valor-lite 0.33.3__py3-none-any.whl → 0.33.4__py3-none-any.whl

valor_lite/detection/__init__.py

@@ -1,11 +1,10 @@
-from .annotation import Bitmask, BoundingBox, Detection
+from .annotation import Bitmask, BoundingBox, Detection, Polygon
 from .computation import (
     compute_detailed_counts,
-    compute_iou,
     compute_metrics,
     compute_ranked_pairs,
 )
-from .manager import DataLoader, Evaluator
+from .manager import DataLoader, Evaluator, compute_iou
 from .metric import (
     AP,
     AR,
@@ -29,6 +28,7 @@ __all__ = [
     "Bitmask",
     "BoundingBox",
     "Detection",
+    "Polygon",
     "MetricType",
     "Counts",
     "Precision",

valor_lite/detection/annotation.py

@@ -2,6 +2,7 @@ from dataclasses import dataclass, field
 
 import numpy as np
 from numpy.typing import NDArray
+from shapely.geometry import Polygon as ShapelyPolygon
 
 
 @dataclass
@@ -24,6 +25,37 @@ class BoundingBox:
         return (self.xmin, self.xmax, self.ymin, self.ymax)
 
 
+@dataclass
+class Polygon:
+    shape: ShapelyPolygon
+    labels: list[tuple[str, str]]
+    scores: list[float] = field(default_factory=list)
+
+    def __post_init__(self):
+        if not isinstance(self.shape, ShapelyPolygon):
+            raise TypeError("shape must be of type shapely.geometry.Polygon.")
+        if len(self.scores) > 0 and len(self.labels) != len(self.scores):
+            raise ValueError(
+                "If scores are defined, there must be a 1:1 pairing with labels."
+            )
+
+    def to_box(self) -> BoundingBox | None:
+
+        if self.shape.is_empty:
+            return None
+
+        xmin, ymin, xmax, ymax = self.shape.bounds
+
+        return BoundingBox(
+            xmin=xmin,
+            xmax=xmax,
+            ymin=ymin,
+            ymax=ymax,
+            labels=self.labels,
+            scores=self.scores,
+        )
+
+
 @dataclass
 class Bitmask:
     mask: NDArray[np.bool_]
@@ -55,8 +87,8 @@ class Bitmask:
 @dataclass
 class Detection:
     uid: str
-    groundtruths: list[BoundingBox]
-    predictions: list[BoundingBox]
+    groundtruths: list[BoundingBox] | list[Bitmask] | list[Polygon]
+    predictions: list[BoundingBox] | list[Bitmask] | list[Polygon]
 
     def __post_init__(self):
         for prediction in self.predictions:

valor_lite/detection/computation.py

@@ -2,7 +2,7 @@ import numpy as np
 from numpy.typing import NDArray
 
 
-def compute_iou(data: NDArray[np.floating]) -> NDArray[np.floating]:
+def compute_bbox_iou(data: NDArray[np.floating]) -> NDArray[np.floating]:
     """
     Computes intersection-over-union (IoU) for axis-aligned bounding boxes.
 
@@ -24,14 +24,12 @@ def compute_iou(data: NDArray[np.floating]) -> NDArray[np.floating]:
     Parameters
     ----------
     data : NDArray[np.floating]
-        A sorted array of classification pairs.
-    label_metadata : NDArray[np.int32]
-        An array containing metadata related to labels.
+        A sorted array of bounding box pairs.
 
     Returns
     -------
     NDArray[np.floating]
-        Compute IoU's.
+        Computed IoU's.
     """
 
     xmin1, xmax1, ymin1, ymax1 = (
@@ -69,6 +67,73 @@ def compute_iou(data: NDArray[np.floating]) -> NDArray[np.floating]:
     return iou
 
 
+def compute_bitmask_iou(data: NDArray[np.floating]) -> NDArray[np.floating]:
+    """
+    Computes intersection-over-union (IoU) for bitmasks.
+
+    Takes data with shape (N, 2):
+
+    Index 0 - first bitmask
+    Index 1 - second bitmask
+
+    Returns data with shape (N, 1):
+
+    Index 0 - IoU
+
+    Parameters
+    ----------
+    data : NDArray[np.floating]
+        A sorted array of bitmask pairs.
+
+    Returns
+    -------
+    NDArray[np.floating]
+        Computed IoU's.
+    """
+    intersection_ = np.array([np.logical_and(x, y).sum() for x, y in data])
+    union_ = np.array([np.logical_or(x, y).sum() for x, y in data])
+
+    return intersection_ / union_
+
+
+def compute_polygon_iou(
+    data: NDArray[np.floating],
+) -> NDArray[np.floating]:
+    """
+    Computes intersection-over-union (IoU) for shapely polygons.
+
+    Takes data with shape (N, 2):
+
+    Index 0 - first polygon
+    Index 1 - second polygon
+
+    Returns data with shape (N, 1):
+
+    Index 0 - IoU
+
+    Parameters
+    ----------
+    data : NDArray[np.floating]
+        A sorted array of polygon pairs.
+
+    Returns
+    -------
+    NDArray[np.floating]
+        Computed IoU's.
+    """
+    intersection_ = np.array(
+        [poly1.intersection(poly2).area for poly1, poly2 in data]
+    )
+    union_ = np.array(
+        [
+            poly1.area + poly2.area - intersection_[i]
+            for i, (poly1, poly2) in enumerate(data)
+        ]
+    )
+
+    return intersection_ / union_
+
+
 def _compute_ranked_pairs_for_datum(
     data: NDArray[np.floating],
     label_metadata: NDArray[np.int32],
@@ -133,7 +198,7 @@ def compute_ranked_pairs(
     Parameters
     ----------
     data : NDArray[np.floating]
-        A sorted array of classification pairs.
+        A sorted array summarizing the IOU calculations of one or more pairs.
     label_metadata : NDArray[np.int32]
         An array containing metadata related to labels.
 
@@ -161,10 +226,10 @@ def compute_ranked_pairs(
 
 
 def compute_metrics(
-    data: np.ndarray,
-    label_metadata: np.ndarray,
-    iou_thresholds: np.ndarray,
-    score_thresholds: np.ndarray,
+    data: NDArray[np.floating],
+    label_metadata: NDArray[np.int32],
+    iou_thresholds: NDArray[np.floating],
+    score_thresholds: NDArray[np.floating],
 ) -> tuple[
     tuple[
         NDArray[np.floating],
@@ -197,7 +262,7 @@ def compute_metrics(
     Parameters
     ----------
     data : NDArray[np.floating]
-        A sorted array of classification pairs.
+        A sorted array summarizing the IOU calculations of one or more pairs.
     label_metadata : NDArray[np.int32]
         An array containing metadata related to labels.
     iou_thresholds : NDArray[np.floating]
@@ -463,7 +528,7 @@ def compute_detailed_counts(
     Parameters
     ----------
     data : NDArray[np.floating]
-        A sorted array of classification pairs.
+        A sorted array summarizing the IOU calculations of one or more pairs.
     label_metadata : NDArray[np.int32]
         An array containing metadata related to labels.
     iou_thresholds : NDArray[np.floating]

valor_lite/detection/manager.py

@@ -3,12 +3,20 @@ from dataclasses import dataclass
 
 import numpy as np
 from numpy.typing import NDArray
+from shapely.geometry import Polygon as ShapelyPolygon
 from tqdm import tqdm
-from valor_lite.detection.annotation import Detection
+from valor_lite.detection.annotation import (
+    Bitmask,
+    BoundingBox,
+    Detection,
+    Polygon,
+)
 from valor_lite.detection.computation import (
+    compute_bbox_iou,
+    compute_bitmask_iou,
     compute_detailed_counts,
-    compute_iou,
     compute_metrics,
+    compute_polygon_iou,
     compute_ranked_pairs,
 )
 from valor_lite.detection.metric import (
@@ -35,7 +43,7 @@ Usage
 -----
 
 loader = DataLoader()
-loader.add_data(
+loader.add_bounding_boxes(
     groundtruths=groundtruths,
     predictions=predictions,
 )
@@ -51,6 +59,103 @@ filtered_metrics = evaluator.evaluate(iou_thresholds=[0.5], filter_mask=filter_m
 """
 
 
+def _get_valor_dict_annotation_key(
+    annotation_type: type[BoundingBox] | type[Polygon] | type[Bitmask],
+) -> str:
+    """Get the correct JSON key to extract a given annotation type."""
+
+    if issubclass(annotation_type, BoundingBox):
+        return "bounding_box"
+    if issubclass(annotation_type, Polygon):
+        return "polygon"
+    else:
+        return "raster"
+
+
+def _get_annotation_representation(
+    annotation_type: type[BoundingBox] | type[Polygon] | type[Bitmask],
+) -> str:
+    """Get the correct representation of an annotation object."""
+
+    representation = (
+        "extrema"
+        if issubclass(annotation_type, BoundingBox)
+        else ("mask" if issubclass(annotation_type, Bitmask) else "shape")
+    )
+
+    return representation
+
+
+def _get_annotation_representation_from_valor_dict(
+    data: list,
+    annotation_type: type[BoundingBox] | type[Polygon] | type[Bitmask],
+) -> tuple[float, float, float, float] | ShapelyPolygon | NDArray[np.bool_]:
+    """Get the correct representation of an annotation object from a valor dictionary."""
+
+    if issubclass(annotation_type, BoundingBox):
+        x = [point[0] for shape in data for point in shape]
+        y = [point[1] for shape in data for point in shape]
+        return (min(x), max(x), min(y), max(y))
+    if issubclass(annotation_type, Polygon):
+        return ShapelyPolygon(data)
+    else:
+        return np.array(data)
+
+
+def _get_annotation_data(
+    keyed_groundtruths: dict,
+    keyed_predictions: dict,
+    annotation_type: type[BoundingBox] | type[Polygon] | type[Bitmask] | None,
+    key=int,
+) -> np.ndarray:
+    """Create an array of annotation pairs for use when calculating IOU. Needed because we unpack bounding box representations, but not bitmask or polygon representations."""
+    if annotation_type == BoundingBox:
+        return np.array(
+            [
+                np.array([*gextrema, *pextrema])
+                for _, _, _, pextrema in keyed_predictions[key]
+                for _, _, gextrema in keyed_groundtruths[key]
+            ]
+        )
+    else:
+        return np.array(
+            [
+                np.array([groundtruth_obj, prediction_obj])
+                for _, _, _, prediction_obj in keyed_predictions[key]
+                for _, _, groundtruth_obj in keyed_groundtruths[key]
+            ]
+        )
+
+
+def compute_iou(
+    data: NDArray[np.floating],
+    annotation_type: type[BoundingBox] | type[Polygon] | type[Bitmask],
+) -> NDArray[np.floating]:
+    """
+    Computes intersection-over-union (IoU) calculations for various annotation types.
+
+    Parameters
+    ----------
+    data : NDArray[np.floating]
+        A sorted array of bounding box, bitmask, or polygon pairs.
+    annotation_type: type[BoundingBox] | type[Polygon] | type[Bitmask]
+        The type of annotation contained in the data.
+
+
+    Returns
+    -------
+    NDArray[np.floating]
+        Computed IoU's.
+    """
+
+    if annotation_type == BoundingBox:
+        return compute_bbox_iou(data=data)
+    elif annotation_type == Bitmask:
+        return compute_bitmask_iou(data=data)
+    else:
+        return compute_polygon_iou(data=data)
+
+
 @dataclass
 class Filter:
     indices: NDArray[np.int32]
@@ -167,46 +272,35 @@ class Evaluator:
                     [self.uid_to_index[uid] for uid in datum_uids],
                     dtype=np.int32,
                 )
-            mask = np.zeros_like(mask_pairs, dtype=np.bool_)
-            mask[
-                np.isin(self._ranked_pairs[:, 0].astype(int), datum_uids)
-            ] = True
-            mask_pairs &= mask
-
-            mask = np.zeros_like(mask_datums, dtype=np.bool_)
-            mask[datum_uids] = True
-            mask_datums &= mask
+            mask_pairs[
+                ~np.isin(self._ranked_pairs[:, 0].astype(int), datum_uids)
+            ] = False
+            mask_datums[~np.isin(np.arange(n_datums), datum_uids)] = False
 
         if labels is not None:
             if isinstance(labels, list):
                 labels = np.array(
                     [self.label_to_index[label] for label in labels]
                 )
-            mask = np.zeros_like(mask_pairs, dtype=np.bool_)
-            mask[np.isin(self._ranked_pairs[:, 4].astype(int), labels)] = True
-            mask_pairs &= mask
-
-            mask = np.zeros_like(mask_labels, dtype=np.bool_)
-            mask[labels] = True
-            mask_labels &= mask
+            mask_pairs[
+                ~np.isin(self._ranked_pairs[:, 4].astype(int), labels)
+            ] = False
+            mask_labels[~np.isin(np.arange(n_labels), labels)] = False
 
         if label_keys is not None:
             if isinstance(label_keys, list):
                 label_keys = np.array(
                     [self.label_key_to_index[key] for key in label_keys]
                 )
-            label_indices = np.where(
-                np.isclose(self._label_metadata[:, 2], label_keys)
-            )[0]
-            mask = np.zeros_like(mask_pairs, dtype=np.bool_)
-            mask[
-                np.isin(self._ranked_pairs[:, 4].astype(int), label_indices)
-            ] = True
-            mask_pairs &= mask
-
-            mask = np.zeros_like(mask_labels, dtype=np.bool_)
-            mask[label_indices] = True
-            mask_labels &= mask
+            label_indices = (
+                np.where(np.isclose(self._label_metadata[:, 2], label_keys))[0]
+                if label_keys.size > 0
+                else np.array([])
+            )
+            mask_pairs[
+                ~np.isin(self._ranked_pairs[:, 4].astype(int), label_indices)
+            ] = False
+            mask_labels[~np.isin(np.arange(n_labels), label_indices)] = False
 
         mask = mask_datums[:, np.newaxis] & mask_labels[np.newaxis, :]
         label_metadata_per_datum = self._label_metadata_per_datum.copy()
@@ -294,7 +388,7 @@ class Evaluator:
             )
             for iou_idx in range(average_precision.shape[0])
             for label_idx in range(average_precision.shape[1])
-            if int(label_metadata[label_idx][0]) > 0
+            if int(label_metadata[label_idx, 0]) > 0
         ]
 
         metrics[MetricType.mAP] = [
@@ -314,7 +408,7 @@ class Evaluator:
                 label=self.index_to_label[label_idx],
             )
             for label_idx in range(self.n_labels)
-            if int(label_metadata[label_idx][0]) > 0
+            if int(label_metadata[label_idx, 0]) > 0
         ]
 
         metrics[MetricType.mAPAveragedOverIOUs] = [
@@ -337,7 +431,7 @@ class Evaluator:
             )
             for score_idx in range(average_recall.shape[0])
             for label_idx in range(average_recall.shape[1])
-            if int(label_metadata[label_idx][0]) > 0
+            if int(label_metadata[label_idx, 0]) > 0
         ]
 
         metrics[MetricType.mAR] = [
@@ -359,7 +453,7 @@ class Evaluator:
                 label=self.index_to_label[label_idx],
             )
             for label_idx in range(self.n_labels)
-            if int(label_metadata[label_idx][0]) > 0
+            if int(label_metadata[label_idx, 0]) > 0
         ]
 
         metrics[MetricType.mARAveragedOverScores] = [
@@ -382,16 +476,17 @@ class Evaluator:
             )
             for iou_idx, iou_threshold in enumerate(iou_thresholds)
             for label_idx, label in self.index_to_label.items()
-            if int(label_metadata[label_idx][0]) > 0
+            if int(label_metadata[label_idx, 0]) > 0
         ]
 
         for label_idx, label in self.index_to_label.items():
+
+            if label_metadata[label_idx, 0] == 0:
+                continue
+
             for score_idx, score_threshold in enumerate(score_thresholds):
                 for iou_idx, iou_threshold in enumerate(iou_thresholds):
 
-                    if label_metadata[label_idx, 0] == 0:
-                        continue
-
                     row = precision_recall[iou_idx][score_idx][label_idx]
                     kwargs = {
                         "label": label,
@@ -475,7 +570,7 @@ class Evaluator:
             return list()
 
         metrics = compute_detailed_counts(
-            self._detailed_pairs,
+            data=self._detailed_pairs,
             label_metadata=self._label_metadata,
             iou_thresholds=np.array(iou_thresholds),
             score_thresholds=np.array(score_thresholds),
@@ -661,12 +756,27 @@ class DataLoader:
             self._evaluator.label_key_to_index[label[0]],
         )
 
-    def _add_data(
+    def _compute_ious_and_cache_pairs(
         self,
         uid_index: int,
         keyed_groundtruths: dict,
         keyed_predictions: dict,
-    ):
+        annotation_type: type[BoundingBox] | type[Polygon] | type[Bitmask],
+    ) -> None:
+        """
+        Compute IOUs between groundtruths and preditions before storing as pairs.
+
+        Parameters
+        ----------
+        uid_index: int
+            The index of the detection.
+        keyed_groundtruths: dict
+            A dictionary of groundtruths.
+        keyed_predictions: dict
+            A dictionary of predictions.
+        annotation_type: type[BoundingBox] | type[Polygon] | type[Bitmask]
+            The type of annotation to compute IOUs for.
+        """
         gt_keys = set(keyed_groundtruths.keys())
         pd_keys = set(keyed_predictions.keys())
         joint_keys = gt_keys.intersection(pd_keys)
@@ -677,14 +787,13 @@ class DataLoader:
         for key in joint_keys:
             n_predictions = len(keyed_predictions[key])
             n_groundtruths = len(keyed_groundtruths[key])
-            boxes = np.array(
-                [
-                    np.array([*gextrema, *pextrema])
-                    for _, _, _, pextrema in keyed_predictions[key]
-                    for _, _, gextrema in keyed_groundtruths[key]
-                ]
+            data = _get_annotation_data(
+                keyed_groundtruths=keyed_groundtruths,
+                keyed_predictions=keyed_predictions,
+                key=key,
+                annotation_type=annotation_type,
             )
-            ious = compute_iou(boxes)
+            ious = compute_iou(data=data, annotation_type=annotation_type)
             mask_nonzero_iou = (ious > 1e-9).reshape(
                 (n_predictions, n_groundtruths)
             )
@@ -780,9 +889,10 @@ class DataLoader:
 
         self.pairs.append(np.array(pairs))
 
-    def add_data(
+    def _add_data(
         self,
         detections: list[Detection],
+        annotation_type: type[Bitmask] | type[BoundingBox] | type[Polygon],
         show_progress: bool = False,
     ):
         """
@@ -792,6 +902,8 @@ class DataLoader:
         ----------
         detections : list[Detection]
             A list of Detection objects.
+        annotation_type : type[Bitmask] | type[BoundingBox] | type[Polygon]
+            The annotation type to process.
         show_progress : bool, default=False
             Toggle for tqdm progress bar.
         """
@@ -817,45 +929,146 @@ class DataLoader:
             # cache labels and annotations
             keyed_groundtruths = defaultdict(list)
             keyed_predictions = defaultdict(list)
+
+            representation_property = _get_annotation_representation(
+                annotation_type=annotation_type
+            )
+
             for gidx, gann in enumerate(detection.groundtruths):
-                self._evaluator.groundtruth_examples[uid_index][
-                    gidx
-                ] = np.array(gann.extrema)
+                if not isinstance(gann, annotation_type):
+                    raise ValueError(
+                        f"Expected {annotation_type}, but annotation is of type {type(gann)}."
+                    )
+
+                if isinstance(gann, BoundingBox):
+                    self._evaluator.groundtruth_examples[uid_index][
+                        gidx
+                    ] = getattr(gann, representation_property)
+                else:
+                    converted_box = gann.to_box()
+                    self._evaluator.groundtruth_examples[uid_index][gidx] = (
+                        getattr(converted_box, "extrema")
+                        if converted_box is not None
+                        else None
+                    )
                 for glabel in gann.labels:
                     label_idx, label_key_idx = self._add_label(glabel)
                     self.groundtruth_count[label_idx][uid_index] += 1
+                    representation = getattr(gann, representation_property)
                     keyed_groundtruths[label_key_idx].append(
                         (
                             gidx,
                             label_idx,
-                            gann.extrema,
+                            representation,
                         )
                     )
+
             for pidx, pann in enumerate(detection.predictions):
-                self._evaluator.prediction_examples[uid_index][
-                    pidx
-                ] = np.array(pann.extrema)
+                if not isinstance(pann, annotation_type):
+                    raise ValueError(
+                        f"Expected {annotation_type}, but annotation is of type {type(pann)}."
+                    )
+
+                if isinstance(pann, BoundingBox):
+                    self._evaluator.prediction_examples[uid_index][
+                        pidx
+                    ] = getattr(pann, representation_property)
+                else:
+                    converted_box = pann.to_box()
+                    self._evaluator.prediction_examples[uid_index][pidx] = (
+                        getattr(converted_box, "extrema")
+                        if converted_box is not None
+                        else None
+                    )
                 for plabel, pscore in zip(pann.labels, pann.scores):
                     label_idx, label_key_idx = self._add_label(plabel)
                     self.prediction_count[label_idx][uid_index] += 1
+                    representation = representation = getattr(
+                        pann, representation_property
+                    )
                     keyed_predictions[label_key_idx].append(
                         (
                             pidx,
                             label_idx,
                             pscore,
-                            pann.extrema,
+                            representation,
                         )
                     )
 
-            self._add_data(
+            self._compute_ious_and_cache_pairs(
                 uid_index=uid_index,
                 keyed_groundtruths=keyed_groundtruths,
                 keyed_predictions=keyed_predictions,
+                annotation_type=annotation_type,
             )
 
-    def add_data_from_valor_dict(
+    def add_bounding_boxes(
+        self,
+        detections: list[Detection],
+        show_progress: bool = False,
+    ):
+        """
+        Adds bounding box detections to the cache.
+
+        Parameters
+        ----------
+        detections : list[Detection]
+            A list of Detection objects.
+        show_progress : bool, default=False
+            Toggle for tqdm progress bar.
+        """
+        return self._add_data(
+            detections=detections,
+            show_progress=show_progress,
+            annotation_type=BoundingBox,
+        )
+
+    def add_polygons(
+        self,
+        detections: list[Detection],
+        show_progress: bool = False,
+    ):
+        """
+        Adds polygon detections to the cache.
+
+        Parameters
+        ----------
+        detections : list[Detection]
+            A list of Detection objects.
+        show_progress : bool, default=False
+            Toggle for tqdm progress bar.
+        """
+        return self._add_data(
+            detections=detections,
+            show_progress=show_progress,
+            annotation_type=Polygon,
+        )
+
+    def add_bitmasks(
+        self,
+        detections: list[Detection],
+        show_progress: bool = False,
+    ):
+        """
+        Adds bitmask detections to the cache.
+
+        Parameters
+        ----------
+        detections : list[Detection]
+            A list of Detection objects.
+        show_progress : bool, default=False
+            Toggle for tqdm progress bar.
+        """
+        return self._add_data(
+            detections=detections,
+            show_progress=show_progress,
+            annotation_type=Bitmask,
+        )
+
+    def _add_data_from_valor_dict(
         self,
         detections: list[tuple[dict, dict]],
+        annotation_type: type[Bitmask] | type[BoundingBox] | type[Polygon],
         show_progress: bool = False,
     ):
         """
@@ -865,20 +1078,14 @@ class DataLoader:
         ----------
         detections : list[tuple[dict, dict]]
             A list of groundtruth, prediction pairs in Valor-format dictionaries.
+        annotation_type : type[Bitmask] | type[BoundingBox] | type[Polygon]
+            The annotation type to process.
         show_progress : bool, default=False
             Toggle for tqdm progress bar.
         """
 
-        def _get_bbox_extrema(
-            data: list[list[list[float]]],
-        ) -> tuple[float, float, float, float]:
-            x = [point[0] for shape in data for point in shape]
-            y = [point[1] for shape in data for point in shape]
-            return (min(x), max(x), min(y), max(y))
-
         disable_tqdm = not show_progress
         for groundtruth, prediction in tqdm(detections, disable=disable_tqdm):
-
             # update metadata
             self._evaluator.n_datums += 1
             self._evaluator.n_groundtruths += len(groundtruth["annotations"])
@@ -898,10 +1105,34 @@ class DataLoader:
             # cache labels and annotations
             keyed_groundtruths = defaultdict(list)
             keyed_predictions = defaultdict(list)
+
+            annotation_key = _get_valor_dict_annotation_key(
+                annotation_type=annotation_type
+            )
+            invalid_keys = list(
+                filter(
+                    lambda x: x != annotation_key,
+                    ["bounding_box", "raster", "polygon"],
+                )
+            )
+
             for gidx, gann in enumerate(groundtruth["annotations"]):
-                self._evaluator.groundtruth_examples[uid_index][
-                    gidx
-                ] = np.array(_get_bbox_extrema(gann["bounding_box"]))
+                if (gann[annotation_key] is None) or any(
+                    [gann[k] is not None for k in invalid_keys]
+                ):
+                    raise ValueError(
+                        f"Input JSON doesn't contain {annotation_type} data, or contains data for multiple annotation types."
+                    )
+                if annotation_type == BoundingBox:
+                    self._evaluator.groundtruth_examples[uid_index][
+                        gidx
+                    ] = np.array(
+                        _get_annotation_representation_from_valor_dict(
+                            gann[annotation_key],
+                            annotation_type=annotation_type,
+                        ),
+                    )
+
                 for valor_label in gann["labels"]:
                     glabel = (valor_label["key"], valor_label["value"])
                     label_idx, label_key_idx = self._add_label(glabel)
@@ -910,13 +1141,29 @@ class DataLoader:
                         (
                             gidx,
                             label_idx,
-                            _get_bbox_extrema(gann["bounding_box"]),
+                            _get_annotation_representation_from_valor_dict(
+                                gann[annotation_key],
+                                annotation_type=annotation_type,
+                            ),
                         )
                     )
             for pidx, pann in enumerate(prediction["annotations"]):
-                self._evaluator.prediction_examples[uid_index][
-                    pidx
-                ] = np.array(_get_bbox_extrema(pann["bounding_box"]))
+                if (pann[annotation_key] is None) or any(
+                    [pann[k] is not None for k in invalid_keys]
+                ):
+                    raise ValueError(
+                        f"Input JSON doesn't contain {annotation_type} data, or contains data for multiple annotation types."
+                    )
+
+                if annotation_type == BoundingBox:
+                    self._evaluator.prediction_examples[uid_index][
+                        pidx
+                    ] = np.array(
+                        _get_annotation_representation_from_valor_dict(
+                            pann[annotation_key],
+                            annotation_type=annotation_type,
+                        )
+                    )
                 for valor_label in pann["labels"]:
                     plabel = (valor_label["key"], valor_label["value"])
                     pscore = valor_label["score"]
@@ -927,16 +1174,41 @@ class DataLoader:
                             pidx,
                             label_idx,
                             pscore,
-                            _get_bbox_extrema(pann["bounding_box"]),
+                            _get_annotation_representation_from_valor_dict(
+                                pann[annotation_key],
+                                annotation_type=annotation_type,
+                            ),
                         )
                     )
 
-            self._add_data(
+            self._compute_ious_and_cache_pairs(
                 uid_index=uid_index,
                 keyed_groundtruths=keyed_groundtruths,
                 keyed_predictions=keyed_predictions,
+                annotation_type=annotation_type,
             )
 
+    def add_bounding_boxes_from_valor_dict(
+        self,
+        detections: list[tuple[dict, dict]],
+        show_progress: bool = False,
+    ):
+        """
+        Adds Valor-format bounding box detections to the cache.
+
+        Parameters
+        ----------
+        detections : list[tuple[dict, dict]]
+            A list of groundtruth, prediction pairs in Valor-format dictionaries.
+        show_progress : bool, default=False
+            Toggle for tqdm progress bar.
+        """
+        return self._add_data_from_valor_dict(
+            detections=detections,
+            show_progress=show_progress,
+            annotation_type=BoundingBox,
+        )
+
     def finalize(self) -> Evaluator:
         """
         Performs data finalization and some preprocessing steps.

{valor_lite-0.33.3.dist-info → valor_lite-0.33.4.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: valor-lite
-Version: 0.33.3
+Version: 0.33.4
 Summary: Compute valor metrics directly in your client.
 License: MIT License
         
@@ -32,6 +32,7 @@ Requires-Dist: Pillow >=9.1.0
 Requires-Dist: tqdm
 Requires-Dist: requests
 Requires-Dist: numpy
+Requires-Dist: shapely
 Requires-Dist: importlib-metadata ; python_version < "3.8"
 Provides-Extra: test
 Requires-Dist: pytest ; extra == 'test'

valor_lite-0.33.4.dist-info/RECORD

@@ -0,0 +1,12 @@
+valor_lite/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+valor_lite/schemas.py,sha256=r4cC10w1xYsA785KmGE4ePeOX3wzEs846vT7QAiVg_I,293
+valor_lite/detection/__init__.py,sha256=taEB7NQBsyCSsMtvDA7E_FhDxMfJB1rax-Rl1ZtRMoE,1017
+valor_lite/detection/annotation.py,sha256=BspLc3SjWXj6qYlGGpzDPHEZ8j7CiFzIL5cNlk0WCAM,2732
+valor_lite/detection/computation.py,sha256=AsF9zb_c7XQ7z3LfOAtMPZDkmuCZmB8HeAMZJlCaO6U,24696
+valor_lite/detection/manager.py,sha256=vnouYdx_Ul9jz_pOYt8xfvdPrNy0S4SB838KXvtS1Bw,45301
+valor_lite/detection/metric.py,sha256=DLqpODJZOG7SCqt7TCgR4am68PQORRCIQW_SXiTb1IA,9473
+valor_lite-0.33.4.dist-info/LICENSE,sha256=M0L53VuwfEEqezhHb7NPeYcO_glw7-k4DMLZQ3eRN64,1068
+valor_lite-0.33.4.dist-info/METADATA,sha256=Eqb7KlTizDcjIV7eWM67zgdbbbVICGURdGrbben2NrI,1865
+valor_lite-0.33.4.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
+valor_lite-0.33.4.dist-info/top_level.txt,sha256=9ujykxSwpl2Hu0_R95UQTR_l07k9UUTSdrpiqmq6zc4,11
+valor_lite-0.33.4.dist-info/RECORD,,

valor_lite-0.33.3.dist-info/RECORD

@@ -1,12 +0,0 @@
-valor_lite/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-valor_lite/schemas.py,sha256=r4cC10w1xYsA785KmGE4ePeOX3wzEs846vT7QAiVg_I,293
-valor_lite/detection/__init__.py,sha256=WHLHwHoKzXTBjkjC6E1_lhqB7gRWkiGWVWPqkKn-yK8,997
-valor_lite/detection/annotation.py,sha256=c45pZD1Pp2vf5GeyW_6Kl9JCx5FoaaktCaaa4q3QDUo,1758
-valor_lite/detection/computation.py,sha256=7PttK0VuOWlhRN92wpLVrGzB7RAdfdZyT3b1aTm_WaI,23214
-valor_lite/detection/manager.py,sha256=ziVnukGs-WrkyBEBBO3LVSv4LTbaWFaWqLWarVosj2c,35807
-valor_lite/detection/metric.py,sha256=DLqpODJZOG7SCqt7TCgR4am68PQORRCIQW_SXiTb1IA,9473
-valor_lite-0.33.3.dist-info/LICENSE,sha256=M0L53VuwfEEqezhHb7NPeYcO_glw7-k4DMLZQ3eRN64,1068
-valor_lite-0.33.3.dist-info/METADATA,sha256=FBpd-wMWv-m37EK5BfFuiVmnJXg4GNzCaJrTDHv4-gE,1842
-valor_lite-0.33.3.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
-valor_lite-0.33.3.dist-info/top_level.txt,sha256=9ujykxSwpl2Hu0_R95UQTR_l07k9UUTSdrpiqmq6zc4,11
-valor_lite-0.33.3.dist-info/RECORD,,