How exactly write a coco annotation style for segmentation and keypoints detection?

[codingforpleasure]

I have read thoroughly the specification for the annotations, as mentioned here:
dataset fields specification

I have annotated the segments and classified them into 5 classes,
There is single class which always holds 2 keypoints for each segment, those keypoints are of different type (let's say "head" and "tail"), I'd like to distinguish between those two types.

I have read here: coco instructions the paragraph of "2. Keypoint Detection"

And in the code I have used the following fields:

1. segmentation (list[list[float]] or dict): the segmentation mask of the instance.
2. keypoints (list[float]): in the format of [x1, y1, v1,…, xn, yn, vn]. v[i] means the visibility of this keypoint. n must be equal to the number of keypoint categories.

and I have used:
MetadataCatalog.get("my_dataset").keypoint_names = ["head", "tail"]

but how do I label each keypoint?
Unfortunately I couldn't find any example :/ .

Can somebody please write/share a short example how to achieve this?

Thank you