Diamond Norm for a Pauli Channel about pygsti HOT 2 CLOSED

Good question - there are a couple things you need to do differently to compute the diamond norm you're after:

1. pyGSTi's diamonddist function operates on the "superoperator" or "Pauli Transfer Matrix" form of the gate, not the Choi form (in which a Pauli channel is diagonal) - so you need to convert your diagonal Choi matrices to the superoperator form using jamiolkowski_iso.
2. pyGSTi's Pauli basis matrices are normalized to 1.0 rather than d (the dimension of the Hilbert space, so 2 for a single qubit) as they are in https://arxiv.org/abs/1109.6887, so you need to convert between normalizations by multiplying by a factor of 2² = 4. The exponent of 2 comes from the fact that each element of pr1 or pr2 is the coefficient of a term like P_i rho P_i in which there are 2 Paulis.

So the following will give you what you're after:

import numpy as np
pr1 = np.array([0.2,0.3,0.2,0.3]) # understood to be in basis of norm=2 Paulis
pr2 = np.array([0.5,0.4,0.05,0.05])

import pygsti
g1 = pygsti.tools.jamiolkowski_iso(4*np.diag(pr1), 'pp','pp')
g2 = pygsti.tools.jamiolkowski_iso(4*np.diag(pr2), 'pp','pp')
pygsti.gatetools.diamonddist(g1,g2,mxBasis='pp')

This gives the 0.8 you expect.

from pygsti.

Thanks, figured it was some different normalization issue.

from pygsti.