dataeval 1.0.4__tar.gz → 1.0.6__tar.gz

{dataeval-1.0.4 → dataeval-1.0.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dataeval
-Version: 1.0.4
+Version: 1.0.6
 Summary: DataEval provides a simple interface to characterize image data and its impact on model performance across classification and object-detection tasks
 Project-URL: Homepage, https://dataeval.ai/
 Project-URL: Repository, https://github.com/aria-ml/dataeval/

{dataeval-1.0.4 → dataeval-1.0.6}/pyproject.toml

@@ -114,10 +114,24 @@ docs = [
   "sphinx-tabs>=3.4.7",
   "Sphinx>=7.2.6,<9.0.0", # sphinx-immaterial <= 0.13.9 is not compatible with sphinx >=9.0
   "torchmetrics>=1.0.0",
-  "torchvision>=0.17.0",
   "markupsafe>=3,<3.0.2",
   "jupytext>=1.19.1",
 ]
+security = [  # keep in sync with [tool.uv.constraint-dependencies]
+  "cryptography>=46.0.5",    # CVE-2026-26007: Missing Subgroup Validation for SECT Curves
+  "filelock>=3.20.3",        # GHSA-w853-jp5j-5j7f, GHSA-qmgc-5h2g-mvrw
+  "onnx>=1.21.0",            # CVE-2026-28500: Untrusted Model Repository Warnings Suppressed by silent=True
+                             # CVE-2026-34445: Malicious ONNX models can crash servers by exploiting unprotected object settings
+                             # CVE-2026-27489: Vulnerable to Path Traversal via Symlink
+                             # GHSA-q56x-g2fj-4rj6: TOCTOU arbitrary file read/write in save_external_dat
+  "pillow>=12.1.1",          # CVE-2026-25990: OOB write via PSD image
+  "poetry>=2.3.3",           # CVE-2026-34591: Poetry Has Wheel Path Traversal Which Can Lead to Arbitrary File Write
+  "protobuf>=6.33.5",        # GHSA-7gcm-g887-7qv7
+  "setuptools>=82.0.0",      # CVE-2026-23949: (jaraco_context) path traversal in tarball()
+                             # CVE-2026-24049: (wheel) privilege escalation via unpack
+  "tornado>=6.5.5",          # CVE-2026-31958: Tornado is vulnerable to DoS due to too many multipart parts
+                             # CVE-2026-35536: Tornado has cookie attribute injection via .RequestHandler.set_cookie
+]
 dev = [
   { include-group = "base" },
   { include-group = "lint" },
@@ -141,9 +155,18 @@ conflicts = [
 ]
 constraint-dependencies = [
     "cryptography>=46.0.5",    # CVE-2026-26007: Missing Subgroup Validation for SECT Curves
+    "filelock>=3.20.3",        # GHSA-w853-jp5j-5j7f, GHSA-qmgc-5h2g-mvrw
+    "onnx>=1.21.0",            # CVE-2026-28500: Untrusted Model Repository Warnings Suppressed by silent=True
+                               # CVE-2026-34445: Malicious ONNX models can crash servers by exploiting unprotected object settings
+                               # CVE-2026-27489: Vulnerable to Path Traversal via Symlink
+                               # GHSA-q56x-g2fj-4rj6: TOCTOU arbitrary file read/write in save_external_dat
     "pillow>=12.1.1",          # CVE-2026-25990: OOB write via PSD image
+    "poetry>=2.3.3",           # CVE-2026-34591: Poetry Has Wheel Path Traversal Which Can Lead to Arbitrary File Write
+    "protobuf>=6.33.5",        # GHSA-7gcm-g887-7qv7
     "setuptools>=82.0.0",      # CVE-2026-23949: (jaraco_context) path traversal in tarball()
                                # CVE-2026-24049: (wheel) privilege escalation via unpack
+    "tornado>=6.5.5",          # CVE-2026-31958: Tornado is vulnerable to DoS due to too many multipart parts
+                               # CVE-2026-35536: Tornado has cookie attribute injection via .RequestHandler.set_cookie
 ]
 
 [[tool.uv.index]]
@@ -201,6 +224,9 @@ version-file = "src/dataeval/_version.py"
 [tool.poetry]
 version = "0.0.0"  # unused
 
+[tool.poetry.dependencies]
+python = ">=3.10,<3.15"
+
 [tool.pyproject2conda.dependencies]
 numpy = { skip = true, packages = "numpy>=1.24.2" }
 scikit-learn = { skip = true, packages = "scikit-learn>=1.5.0" }
@@ -262,20 +288,23 @@ exclude = [
   ".jupyter_cache",
   "*env*",
   "output",
+  "_build",
   "build",
   ".nox",
   ".tox",
+  "prototype",
   "src/dataeval/_version.py",
 ]
 line-length = 120
 indent-width = 4
 target-version = "py310"
+extend-include = ["*.ipynb"]
 
 [tool.ruff.lint]
 select = ["F", "E", "W", "C90", "I", "N", "D", "UP", "YTT", "ANN", "S", "BLE", "B", "A",
           "COM", "C4", "T10", "ISC", "ICN", "PYI", "PT", "Q", "RSE", "RET", "SLF", "SIM",
-          "TID252", "ARG", "FIX", "PD", "FLY", "NPY", "RUF100", "PERF"]
-ignore = ["ANN401", "COM812", "NPY002", "SLF001"]
+          "TID252", "ARG", "FIX", "PD", "FLY", "NPY", "RUF027", "RUF100", "PERF"]
+ignore = ["ANN101", "ANN102", "ANN401", "C408", "C416", "COM812", "NPY002", "SLF001"]
 fixable = ["ALL"]
 unfixable = []
 dummy-variable-rgx = "^(_+|(_+[a-zA-Z0-9_]*[a-zA-Z0-9]+?))$"
@@ -287,10 +316,14 @@ builtins-strict-checking = false
 [tool.ruff.lint.isort]
 known-first-party = ["dataeval"]
 
+[tool.ruff.lint.mccabe]
+max-complexity = 5
+
 [tool.ruff.lint.pydocstyle]
 convention = "numpy"
 
 [tool.ruff.format]
+preview = true
 quote-style = "double"
 indent-style = "space"
 skip-magic-trailing-comma = false

{dataeval-1.0.4 → dataeval-1.0.6}/src/dataeval/_embeddings.py

@@ -509,7 +509,7 @@ class Embeddings(Array, FeatureExtractor):
             images.append(image)
         return images
 
-    def _batch(self, indices: Sequence[int]) -> Iterator[NDArray[Any]]:
+    def _batch(self, indices: Sequence[int]) -> Iterator[NDArray[Any]]:  # noqa: C901
         """Process indices in batches using the extractor."""
         if self._dataset is None:
             raise NotFittedError("No dataset bound. Call bind() first.")
@@ -559,7 +559,7 @@ class Embeddings(Array, FeatureExtractor):
             batch_indices = list(indices[batch_start : batch_start + self._batch_size])
             yield self._embeddings[batch_indices]
 
-    def __getitem__(self, key: int | Iterable[int] | slice, /) -> NDArray[Any]:
+    def __getitem__(self, key: int | Iterable[int] | slice, /) -> NDArray[Any]:  # noqa: C901
         """
         Access embeddings by index, indices or slice.
 

{dataeval-1.0.4 → dataeval-1.0.6}/src/dataeval/_experimental.py

@@ -13,7 +13,7 @@ from dataeval.exceptions import DeprecatedWarning, ExperimentalWarning
 F = TypeVar("F", bound=Callable[..., Any])
 
 
-def _make_warning_message(
+def _make_warning_message(  # noqa: C901
     name: str,
     kind: str,
     *,
@@ -51,7 +51,7 @@ def _prepend_doc_note(doc: str | None, note: str) -> str:
 def experimental(_target: F) -> F: ...
 @overload
 def experimental(*, alternative: str | None = None, details: str | None = None) -> Callable[[F], F]: ...
-def experimental(
+def experimental(  # noqa: C901
     _target: F | None = None,
     *,
     alternative: str | None = None,
@@ -72,7 +72,7 @@ def experimental(
         def my_func(): ...
     """
 
-    def decorator(target: F) -> F:
+    def decorator(target: F) -> F:  # noqa: C901
         name = getattr(target, "__qualname__", getattr(target, "__name__", str(target)))
         msg = _make_warning_message(name, "experimental", alternative=alternative, details=details)
         warned = False
@@ -118,7 +118,7 @@ def deprecated(
     alternative: str | None = None,
     details: str | None = None,
 ) -> Callable[[F], F]: ...
-def deprecated(
+def deprecated(  # noqa: C901
     _target: F | None = None,
     *,
     since: str | None = None,
@@ -141,7 +141,7 @@ def deprecated(
         def old_func(): ...
     """
 
-    def decorator(target: F) -> F:
+    def decorator(target: F) -> F:  # noqa: C901
         name = getattr(target, "__qualname__", getattr(target, "__name__", str(target)))
         msg = _make_warning_message(
             name,

{dataeval-1.0.4 → dataeval-1.0.6}/src/dataeval/_metadata.py

@@ -282,7 +282,7 @@ class Metadata(Array, FeatureExtractor):
             raise NotFittedError("No dataset bound. Call bind() first.")
         yield from self.factor_data
 
-    def __getitem__(self, index: int | str | slice) -> Array:
+    def __getitem__(self, index: int | str | slice) -> Array:  # noqa: C901
         """Get binned metadata for specific indices or factors.
 
         Parameters
@@ -650,7 +650,7 @@ class Metadata(Array, FeatureExtractor):
         -------
         Sequence[str]
             List of factor names that passed filtering and preprocessing steps.
-            Order matches columns in factor_data and binned_data.
+            Order matches columns in factor_data.
 
         Notes
         -----
@@ -934,7 +934,7 @@ class Metadata(Array, FeatureExtractor):
         factor = factor[0] if isinstance(factor, tuple) else factor
         return factor in self.include if self.include else factor not in self.exclude
 
-    def _reset_bins(self, cols: Iterable[str] | None = None) -> None:
+    def _reset_bins(self, cols: Iterable[str] | None = None) -> None:  # noqa: C901
         if self._is_binned:
             columns = self._dataframe.columns
             for col in cols or columns:
@@ -1006,7 +1006,7 @@ class Metadata(Array, FeatureExtractor):
             )
         return target_rows
 
-    def _get_target_factor_values(
+    def _get_target_factor_values(  # noqa: C901
         self,
         factor_name: str,
         factor_values: Any,
@@ -1152,7 +1152,7 @@ class Metadata(Array, FeatureExtractor):
             self._structure()
         return bool(self._has_targets)
 
-    def _process_targets(
+    def _process_targets(  # noqa: C901
         self,
         raw: list,
         labels: list,
@@ -1284,7 +1284,7 @@ class Metadata(Array, FeatureExtractor):
         existing = self._factors if hasattr(self, "_factors") else {}
         self._factors = {k: existing.get(k) for k in usable_factors}
 
-    def _structure(
+    def _structure(  # noqa: C901
         self,
         *,
         progress_callback: ProgressCallback | None = None,
@@ -1478,7 +1478,7 @@ class Metadata(Array, FeatureExtractor):
         df = self._add_column_with_padding(df, col_dg, ordinal.astype(np.int64), is_od)
         return df, FactorInfo("discrete", is_digitized=True)
 
-    def _bin(
+    def _bin(  # noqa: C901
         self,
         *,
         progress_callback: ProgressCallback | None = None,
@@ -1523,7 +1523,7 @@ class Metadata(Array, FeatureExtractor):
         self._factors.update(factor_info)
         self._is_binned = True
 
-    def add_factors(
+    def add_factors(  # noqa: C901
         self,
         factors: Mapping[str, Array1D[Any]],
         level: Literal["image", "target", "auto"] = "auto",

dataeval-1.0.6/src/dataeval/_version.py

@@ -0,0 +1,24 @@
+# file generated by vcs-versioning
+# don't change, don't track in version control
+from __future__ import annotations
+
+__all__ = [
+    "__version__",
+    "__version_tuple__",
+    "version",
+    "version_tuple",
+    "__commit_id__",
+    "commit_id",
+]
+
+version: str
+__version__: str
+__version_tuple__: tuple[int | str, ...]
+version_tuple: tuple[int | str, ...]
+commit_id: str | None
+__commit_id__: str | None
+
+__version__ = version = '1.0.6'
+__version_tuple__ = version_tuple = (1, 0, 6)
+
+__commit_id__ = commit_id = None

{dataeval-1.0.4 → dataeval-1.0.6}/src/dataeval/bias/_balance.py

@@ -22,28 +22,30 @@ class BalanceOutput(DictOutput):
     """
     Output class for the :class:`.Balance` :term:`bias<Bias>` evaluator.
 
-    Contains three polars DataFrames with mutual information scores and threshold flags.
+    Contains three polars DataFrames with normalized mutual information scores and threshold flags.
 
     Attributes
     ----------
     balance : pl.DataFrame
-        DataFrame with global class-to-factor mutual information:
+        DataFrame with global class-to-factor normalized mutual information:
 
-        - factor_name: str - Name of the metadata factor
-        - mi_value: float - Mutual information value between this factor and class labels
+        - factor_name: str - Name of the metadata factor. Includes "class_label"
+          which represents the self-information (always 1.0).
+        - mi_value: float - Normalized mutual information value between this
+          factor and class labels
     factors : pl.DataFrame
-        DataFrame with inter-factor mutual information correlations:
+        DataFrame with inter-factor normalized mutual information correlations:
 
         - factor1: str - Name of the first factor
         - factor2: str - Name of the second factor
-        - mi_value: float - Mutual information value
+        - mi_value: float - Normalized mutual information value
         - is_correlated: bool - True if mi_value > factor_correlation_threshold
     classwise : pl.DataFrame
-        DataFrame with per-class-to-factor mutual information:
+        DataFrame with per-class-to-factor normalized mutual information:
 
         - class_name: str - Name of the class
         - factor_name: str - Name of the metadata factor
-        - mi_value: float - Mutual information value
+        - mi_value: float - Normalized mutual information value
         - is_imbalanced: bool - True if mi_value > class_imbalance_threshold
     """
 
@@ -58,21 +60,21 @@ class BalanceOutput(DictOutput):
 
 class Balance(Evaluator):
     """
-    Computes mutual information (MI) between factors (class label, metadata, label/image properties).
+    Computes normalized mutual information (NMI) between factors (class label, metadata, label/image properties).
 
     Identifies imbalanced classes and highly correlated metadata factors based on
-    mutual information thresholds.
+    NMI thresholds.
 
     Parameters
     ----------
     num_neighbors : int, default 5
         Number of points to consider as neighbors
     class_imbalance_threshold : float, default 0.3
-        Threshold for identifying imbalanced classes. Classes with MI above this
+        Threshold for identifying imbalanced classes. Classes with NMI above this
         threshold with any metadata factor are considered imbalanced.
     factor_correlation_threshold : float, default 0.5
         Threshold for identifying highly correlated metadata factors. Factor pairs
-        with MI above this threshold are considered highly correlated.
+        with NMI above this threshold are considered highly correlated.
 
     Attributes
     ----------
@@ -89,7 +91,8 @@ class Balance(Evaluator):
     -----
     We use `mutual_info_classif` from sklearn since class label is categorical.
     `mutual_info_classif` outputs are consistent up to O(1e-4) and depend on a random
-    seed. MI is computed differently for categorical and continuous variables.
+    seed. MI is computed differently for categorical and continuous variables, and
+    in all cases normalized or transformed to [0, 1] prior to being returned.
 
     Examples
     --------
@@ -147,9 +150,9 @@ class Balance(Evaluator):
         super().__init__(locals())
 
     @set_metadata(state=["num_neighbors", "class_imbalance_threshold", "factor_correlation_threshold"])
-    def evaluate(self, data: AnnotatedDataset[Any] | MetadataLike) -> BalanceOutput:
+    def evaluate(self, data: AnnotatedDataset[Any] | MetadataLike) -> BalanceOutput:  # noqa: C901
         """
-        Compute mutual information between factors and identify imbalanced classes.
+        Compute normalized mutual information between factors and identify imbalanced classes.
 
         Parameters
         ----------
@@ -160,7 +163,7 @@ class Balance(Evaluator):
         Returns
         -------
         BalanceOutput
-            Three DataFrames containing MI scores and threshold flags:
+            Three DataFrames containing NMI scores and threshold flags:
 
             - balance: Global class-to-factor mutual information
             - factors: Inter-factor mutual information
@@ -168,7 +171,7 @@ class Balance(Evaluator):
 
         Example
         -------
-        Return balance (mutual information) of factors with class_labels
+        Return balance (NMI) of factors with class_labels
 
         >>> from dataeval import Metadata
         >>> metadata = Metadata(dataset)

{dataeval-1.0.4 → dataeval-1.0.6}/src/dataeval/bias/_diversity.py

@@ -56,7 +56,7 @@ class Diversity(Evaluator):
     Through standard histogram binning, for continuous variables.
 
     The method specified defines diversity as the inverse Simpson diversity index linearly rescaled to
-    the unit interval, or the normalized form of the Shannon entropy.
+    the unit interval [0, 1], or the normalized form of the Shannon entropy.
 
     diversity = 1 implies that samples are evenly distributed across a particular factor
     diversity = 0 implies that all samples belong to one category/bin
@@ -66,7 +66,9 @@ class Diversity(Evaluator):
     Parameters
     ----------
     method : "simpson" or "shannon", default "simpson"
-        The methodology used for defining diversity
+        The methodology used for defining diversity. When "simpson" is used,
+        the index is linearly rescaled so that 1.0 represents maximum diversity
+        (even distribution) and 0.0 represents minimum diversity (all samples in one bin).
     threshold : float, default 0.5
         Threshold for identifying low diversity. Factors with diversity values
         at or below this threshold are flagged as having low diversity.
@@ -135,7 +137,7 @@ class Diversity(Evaluator):
         super().__init__(locals())
 
     @set_metadata(state=["method", "threshold"])
-    def evaluate(self, data: AnnotatedDataset[Any] | MetadataLike) -> DiversityOutput:
+    def evaluate(self, data: AnnotatedDataset[Any] | MetadataLike) -> DiversityOutput:  # noqa: C901
         """
         Compute diversity and classwise diversity for the dataset.
 

{dataeval-1.0.4 → dataeval-1.0.6}/src/dataeval/bias/_parity.py

@@ -118,8 +118,6 @@ class Parity(Evaluator):
 
     >>> config = Parity.Config(score_threshold=0.4, p_value_threshold=0.01)
     >>> parity = Parity(config=config)
-
-    output = parity(metadata.binned_data, metadata.class_labels.tolist())
     """
 
     class Config(EvaluatorConfig):

{dataeval-1.0.4 → dataeval-1.0.6}/src/dataeval/config.py

@@ -23,7 +23,7 @@ from pydantic import BaseModel, ConfigDict, field_validator
 
 from dataeval.protocols import DeviceLike
 
-### GLOBAL CONFIG ###
+# GLOBAL CONFIG ###
 
 
 class GlobalConfig(BaseModel):
@@ -77,7 +77,7 @@ class GlobalConfig(BaseModel):
 _config = GlobalConfig()
 
 
-### CONTEXT MANAGER ###
+# CONTEXT MANAGER ###
 
 
 class _ConfigContextManager:
@@ -96,7 +96,7 @@ class _ConfigContextManager:
         setattr(_config, self._attr_name, self._old)
 
 
-### FUNCS ###
+# FUNCS ###
 
 
 def _todevice(device: DeviceLike) -> torch.device:

{dataeval-1.0.4 → dataeval-1.0.6}/src/dataeval/core/_ber.py

@@ -78,6 +78,8 @@ def ber_mst(embeddings: ArrayND[float], class_labels: Array1D[int]) -> BERResult
     """
     Estimate Multi-class :term:`Bayes error rate<Bayes Error Rate (BER)>` using a minimum spanning tree.
 
+    BER bounds the irreducible classification error given the current feature
+    representation — the error attributable to class overlap in embedding space.
     Uses FR with a minimum spanning tree (MST) test statistic basis.
 
     Parameters
@@ -137,7 +139,13 @@ def ber_knn(embeddings: ArrayND[float], class_labels: Array1D[int], k: int) -> B
     """
     Estimate Multi-class :term:`Bayes error rate<Bayes Error Rate (BER)>` using KNN.
 
-    Uses KNN test statistic basis.
+    BER bounds the irreducible classification error given the current feature
+    representation — the error attributable to class overlap in embedding space.
+    Uses KNN test statistic basis. The estimator's behavior depends on the value of k:
+    - k=1: Uses 1-NN for the lower bound and 2-NN for the upper bound.
+    - k=2: Uses 2-NN for the lower bound and 3-NN for the upper bound.
+    - 2<k<=5: Uses k-NN for the lower bound and (k+1)-NN for the upper bound.
+    - k>5: Only available for binary classification; uses k-NN for both bounds with specialized asymptotic weights.
 
     Parameters
     ----------
@@ -146,7 +154,7 @@ def ber_knn(embeddings: ArrayND[float], class_labels: Array1D[int], k: int) -> B
     class_labels : Array1D[int]
         Array of class labels for each image. Can be a 1D list, or array-like object.
     k : int
-        Number of nearest neighbors for KNN estimator
+        Number of nearest neighbors for KNN estimator. Should be between 1 and the number of samples.
 
     Returns
     -------

{dataeval-1.0.4 → dataeval-1.0.6}/src/dataeval/core/_bin.py

@@ -92,7 +92,7 @@ def bin_data(data: NDArray[Any], bin_method: str) -> NDArray[np.intp]:
     return np.digitize(data, bin_edges)
 
 
-def is_continuous(data: NDArray[np.number[Any]], image_indices: NDArray[np.number[Any]] | None = None) -> bool:
+def is_continuous(data: NDArray[np.number[Any]], image_indices: NDArray[np.number[Any]] | None = None) -> bool:  # noqa: C901
     """
     Determine whether the data is continuous or discrete using the Wasserstein distance.
 
@@ -144,7 +144,7 @@ def is_continuous(data: NDArray[np.number[Any]], image_indices: NDArray[np.numbe
     return bool(shift < DISCRETE_MIN_WD)  # if NNN is close enough to uniform, consider the sample continuous.
 
 
-def _bin_by_clusters(data: NDArray[np.number[Any]]) -> NDArray[np.float64]:
+def _bin_by_clusters(data: NDArray[np.number[Any]]) -> NDArray[np.float64]:  # noqa: C901
     """
     Bin continuous data by using the Clusterer to identify clusters.
 

{dataeval-1.0.4 → dataeval-1.0.6}/src/dataeval/core/_clusterer.py

@@ -99,7 +99,7 @@ class _Clusters:
         prob: NDArray[np.float64] = exp / np.sum(exp)
         return prob
 
-    def _sort_by_weights(self, embeddings: NDArray[np.float64]) -> NDArray[np.intp]:
+    def _sort_by_weights(self, embeddings: NDArray[np.float64]) -> NDArray[np.intp]:  # noqa: C901
         """Sort samples using complexity-based weighted sampling."""
         labels = self._get_labels(embeddings)
         pr = self._complexity(embeddings)
@@ -241,7 +241,7 @@ class _HDBSCANSorter:
         n_samples_per_cluster = np.bincount(labels)
         _logger.debug(
             "HDBSCAN clustering complete: %d clusters, samples per cluster: min=%d, max=%d, mean=%.1f",
-            clst.unique_clusters,
+            len(clst.unique_clusters),
             np.min(n_samples_per_cluster),
             np.max(n_samples_per_cluster),
             np.mean(n_samples_per_cluster),
@@ -356,7 +356,7 @@ class _HDBSCAN:
         self.cluster_selection_epsilon = 0.0
         self.cluster_selection_method = "eom"
 
-    def fit(self, embeddings: NDArray[np.floating]) -> "_HDBSCAN":
+    def fit(self, embeddings: NDArray[np.floating]) -> "_HDBSCAN":  # noqa: C901
         """
         Find clusters based on hierarchical density-based clustering.
 
@@ -541,7 +541,7 @@ class ClusterStats(TypedDict):
     nearest_cluster_idx: NDArray[np.int64]
 
 
-def compute_cluster_stats(
+def compute_cluster_stats(  # noqa: C901
     embeddings: NDArray[np.floating],
     cluster_labels: _Clusters | NDArray[np.int64],
 ) -> ClusterStats:
@@ -642,7 +642,7 @@ def compute_cluster_stats(
     )
 
 
-def cluster(
+def cluster(  # noqa: C901
     embeddings: ArrayND[float],
     algorithm: Literal["kmeans", "hdbscan"] = "hdbscan",
     n_clusters: int | None = None,

{dataeval-1.0.4 → dataeval-1.0.6}/src/dataeval/core/_compute_ratios.py

@@ -101,7 +101,7 @@ def _build_image_lookup(source_indices: Sequence[SourceIndex]) -> dict[tuple[int
     return lookup
 
 
-def _calculate_ratio_for_stat(
+def _calculate_ratio_for_stat(  # noqa: C901
     stat_name: str,
     box_value: Any,
     img_value: Any,
@@ -160,7 +160,7 @@ def _calculate_ratio_for_stat(
         return box_value
 
 
-def _validate_separate_inputs(
+def _validate_separate_inputs(  # noqa: C901
     stats_output: StatsResult,
     box_stats_output: StatsResult,
 ) -> tuple[Sequence[SourceIndex], Sequence[SourceIndex]]:
@@ -241,7 +241,7 @@ def _validate_unified_input(source_indices: Sequence[SourceIndex]) -> None:
         )
 
 
-def compute_ratios(
+def compute_ratios(  # noqa: C901
     stats_output: StatsResult,
     *,
     target_stats_output: StatsResult | None = None,

{dataeval-1.0.4 → dataeval-1.0.6}/src/dataeval/core/_compute_stats.py

@@ -158,7 +158,7 @@ def _determine_channel_indices(calculator_output: list[dict[str, list[Any]]], nu
     return sorted(channel_indices_needed, key=lambda x: -1 if x is None else x)
 
 
-def _reconcile_stats(
+def _reconcile_stats(  # noqa: C901
     calculator_output: list[dict[str, list[Any]]],
     sorted_channels: list[int | None],
     empty_values_map: dict[str, Any],
@@ -344,7 +344,7 @@ def _aggregate_batch(
 _UNSET = object()
 
 
-def compute_stats(
+def compute_stats(  # noqa: C901
     data: Iterable[ArrayLike] | Dataset[ArrayLike] | Dataset[tuple[ArrayLike, Any, Any]],
     *,
     boxes: Iterable[Iterable[BoxLike] | None] | None = None,
@@ -544,7 +544,7 @@ def compute_stats(
     )
 
 
-def combine_stats_results(
+def combine_stats_results(  # noqa: C901
     results: StatsResult | Sequence[StatsResult],
 ) -> tuple[StatsMap, list[SourceIndex], list[int]]:
     """Combine one or more StatsResults into unified stats, source_index, and dataset_steps.