Navigational features and the sf package

[roaldarbol]

We should attempt to rely on the super well-maintained sf package. It contains a lot of really nice function which could be leveraged really quickly.

• First of all, it contains all the information needed to represent geometries, such as arenas.
• It can create a bounding box (st_bbox) from a vector of points (need to find a way to create a vector of points (st_point) form two vectors (x and y coordinates). Can also create circles (https://r-spatial.github.io/sf/reference/st_nearest_feature.html), so maybe find good way to do that from animovement (and here st_buffer can be used to generate zones)
• It contains relational functions like is within, distance to, nearest point, nearest feature, overlaps with and more.
• st_make_grid which can be useful for something like an exploratory behaviour test.
• st_buffer which can also be used to make zones (https://r-spatial.github.io/sf/articles/sf1.html#geometrycollection).
• st_centroid can be used to generate a centroid for top view pose estimation animals.

[roaldarbol]

See neuroinformatics-unit/movement#377 for discussion of ROI.