tabulate data makes target locations into a list

The line
```
target_location = get_target_locations(preproc_dir, subject, te, date, target_idx).tolist()
```
in `data.bmi3d.tabulate_behavior_data` makes each target location a python list [x,y,z], when really it should be a a list of np.array target locations np.array([x,y,z])

This is annoying when you try to extract a list of unique targets, for example. you have to use:
```
target_locations = np.array([np.array(loc) for loc in df['target_location'].to_numpy()])
unique_targets, target_idx = np.unique(target_locations, return_index=True, axis=0)
```

instead of the more simple
```
unique_targets, target_idx = np.unique(df['target_location'], return_index=True, axis=0)
```