiles")[:, -2],
-        "sharpness": lambda x: np.std(np.vectorize(edge_filter, signature="(m,n)->(m,n)")(x.image), axis=(1, 2)),
+        "brightness": lambda x: x.get("percentiles")[:, 1],
         "contrast": lambda x: np.nan_to_num(
             (np.max(x.get("percentiles"), axis=1) - np.min(x.get("percentiles"), axis=1))
             / np.mean(x.get("percentiles"), axis=1)
         ),
-        "darkness": lambda x: x.get("percentiles")[:, 1],
+        "darkness": lambda x: x.get("percentiles")[:, -2],
         "missing": lambda x: np.count_nonzero(np.isnan(x.image), axis=(1, 2)) / np.prod(x.shape[-2:]),
+        "sharpness": lambda x: np.std(np.vectorize(edge_filter, signature="(m,n)->(m,n)")(x.image), axis=(1, 2)),
         "zeros": lambda x: np.count_nonzero(x.image == 0, axis=(1, 2)) / np.prod(x.shape[-2:]),
         "percentiles": lambda x: np.nanpercentile(x.scaled, q=QUARTILES, axis=1).T,
     }
@@ -113,9 +113,10 @@ def visualstats(
 
     >>> results = visualstats(images)
     >>> print(results.brightness)
-    [0.0737 0.607  0.0713 0.1046 0.138  0.1713 0.2046 0.2379 0.2712 0.3047
-     0.338  0.3713 0.4045 0.438  0.4712 0.5044 0.538  0.5713 0.6045 0.638
-     0.6714 0.7046 0.738  0.7715 0.8047 0.838  0.871  0.905  0.938  0.971 ]
+    [0.02246 0.5557  0.06805 0.1014  0.1348  0.1681  0.2014  0.2347  0.268
+     0.3015  0.3347  0.3682  0.4014  0.4348  0.468   0.5015  0.5347  0.568
+     0.6016  0.635   0.668   0.701   0.735   0.768   0.8013  0.835   0.868
+     0.9014  0.9346  0.9683 ]
     >>> print(results.contrast)
     [2.041 1.332 1.293 1.279 1.272 1.268 1.265 1.263 1.261 1.26  1.259 1.258
      1.258 1.257 1.257 1.256 1.256 1.255 1.255 1.255 1.255 1.254 1.254 1.254

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/models/tensorflow/autoencoder.py

@@ -12,9 +12,9 @@ from __future__ import annotations
 
 from typing import Callable, cast
 
-import keras
 import tensorflow as tf
-from keras.layers import (
+import tf_keras as keras
+from tf_keras.layers import (
     Dense,
     Flatten,
     Layer,

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/models/tensorflow/losses.py

@@ -11,11 +11,11 @@ from __future__ import annotations
 from typing import Literal, cast
 
 import tensorflow as tf
-from keras.layers import Flatten
 from numpy.typing import NDArray
 from tensorflow_probability.python.distributions.mvn_diag import MultivariateNormalDiag
 from tensorflow_probability.python.distributions.mvn_tril import MultivariateNormalTriL
 from tensorflow_probability.python.stats import covariance
+from tf_keras.layers import Flatten
 
 from dataeval._internal.models.tensorflow.gmm import gmm_energy, gmm_params
 

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/models/tensorflow/pixelcnn.py

@@ -13,9 +13,9 @@ from __future__ import annotations
 import functools
 import warnings
 
-import keras
 import numpy as np
 import tensorflow as tf
+import tf_keras as keras
 from tensorflow_probability.python.bijectors import bijector
 from tensorflow_probability.python.distributions import (
     categorical,

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/models/tensorflow/trainer.py

@@ -10,9 +10,9 @@ from __future__ import annotations
 
 from typing import Callable, Iterable, cast
 
-import keras
 import numpy as np
 import tensorflow as tf
+import tf_keras as keras
 from numpy.typing import NDArray
 
 

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/_internal/models/tensorflow/utils.py

@@ -11,11 +11,13 @@ from __future__ import annotations
 import math
 from typing import Callable, Union, cast
 
-import keras as keras
 import numpy as np
 import tensorflow as tf
-from keras import Sequential
-from keras.layers import (
+import tf_keras as keras
+from numpy.typing import NDArray
+from tensorflow._api.v2.nn import relu, softmax, tanh
+from tf_keras import Sequential
+from tf_keras.layers import (
     Conv2D,
     Conv2DTranspose,
     Dense,
@@ -23,8 +25,6 @@ from keras.layers import (
     InputLayer,
     Reshape,
 )
-from numpy.typing import NDArray
-from tensorflow._api.v2.nn import relu, softmax, tanh
 
 from dataeval._internal.models.tensorflow.autoencoder import AE, AEGMM, VAE, VAEGMM
 from dataeval._internal.models.tensorflow.pixelcnn import PixelCNN

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/utils/tensorflow/__init__.py

@@ -6,6 +6,12 @@ as well as constructors which allow for customization of the encoder, decoder an
 layers used by the model.
 """
 
+from dataeval import _IS_TENSORFLOW_AVAILABLE
+
 from . import loss, models, recon
 
-__all__ = ["loss", "models", "recon"]
+__all__ = []
+
+
+if _IS_TENSORFLOW_AVAILABLE:
+    __all__ = ["loss", "models", "recon"]

dataeval-0.71.1/src/dataeval/utils/tensorflow/loss/__init__.py

@@ -0,0 +1,7 @@
+from dataeval import _IS_TENSORFLOW_AVAILABLE
+from dataeval._internal.models.tensorflow.losses import Elbo, LossGMM
+
+__all__ = []
+
+if _IS_TENSORFLOW_AVAILABLE:
+    __all__ += ["Elbo", "LossGMM"]

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/utils/tensorflow/models/__init__.py

@@ -1,5 +1,9 @@
+from dataeval import _IS_TENSORFLOW_AVAILABLE
 from dataeval._internal.models.tensorflow.autoencoder import AE, AEGMM, VAE, VAEGMM
 from dataeval._internal.models.tensorflow.pixelcnn import PixelCNN
 from dataeval._internal.models.tensorflow.utils import create_model
 
-__all__ = ["create_model", "AE", "AEGMM", "PixelCNN", "VAE", "VAEGMM"]
+__all__ = []
+
+if _IS_TENSORFLOW_AVAILABLE:
+    __all__ += ["create_model", "AE", "AEGMM", "PixelCNN", "VAE", "VAEGMM"]

{dataeval-0.71.0 → dataeval-0.71.1}/src/dataeval/utils/torch/__init__.py

@@ -5,8 +5,17 @@ While these metrics can take in custom models, DataEval provides utility classes
 to create a seamless integration between custom models and DataEval's metrics.
 """
 
+from dataeval import _IS_TORCH_AVAILABLE, _IS_TORCHVISION_AVAILABLE
 from dataeval._internal.utils import read_dataset
 
-from . import models, trainer
+__all__ = []
 
-__all__ = ["read_dataset", "models", "trainer"]
+if _IS_TORCH_AVAILABLE:
+    from . import models, trainer
+
+    __all__ += ["read_dataset", "models", "trainer"]
+
+if _IS_TORCHVISION_AVAILABLE:
+    from . import datasets
+
+    __all__ += ["datasets"]

dataeval-0.71.1/src/dataeval/utils/torch/datasets/__init__.py

@@ -0,0 +1,12 @@
+"""
+Provide access to common Torch datasets used for computer vision
+"""
+
+from dataeval import _IS_TORCHVISION_AVAILABLE
+
+__all__ = []
+
+if _IS_TORCHVISION_AVAILABLE:
+    from dataeval._internal.datasets import CIFAR10, MNIST, VOCDetection
+
+    __all__ += ["CIFAR10", "MNIST", "VOCDetection"]

dataeval-0.71.1/src/dataeval/utils/torch/models/__init__.py

@@ -0,0 +1,11 @@
+from dataeval import _IS_TORCH_AVAILABLE
+from dataeval._internal.models.pytorch.autoencoder import (
+    AriaAutoencoder,
+    Decoder,
+    Encoder,
+)
+
+__all__ = []
+
+if _IS_TORCH_AVAILABLE:
+    __all__ += ["AriaAutoencoder", "Decoder", "Encoder"]

dataeval-0.71.1/src/dataeval/utils/torch/trainer/__init__.py

@@ -0,0 +1,7 @@
+from dataeval import _IS_TORCH_AVAILABLE
+from dataeval._internal.models.pytorch.autoencoder import AETrainer
+
+__all__ = []
+
+if _IS_TORCH_AVAILABLE:
+    __all__ += ["AETrainer"]

dataeval-0.71.0/src/dataeval/utils/tensorflow/loss/__init__.py

@@ -1,3 +0,0 @@
-from dataeval._internal.models.tensorflow.losses import Elbo, LossGMM
-
-__all__ = ["Elbo", "LossGMM"]

dataeval-0.71.0/src/dataeval/utils/torch/datasets/__init__.py

@@ -1,7 +0,0 @@
-"""
-Provide access to common Torch datasets used for computer vision
-"""
-
-from dataeval._internal.datasets import CIFAR10, MNIST, VOCDetection
-
-__all__ = ["CIFAR10", "MNIST", "VOCDetection"]

dataeval-0.71.0/src/dataeval/utils/torch/models/__init__.py

@@ -1,7 +0,0 @@
-from dataeval._internal.models.pytorch.autoencoder import (
-    AriaAutoencoder,
-    Decoder,
-    Encoder,
-)
-
-__all__ = ["AriaAutoencoder", "Decoder", "Encoder"]

dataeval-0.71.0/src/dataeval/utils/torch/trainer/__init__.py

@@ -1,3 +0,0 @@
-from dataeval._internal.models.pytorch.autoencoder import AETrainer
-
-__all__ = ["AETrainer"]