updating some details and discussion in example 1

[StevePny]

No description provided.

[StevePny]

@kysolvik I've updated for the xarray refactor - it could use your review when you're ready.

[kysolvik]

@StevePny Looks good to me! Two potential things before merging:

1. Instead of help(data) at the top, maybe we should just do help(data.Lorenz63) since the full data module is quite large now and printing out all the docstrings/help info is quite long. I think I probably wrote help(data) a while ago and just never updated it.
2. When plotting, we're indexing the ds['x'].data array object, instead of indexing the dataset itself. For example:

plt.plot(ds_l63['x'].data[:,0], ds_l63['x'].data[:,2], label="Original", color='gray')
plt.plot(ds_l63['x'].data[0,0], ds_l63['x'].data[0,2], marker='o', color='gray')

could be rewritten as:

plt.plot(ds_l63['x'].isel(index=0).data,  ds_l63['x'].isel(index=2).data label="Original", color='gray')
plt.plot(ds_l63['x'].isel(time=0, index=0), data ds_l63['x'].isel(time=0, index=2), marker='o', color='gray')

Either works, but I wonder if we should steer people towards using xarray's indexing instead of indexing the resulting numpy/jax array? What do you think?