dataeval 1.0.4__tar.gz → 1.0.5__tar.gz

{dataeval-1.0.4 → dataeval-1.0.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dataeval
-Version: 1.0.4
+Version: 1.0.5
 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-1.0.5}/pyproject.toml

@@ -118,6 +118,14 @@ docs = [
   "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
+  "pillow>=12.1.1",          # CVE-2026-25990: OOB write via PSD image
+  "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
+]
 dev = [
   { include-group = "base" },
   { include-group = "lint" },
@@ -141,7 +149,9 @@ 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
     "pillow>=12.1.1",          # CVE-2026-25990: OOB write via PSD image
+    "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
 ]
@@ -262,20 +272,23 @@ 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]+?))$"
@@ -287,6 +300,9 @@ 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"
 

{dataeval-1.0.4 → dataeval-1.0.5}/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-1.0.5}/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-1.0.5}/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
@@ -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-1.0.5/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 = '1.0.5'
+__version_tuple__ = version_tuple = (1, 0, 5)
+
+__commit_id__ = commit_id = None

{dataeval-1.0.4 → dataeval-1.0.5}/src/dataeval/bias/_balance.py

@@ -147,7 +147,7 @@ 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.
 

{dataeval-1.0.4 → dataeval-1.0.5}/src/dataeval/bias/_diversity.py

@@ -135,7 +135,7 @@ 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-1.0.5}/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-1.0.5}/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-1.0.5}/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)
@@ -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-1.0.5}/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-1.0.5}/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.

{dataeval-1.0.4 → dataeval-1.0.5}/src/dataeval/core/_fast_hdbscan/_cluster_trees.py

@@ -315,9 +315,9 @@ def cluster_tree_from_condensed_tree_bcubed(condensed_tree, cluster_tree, label_
     mask1 = condensed_tree.child_size > 1
     mask2 = condensed_tree.child_size == 1
     mask3 = np.array([child in label_indices_list for child in condensed_tree.child])
-    mask4 = np.array(
-        [parent in cluster_tree_parents for parent in condensed_tree.parent]
-    )  # check that it's not a leaf cluster
+    mask4 = np.array([
+        parent in cluster_tree_parents for parent in condensed_tree.parent
+    ])  # check that it's not a leaf cluster
 
     mask = mask1 | (mask2 & mask3 & mask4)
 
@@ -501,9 +501,9 @@ def extract_clusters_bcubed(
                 unselected_nodes,
             )
 
-    return np.asarray(
-        [node for node, selected in selected_clusters.items() if (selected and (node not in virtual_nodes))]
-    )
+    return np.asarray([
+        node for node, selected in selected_clusters.items() if (selected and (node not in virtual_nodes))
+    ])
 
 
 @numba.njit(cache=True)

{dataeval-1.0.4 → dataeval-1.0.5}/src/dataeval/core/_fast_hdbscan/_mst.py

@@ -80,7 +80,7 @@ def _expand_tree(
 
 
 @numba.njit(locals={"i": numba.types.intp, "nbr": numba.types.intp, "dist": numba.types.float32}, cache=True)
-def _update_tree(
+def _update_tree(  # noqa: C901
     tree: NDArray[np.float32],
     total_edge: int,
     disjoint_set: tuple[NDArray[np.int64], NDArray[np.int64]],

{dataeval-1.0.4 → dataeval-1.0.5}/src/dataeval/core/_label_stats.py

@@ -54,7 +54,7 @@ class LabelStatsResult(TypedDict):
     empty_image_count: int
 
 
-def label_stats(
+def label_stats(  # noqa: C901
     class_labels: Iterable[int],
     item_indices: Iterable[int] | None = None,
     index2label: Mapping[int, str] | None = None,

{dataeval-1.0.4 → dataeval-1.0.5}/src/dataeval/core/_metadata_insights.py

@@ -68,7 +68,7 @@ def _calc_median_deviations(reference: NDArray[Any], test: NDArray[Any]) -> NDAr
     return np.abs(np.where(test_dev >= 0, test_dev / pscale, test_dev / nscale))  # (S_t, F)
 
 
-def factor_deviation(
+def factor_deviation(  # noqa: C901
     reference_factors: Mapping[str, NDArray[Any]],
     test_factors: Mapping[str, NDArray[Any]],
     indices: SequenceLike[int],
@@ -183,7 +183,7 @@ def factor_deviation(
     return results
 
 
-def factor_predictors(
+def factor_predictors(  # noqa: C901
     factors: Mapping[str, NDArray[Any]],
     indices: SequenceLike[int],
     discrete_features: list[bool] | None = None,

{dataeval-1.0.4 → dataeval-1.0.5}/src/dataeval/core/_mutual_info.py

@@ -74,7 +74,7 @@ def _merge_labels_and_factors(
     return data, discrete_features
 
 
-def mutual_info(
+def mutual_info(  # noqa: C901
     class_labels: Array1D[int],
     factor_data: Array2D[int | float],
     discrete_features: Array1D[bool] | None = None,
@@ -124,13 +124,11 @@ def mutual_info(
 
     >>> rng = np.random.default_rng(175)
     >>> class_labels = rng.choice([0, 1, 2], size=100)
-    >>> factor_data = np.column_stack(
-    ...     [
-    ...         rng.choice([25, 35, 45, 55], size=100),  # age
-    ...         rng.choice([50000, 65000, 80000], size=100),  # income
-    ...         rng.choice([0, 1], size=100),  # gender
-    ...     ]
-    ... )
+    >>> factor_data = np.column_stack([
+    ...     rng.choice([25, 35, 45, 55], size=100),  # age
+    ...     rng.choice([50000, 65000, 80000], size=100),  # income
+    ...     rng.choice([0, 1], size=100),  # gender
+    ... ])
     >>> result = mutual_info(class_labels=class_labels, factor_data=factor_data)
     >>> result["class_to_factor"]
     array([1.   , 0.034, 0.026, 0.004])
@@ -250,13 +248,11 @@ def mutual_info_classwise(
 
     >>> rng = np.random.default_rng(175)
     >>> class_labels = rng.choice([0, 1, 2], size=100)
-    >>> factor_data = np.column_stack(
-    ...     [
-    ...         rng.choice([25, 35, 45, 55], size=100),  # age
-    ...         rng.choice([50000, 65000, 80000], size=100),  # income
-    ...         rng.choice([0, 1], size=100),  # gender
-    ...     ]
-    ... )
+    >>> factor_data = np.column_stack([
+    ...     rng.choice([25, 35, 45, 55], size=100),  # age
+    ...     rng.choice([50000, 65000, 80000], size=100),  # income
+    ...     rng.choice([0, 1], size=100),  # gender
+    ... ])
     >>> mutual_info_classwise(class_labels=class_labels, factor_data=factor_data)
     array([[1.000e+00, 2.077e-02, 2.296e-03, 7.317e-04],
            [1.000e+00, 4.893e-02, 2.451e-02, 4.362e-03],

{dataeval-1.0.4 → dataeval-1.0.5}/src/dataeval/core/_parity.py

@@ -42,7 +42,7 @@ class ParityResult(TypedDict):
 
 
 @experimental
-def parity(
+def parity(  # noqa: C901
     factor_data: Array2D[int],
     class_labels: Array1D[int],
 ) -> ParityResult:

{dataeval-1.0.4 → dataeval-1.0.5}/src/dataeval/core/_uap.py

@@ -57,16 +57,14 @@ def uap(labels: Array2D[int], scores: Array2D[float]) -> float:
     0.8333333333333333
 
     >>> y_true = np.array([0, 0, 1, 1, 2, 2])
-    >>> y_scores = np.array(
-    ...     [
-    ...         [0.7, 0.2, 0.1],
-    ...         [0.4, 0.3, 0.3],
-    ...         [0.1, 0.8, 0.1],
-    ...         [0.2, 0.3, 0.5],
-    ...         [0.4, 0.4, 0.2],
-    ...         [0.1, 0.2, 0.7],
-    ...     ]
-    ... )
+    >>> y_scores = np.array([
+    ...     [0.7, 0.2, 0.1],
+    ...     [0.4, 0.3, 0.3],
+    ...     [0.1, 0.8, 0.1],
+    ...     [0.2, 0.3, 0.5],
+    ...     [0.4, 0.4, 0.2],
+    ...     [0.1, 0.2, 0.7],
+    ... ])
     >>> uap(y_true, y_scores)
     0.7777777777777777
     """

{dataeval-1.0.4 → dataeval-1.0.5}/src/dataeval/extractors/_onnx.py

@@ -156,7 +156,7 @@ class OnnxExtractor(ReprMixin):
             return [transforms]
         return list(transforms)
 
-    def _load_model(self) -> None:
+    def _load_model(self) -> None:  # noqa: C901
         """Load the ONNX model and validate configuration."""
         session_cls = _get_inference_session()
         providers = _get_execution_providers()
@@ -220,7 +220,7 @@ class OnnxExtractor(ReprMixin):
         # Ensure float32 for ONNX
         return result.astype(np.float32)
 
-    def __call__(self, data: Any) -> Array:
+    def __call__(self, data: Any) -> Array:  # noqa: C901
         """
         Extract features from a batch of images.
 

{dataeval-1.0.4 → dataeval-1.0.5}/src/dataeval/extractors/_torch.py

@@ -137,7 +137,7 @@ class TorchExtractor(ReprMixin):
 
         return modules_dict[layer_name]
 
-    def __call__(self, data: Any) -> Array:
+    def __call__(self, data: Any) -> Array:  # noqa: C901
         """
         Extract features from a batch of images.
 

{dataeval-1.0.4 → dataeval-1.0.5}/src/dataeval/performance/_output.py

@@ -167,7 +167,7 @@ class SufficiencyOutput(DictOutput):
 
         return self._build_dataframe(proj_array, output)
 
-    def inv_project(self, targets: Mapping[str, ArrayLike] | ArrayLike, n_iter: int = 1000) -> pl.DataFrame:
+    def inv_project(self, targets: Mapping[str, ArrayLike] | ArrayLike, n_iter: int = 1000) -> pl.DataFrame:  # noqa: C901
         """
         Compute training samples needed to achieve target metric values.
 
@@ -214,7 +214,7 @@ class SufficiencyOutput(DictOutput):
         return self._build_inv_project_dataframe(results)
 
     @staticmethod
-    def _build_inv_project_dataframe(
+    def _build_inv_project_dataframe(  # noqa: C901
         results: Mapping[str, tuple[NDArray[Any], NDArray[np.int64]]],
     ) -> pl.DataFrame:
         """Build wide-format DataFrame from inverse projection results."""
@@ -405,7 +405,7 @@ def linear_initialization(metric: NDArray[Any], sizes: NDArray[Any], bounds: Con
     return np.array([scale, -negative_exponent, asymptote], dtype=np.float64)
 
 
-def calc_params(p_i: NDArray[Any], n_i: NDArray[Any], niter: int, unit_interval: bool) -> NDArray[np.float64]:
+def calc_params(p_i: NDArray[Any], n_i: NDArray[Any], niter: int, unit_interval: bool) -> NDArray[np.float64]:  # noqa: C901
     """
     Retrieve the inverse power curve coefficients for the line of best fit.