Merge branch 'release/v1.3.0' of https://github.com/voxel51/fiftyone …

…into develop

fiftyone/core/collections.py

@@ -7680,6 +7680,22 @@ def count(self, field_or_expr=None, expr=None, safe=False):
         Returns:
             the count
         """
+
+        # Optimization: use estimated document count when possible
+        if self._is_full_collection() and (
+            expr is None
+            and (
+                field_or_expr is None
+                or (
+                    etau.is_str(field_or_expr)
+                    and field_or_expr == "frames"
+                    and self._has_frame_fields()
+                )
+            )
+        ):
+            frames = field_or_expr == "frames"
+            return self._dataset._estimated_count(frames=frames)
+
         make = lambda field_or_expr: foa.Count(
             field_or_expr, expr=expr, safe=safe
         )
@@ -10556,6 +10572,22 @@ def _has_frame_fields(self):
     def _handle_id_fields(self, field_name):
         return _handle_id_fields(self, field_name)
 
+    def _is_full_collection(self):
+        if isinstance(self, fod.Dataset) and self.media_type != fom.GROUP:
+            return True
+
+        # pylint:disable=no-member
+        if (
+            isinstance(self, fov.DatasetView)
+            and self._dataset.media_type == fom.GROUP
+            and len(self._stages) == 1
+            and isinstance(self._stages[0], fos.SelectGroupSlices)
+            and self._pipeline() == []
+        ):
+            return True
+
+        return False
+
     def _is_label_field(self, field_name, label_type_or_types):
         try:
             label_type = self._get_label_field_type(field_name)

fiftyone/core/dataset.py

@@ -360,6 +360,15 @@ def __deepcopy__(self, memo):
     def __len__(self):
         return self.count()
 
+    def _estimated_count(self, frames=False):
+        if frames:
+            if self._frame_collection is None:
+                return None
+
+            return self._frame_collection.estimated_document_count()
+
+        return self._sample_collection.estimated_document_count()
+
     def __getitem__(self, id_filepath_slice):
         if isinstance(id_filepath_slice, numbers.Integral):
             raise ValueError(