274 building xppauli classes (part of 257) #281
Conversation
| @@ -104,6 +108,8 @@ def __init__( | |||
| matrix = nmatrix | |||
|
|
|||
| self.matrix = matrix | |||
| self.precision = precision | |||
| # TODO should _num_paulis be renamed to _num_xppaulis in this and derived classes? | |||
There was a problem hiding this comment.
As we are separating the Pauli and XPPauli classes I would say it is better to change the name as you suggest.
| BaseXPPauli: _description_ | ||
| """ | ||
|
|
||
| assert a.precision == b.precision, QiskitError( |
There was a problem hiding this comment.
This should not be in a "_"-method but a higher method compose which probably belongs in XPPauli and XPPauliList (different for each). Assertions have the nice property that they have zero impact with DEBUG is set correctly but some people feel that they should not be used for format requirements and are more to detect algorithm and internal faults.
There was a problem hiding this comment.
See previous comment.
Replaced assert with if...raise.
| """(TODO improve doc): This is the equivalent of XPRound from Mark's | ||
| code. It converts the XPPauli operator into unique vector form, ie | ||
| phase_exp in Z modulo 2*precision, x in Z_2, z in Z modulo | ||
| precision.""" |
There was a problem hiding this comment.
Add explicit examples ...
| return np.sum(np.logical_or(self.x, self.z), axis=-1) | ||
|
|
||
| def is_diagonal(self): | ||
| """_summary_""" |
qiskit_qec/operators/random.py
Outdated
| @@ -196,6 +197,38 @@ def random_clifford(num_qubits, seed=None): | |||
| return Clifford(StabilizerTable(table, phase)) | |||
|
|
|||
|
|
|||
| def random_xppauli(num_qubits, precision=None, seed=None): | |||
| Args: | ||
| num_qubits (int): the number of qubits. | ||
| precision (int): Precision of XP operators. Must be an integer | ||
| greater than or equal to 2. |
| # TODO | ||
| pass | ||
|
|
||
| def compose(self, other, qargs=None, front=False, inplace=False): |
| from Mark's code. It returns the count of qubits where either z or x | ||
| component is nonzero.""" | ||
| # TODO Since 'distance' has a specific meaning in QECCs, for now, the | ||
| # name 'weight' has been used for this function. |
There was a problem hiding this comment.
Agree that weight is better than 'distance'
…ppaulilist 3. added test
|
In addition to implementing review comments, I've also added tests for |
ghost
merged commit 
Summary
Partial fix for #274 (part of #257 ). This PR contains an initial implementation of
XPPauliandXPPauliListclasses.Details and comments
This is an initial implementation of XP operator algebra (from Mark Webster's package ) for
XPPauliandXPPauliList. This is a work in progress.XPPauliandXPPauliListand associated test files.random_xppauliin operators/random.pyBaseXPPauli(XP* are function names from Mark's code, name in brackets is the new name):XPSetN(rescale_precision)XPDistance(weight)XPisDiag(is_diagonal)XPDegree(degree)XPPower(power)XPMul(compose)XPRound(unique_vector_rep)XPD(antisymmetric_op)XPSetNSingle,XPSquare,XPdeg).XPPauliandXPPauliList. Some tests are yet to be added forXPPauliListfor above implemented functions.Highly appreciate your feedback!