feat: add scalar viz and fix conversion

moai/conf/model/monads/geometry/rotation/roma_unitquat_to_rotmat.yaml

@@ -1,3 +1,3 @@
-# @package model.monads._name_
+# @package model.monads.roma_unitquat_to_rotmat
 
-_target_: moai.monads.geometry.rotations.Quaternion2RotationVector
+_target_: moai.monads.geometry.rotations.Quaternion2RotationMatrix

moai/conf/model/monitors/visualization/rerun/scalar.yaml

@@ -0,0 +1,4 @@
+# @package model.monitors.scalar
+
+_target_: moai.visualization.rerun.scalars.scalar
+_partial_: true

moai/monads/geometry/rotations/conversions.py

@@ -31,7 +31,7 @@ def __init__(
     ):
         super().__init__()
         if to == "mat":
-            self.convert_func = kn.geometry.conversions.angle_axis_to_rotation_matrix
+            self.convert_func = kn.geometry.conversions.axis_angle_to_rotation_matrix
         elif to == "quat":
             # (x, y, z, w) format of quaternion
             self.convert_func = kn.geometry.conversions.angle_axis_to_quaternion

moai/visualization/rerun/scalars.py

@@ -1,13 +1,12 @@
 import logging
 import typing
 
+import colour
+import numpy as np
 import torch
 
 from moai.core.execution.constants import Constants as C
 
-# from moai.utils.color.colormap import colormap
-
-
 log = logging.getLogger(__name__)
 
 try:
@@ -19,9 +18,31 @@
 
 __all__ = [
     "optimization_losses",
+    "scalar",
 ]
 
 
+def scalar(
+    value: np.ndarray,
+    path: str,
+    color: str,
+    optimization_step: typing.Optional[int] = None,
+    lightning_step: typing.Optional[int] = None,
+    iteration: typing.Optional[int] = None,
+):
+    if optimization_step is not None:
+        rr.set_time_sequence("optimization_step", optimization_step)
+    elif lightning_step is not None:
+        rr.set_time_sequence("lightning_step", lightning_step)
+    elif iteration is not None:
+        rr.set_time_sequence("iteration", iteration)
+    color = colour.Color(color)
+    rr.log(
+        path,
+        rr.Scalar(float(value)),
+    )
+
+
 def optimization_losses(
     tensors: typing.Dict[str, torch.Tensor],
     path: str,