Create pipeline.utils.manualsegmenter.py

[Neuro-CS]

Added a manual segmentation iPython widget to pipeline.utils. To use it, run the following in a Jupyter notebook:

%matplotlib ipympl

from pipeline.utils import manualsegmenter as ms
key = dict(animal_id=0, session=0, scan_idx=0)  # Your key here
ms.ManualSegmentationWidget({**key, "field": 1, "channel": 1})

Two critical limitations:

1. Segmentations cannot be updated. The class can add manually segmented masks only if no manual segmentation currently exists
2. The classification dropdown menu will apply the selected classification to ALL masks on insertion. You cannot specify different classifications on a per-mask basis.

To draw multiple, disconnected shapes and have them belong to the same mask, click the "Multi-shape mode" button and then draw all pieces of your mask. Once done, click "Finish Mask" to add the pieces as a new mask.