When creating an asymmetric total order basis, the resulting index set plot displays as symmetric with the larger input order.
basis = eq.Basis('total-order', orders=[5,15])
basis.plot_index_set()
The following might work:
total_order = basis.elements
# To be added to basis.py in total_order_basis()
bool_entries = np.ones((total_order.shape[0]), dtype=bool) # True
for j in range(total_order.shape[0]):
for d in range(total_order.shape[1]):
if total_order[j,d] > basis.orders[d]:
bool_entries[j] = False
total_order_new = total_order[bool_entries,:]
# Check
basis.elements = total_order_new
basis.plot_index_set()
Potential suggestion for vectorised approach using numpy.
basis = eq.Basis('total-order', orders=[5,15])
total_order = basis.elements
valid = np.prod(total_order <= basis.orders, axis=1).astype(bool)
total_order_new = total_order[valid, :]
# Check
basis.elements = total_order_new
basis.plot_index_set()
1 Like