use 2d version and edgetam-dimos

Author: paul-nechifor

dimos/agents/skills/person_follow.py

@@ -61,7 +61,7 @@ def __init__(
         self,
         camera_info: CameraInfo,
         global_config: GlobalConfig,
-        use_3d_navigation: bool = True,
+        use_3d_navigation: bool = False,
     ) -> None:
         super().__init__()
         self._global_config: GlobalConfig = global_config

dimos/models/segmentation/edge_tam.py

@@ -36,7 +36,7 @@
 from dimos.utils.logging_config import setup_logger
 
 if TYPE_CHECKING:
-    from sam2.sam2_video_predictor import SAM2VideoPredictor
+    from sam2.sam2_video_predictor import SAM2VideoPredictor  # type: ignore[import-untyped]
 
 os.environ['TQDM_DISABLE'] = '1'
 

dimos/simulation/mujoco/person_on_track.py

@@ -12,8 +12,11 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from typing import Any
+
 import mujoco
 import numpy as np
+from numpy.typing import NDArray
 
 from dimos.core.transport import LCMTransport
 from dimos.msgs.geometry_msgs import Pose
@@ -145,7 +148,7 @@ def tick(self) -> None:
         # Publish pose
         self._publish_pose(self._current_pos, heading + np.pi)
 
-    def _publish_pose(self, pos: np.ndarray, heading: float) -> None:
+    def _publish_pose(self, pos: NDArray[np.floating[Any]], heading: float) -> None:
         c, s = np.cos(heading / 2), np.sin(heading / 2)
         pose = Pose(
             position=[pos[0], pos[1], 0.0],

pyproject.toml

@@ -101,7 +101,7 @@ misc = [
     # Perception Dependencies
     "scikit-learn",
     "timm>=1.0.15",
-    "edgetam @ git+https://github.com/facebookresearch/EdgeTAM.git",
+    "edgetam-dimos",
     "opencv-contrib-python==4.10.0.84",
     "omegaconf>=2.3.0",
     "hydra-core>=1.3.0",