dataeval 0.69.2__py3-none-any.whl → 0.69.3__py3-none-any.whl

dataeval/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.69.2"
+__version__ = "0.69.3"
 
 from importlib.util import find_spec
 

dataeval/_internal/metrics/diversity.py

@@ -130,7 +130,10 @@ def diversity_simpson(
         p_i = cnts / cnts.sum()
         # inverse Simpson index normalized by (number of bins)
         s_0 = 1 / np.sum(p_i**2) / num_bins[col]
-        ev_index[col] = (s_0 * num_bins[col] - 1) / (num_bins[col] - 1)
+        if num_bins[col] == 1:
+            ev_index[col] = 0
+        else:
+            ev_index[col] = (s_0 * num_bins[col] - 1) / (num_bins[col] - 1)
     return ev_index
 
 

dataeval/_internal/metrics/parity.py

@@ -348,6 +348,7 @@ def parity(
     chi_scores = np.zeros(len(factors))
     p_values = np.zeros(len(factors))
     n_cls = len(np.unique(labels))
+    not_enough_data = {}
     for i, (current_factor_name, factor_values) in enumerate(factors.items()):
         unique_factor_values = np.unique(factor_values)
         contingency_matrix = np.zeros((len(unique_factor_values), n_cls))
@@ -361,13 +362,12 @@ def parity(
                 with_both = np.bitwise_and((labels == label), factor_values == factor_value)
                 contingency_matrix[fi, label] = np.sum(with_both)
                 if 0 < contingency_matrix[fi, label] < 5:
-                    warnings.warn(
-                        f"Factor {current_factor_name} value {factor_value} co-occurs "
-                        f"only {contingency_matrix[fi, label]} times with label {label}. "
-                        "This can cause inaccurate chi_square calculation. Recommend"
-                        "ensuring each label occurs either 0 times or at least 5 times. "
-                        "Alternatively, digitize any continuous-valued factors "
-                        "into fewer bins."
+                    if current_factor_name not in not_enough_data:
+                        not_enough_data[current_factor_name] = {}
+                    if factor_value not in not_enough_data[current_factor_name]:
+                        not_enough_data[current_factor_name][factor_value] = []
+                    not_enough_data[current_factor_name][factor_value].append(
+                        (label, int(contingency_matrix[fi, label]))
                     )
 
         # This deletes rows containing only zeros,
@@ -381,4 +381,23 @@ def parity(
         chi_scores[i] = chi2
         p_values[i] = p
 
+    if not_enough_data:
+        factor_msg = []
+        for factor, fact_dict in not_enough_data.items():
+            stacked_msg = []
+            for key, value in fact_dict.items():
+                msg = []
+                for item in value:
+                    msg.append(f"label {item[0]}: {item[1]} occurrences")
+                flat_msg = "\n\t\t".join(msg)
+                stacked_msg.append(f"value {key} - {flat_msg}\n\t")
+            factor_msg.append(factor + " - " + "".join(stacked_msg))
+
+        message = "\n".join(factor_msg)
+
+        warnings.warn(
+            f"The following factors did not meet the recommended 5 occurrences for each value-label combination. \nRecommend rerunning parity after adjusting the following factor-value-label combinations: \n{message}",  # noqa: E501
+            UserWarning,
+        )
+
     return ParityOutput(chi_scores, p_values)

{dataeval-0.69.2.dist-info → dataeval-0.69.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: dataeval
-Version: 0.69.2
+Version: 0.69.3
 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

{dataeval-0.69.2.dist-info → dataeval-0.69.3.dist-info}/RECORD

@@ -1,4 +1,4 @@
-dataeval/__init__.py,sha256=NUQixSNyEc0GiI7YgbfY9IL0OEkIN9kdbrOGAB041Ig,590
+dataeval/__init__.py,sha256=4JtJRUfhO_kYbjWDhzY5niIvmLb8K_3sCL-wbcZ_mUU,590
 dataeval/_internal/detectors/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 dataeval/_internal/detectors/clusterer.py,sha256=hJwELUeAdZZ3OVLIfwalw2P7Zz13q2ZqrV6gx90s44E,20695
 dataeval/_internal/detectors/drift/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -25,8 +25,8 @@ dataeval/_internal/metrics/balance.py,sha256=eAHvgjiGCH893XSQLqh9j9wgvAECoNPVT8k
 dataeval/_internal/metrics/ber.py,sha256=Onsi47AbT9rMvng-Pbu8LIrYRfLpI13En1FxkFoMKQs,4668
 dataeval/_internal/metrics/coverage.py,sha256=EZVES1rbZW2j_CtQv1VFfSO-UmWcrt5nmqxDErtrG14,3473
 dataeval/_internal/metrics/divergence.py,sha256=nmMUfr9FGnH798eb6xzEiMj4C42rQVthh5HeexiY6EE,4119
-dataeval/_internal/metrics/diversity.py,sha256=nGjYQ-NLjb8mPt1PAYnvkWH4D58kjM39IPs2FULfis4,7503
-dataeval/_internal/metrics/parity.py,sha256=suv1Pf7gPj0_NxsS0_M6ewfUndsFJyEhbt5NPp6ktMI,15457
+dataeval/_internal/metrics/diversity.py,sha256=_oT0FHsgfLOoe_TLD2Aax4r4jmH6WnOPVIkcl_YjaoY,7582
+dataeval/_internal/metrics/parity.py,sha256=VszQNbHWjct2bCqrIXUZC_qFi4ZIq2Lm-vs-DiarBFo,16244
 dataeval/_internal/metrics/stats.py,sha256=ILKteVMGjrp1s2CECPL_hbLsijIKR2d6II2-8w9oxW8,18105
 dataeval/_internal/metrics/uap.py,sha256=w-wvXXnX16kUq-weaZD2SrJi22LJ8EjOFbOhPxeGejI,2043
 dataeval/_internal/metrics/utils.py,sha256=mSYa-3cHGcsQwPr7zbdpzrnK_8jIXCiAcu2HCcvrtaY,13007
@@ -67,7 +67,7 @@ dataeval/torch/models/__init__.py,sha256=YnDnePYpRIKHyYn3F5qR1OObMSb-g0FGvI8X-uT
 dataeval/torch/trainer/__init__.py,sha256=Te-qElt8h-Zv8NN0r-VJOEdCPHTQ2yO3rd2MhRiZGZs,93
 dataeval/utils/__init__.py,sha256=ExQ1xj62MjcM9uIu1-g1P2fW0EPJpcIofnvxjQ908c4,172
 dataeval/workflows/__init__.py,sha256=gkU2B6yUiefexcYrBwqfZKNl8BvX8abUjfeNvVBXF4E,186
-dataeval-0.69.2.dist-info/LICENSE.txt,sha256=Kpzcfobf1HlqafF-EX6dQLw9TlJiaJzfgvLQFukyXYw,1060
-dataeval-0.69.2.dist-info/METADATA,sha256=_9rVrbIh4EPYStZtOUYnB-Xo3cZ5JMUAf06TqDKvrZs,4217
-dataeval-0.69.2.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
-dataeval-0.69.2.dist-info/RECORD,,
+dataeval-0.69.3.dist-info/LICENSE.txt,sha256=Kpzcfobf1HlqafF-EX6dQLw9TlJiaJzfgvLQFukyXYw,1060
+dataeval-0.69.3.dist-info/METADATA,sha256=dyyl60cjz6n7gRgYMZs9gCOdqpc9UbSV4LFCD8rJNCM,4217
+dataeval-0.69.3.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+dataeval-0.69.3.dist-info/RECORD,,