Annotator2d object has no attribute 'layers'

[Sagnik700]

I am using micro-sam 1.7.0 from the source. This issue occurred while trying to annotate using Annotator 2D in napari.

After completing annotating one image and saving the committed objects layer, I deleted the raw image and committed objects layers from the layer list and imported the next raw image. After clicking on Compute Embeddings, the error prompt

'Annotator2d' object has no attribute 'layers'

came, and behind it the prompt saying 'Computer Embeddings for ....' came up as well, which signified that the embeddings were computed successfully. No other issues were faced during the annotation phase, and the committed objects layer got regenerated as well.

The error trace:

AttributeError Traceback (most recent call last)
File E:\source\micro-sam\micro_sam\sam_annotator_[annotator.py:204](http://annotator.py:204/), in _AnnotatorBase._update_image(self=<micro_sam.sam_annotator.annotator_2d.Annotator2d object>, segmentation_result=False)
201 self._shape = state.image_shape
203 # Before we reset the layers, we ensure all expected layers exist.
--> 204 self._require_layers()
self = <micro_sam.sam_annotator.annotator_2d.Annotator2d object at 0x000002060A958200> 206 # Update the image scale.
207 scale = state.image_scale

File E:\source\micro-sam\micro_sam\sam_annotator_[annotator.py:53](http://annotator.py:53/), in _AnnotatorBase._require_layers(self=<micro_sam.sam_annotator.annotator_2d.Annotator2d object>, layer_choices=None)
51 self._viewer.layers["committed_objects"].new_colormap()
52 if image_scale is not None:
---> 53 self.layers["committed_objects"].scale = image_scale
image_scale = (np.float64(1.0), np.float64(1.0))
self = <micro_sam.sam_annotator.annotator_2d.Annotator2d object at 0x000002060A958200> 55 # Add the point layer for point prompts.
56 self._point_labels = ["positive", "negative"]

AttributeError: 'Annotator2d' object has no attribute 'layers'

Image showing the issue:

[constantinpape]

Hi Sagnik,

the issue occurs because you are deleting the committed_objects layer. It works if you don't do this. Deleting it is also not necessary, it will be reset if you retun Compute Embeddings.

We will see if we can fix this, i.e. recover committed_objects being deleted, but since everything works as expected otherwise it's not a high priority.

A general comment: if you want to annotate many images in a fast manner we recommend to use the Image Series Annotator: https://computational-cell-analytics.github.io/micro-sam/micro_sam.html#image-series-annotator