Loading file with multiple skeletons destroys node name list

[SkepticRaven]

The read_skeletons function fails when multiple skeletons without the same nodes is read.

Summary:
The function that reads in skeletons from .slp files will read in all the names of nodes, then sort + subsets each skeleton. Due to a re-used variable name (node_names in L210 linked below), the full list of nodes becomes the first skeletons subset. Causes either IndexError if indices of second skeleton happen to point outside first skeletons sub-list or simply carries forward incorrect node names.

The offending lines are here:
Node names defined outside loop:

sleap-io/sleap_io/io/slp.py

Line 190 in 9a04f7f

node_names = [x["name"] for x in metadata["nodes"]]

Full list modified to sub-list here:

sleap-io/sleap_io/io/slp.py

Line 210 in 9a04f7f

node_names = [node_names[i] for i in skeleton_node_inds]

Sub-list used later (once here):

sleap-io/sleap_io/io/slp.py

Line 214 in 9a04f7f

for name in node_names:

Using a second variable (eg sorted_node_names) in 210 and 214 appears to fix this.