Prepare 0.20.0 release #7828

mtreinish · 2022-03-28T21:56:21Z

Summary

To prepare for the upcoming 0.20.0 release this commit moves all the
release notes into a 0.20 subdirectory to keep them separate from future
development notes post-release. It also adds a release note with the
prelude section providing the high level overview of the release. When
this merges it should be the commit used to tag the 0.20.0 release.

Details and comments

Wait for the remaining open 0.20 PRs to merge or be deferred
Finish editing and fixing release notes

coveralls · 2022-03-28T22:25:52Z

Pull Request Test Coverage Report for Build 2073151203

0 of 0 changed or added relevant lines in 0 files are covered.
3 unchanged lines in 1 file lost coverage.
Overall coverage decreased (-0.005%) to 83.696%

Files with Coverage Reduction	New Missed Lines	%
qiskit/pulse/library/waveform.py	3	89.36%

Totals
Change from base Build 2073149689:	-0.005%
Covered Lines:	53970
Relevant Lines:	64483

💛 - Coveralls

tox.ini

releasenotes/notes/0.20/Operation-abstract-base-class-99f46180c7ddaa6b.yaml

…asses In Qiskit#7762 the scheduling passes in the transpiler were completely rewritten to operate in a model where scheduling and alignment happen as metadata on the circuit until a single padding pass is called to adapt the circuit to fit the scheduling metadata. As part of that rework a new pass ConstrainedReschedule was added to perform the alignment for measurement as well as all instructions via this new mechanism. However the way the migration was done was a breaking change and should have been additive so that we can deprecate the old path in 0.21 and allow users time to migrate to the new approach. Additionally as part of Qiskit#7762 the AlignMeasures pass was deprecated and its implementation was replaced with just returning an equivalent ConstraintedReschedule pass. This was problematic for two reasons though, first the two passes were not equivalent and this a breaking change and second it violates the deprecation policy. For the breaking change the AlignMeasures and ConstrainedReschedule while performing the same function do it in a different manner. I assume the deprecation was done this way because AlignMeasures is incompatible with the While the scheduling pass changes were also breaking to do the migration gracefully would require duplicating their functionality and deprecating the old This is also violating the deprecation policy which is defined here: https://qiskit.org/documentation/contributing_to_qiskit.html#deprecation-policy as the alternative wasn't available for a release prior to the start of the deprecation (see 2a in the above link). This commit adds back the breaking classes and renames the new scheduling classes ALAPScheduleAnalysis and ASAPScheduleAnalysis to differentiate them from the previous implementations. This also reverts the deprecation and changes to the AlignMeasures pass until we can do it correctly at the appropriate time. The class is restored to its previous state but instead a PendingDeprecationWarning is now emitted to warn users of the pending deprecation and removal of the class. Additionally, a warning in the release notes will be added as part of Qiskit#7828 to document the incompatibility with the new behavior of the scheduling and alignment passes.

mtreinish · 2022-03-30T16:42:16Z

I've finished a first pass through the release notes. I still need to review it all again both for typos I likely introduced and any notes I might have missed (along with waiting for the rest of the pending release PRs to merge).

jakelishman · 2022-03-30T17:26:52Z

Ok, I'll finish going through them now - I have a few changes from the ones I looked at last night, but nothing major (obviously).

releasenotes/notes/0.20/prepare-0.20-79918ed0fc5b496e.yaml

releasenotes/notes/0.20/primitives-fb4515ec0f4cbd8e.yaml

jakelishman

I left my suggested changes in mtreinish#17

releasenotes/notes/0.20/Operation-abstract-base-class-99f46180c7ddaa6b.yaml

…asses (#7835) * Revert deprecation and breaking changes of scheduling and alignment passes In #7762 the scheduling passes in the transpiler were completely rewritten to operate in a model where scheduling and alignment happen as metadata on the circuit until a single padding pass is called to adapt the circuit to fit the scheduling metadata. As part of that rework a new pass ConstrainedReschedule was added to perform the alignment for measurement as well as all instructions via this new mechanism. However the way the migration was done was a breaking change and should have been additive so that we can deprecate the old path in 0.21 and allow users time to migrate to the new approach. Additionally as part of #7762 the AlignMeasures pass was deprecated and its implementation was replaced with just returning an equivalent ConstraintedReschedule pass. This was problematic for two reasons though, first the two passes were not equivalent and this a breaking change and second it violates the deprecation policy. For the breaking change the AlignMeasures and ConstrainedReschedule while performing the same function do it in a different manner. I assume the deprecation was done this way because AlignMeasures is incompatible with the While the scheduling pass changes were also breaking to do the migration gracefully would require duplicating their functionality and deprecating the old This is also violating the deprecation policy which is defined here: https://qiskit.org/documentation/contributing_to_qiskit.html#deprecation-policy as the alternative wasn't available for a release prior to the start of the deprecation (see 2a in the above link). This commit adds back the breaking classes and renames the new scheduling classes ALAPScheduleAnalysis and ASAPScheduleAnalysis to differentiate them from the previous implementations. This also reverts the deprecation and changes to the AlignMeasures pass until we can do it correctly at the appropriate time. The class is restored to its previous state but instead a PendingDeprecationWarning is now emitted to warn users of the pending deprecation and removal of the class. Additionally, a warning in the release notes will be added as part of #7828 to document the incompatibility with the new behavior of the scheduling and alignment passes. * More release note fixes * Fix lint * Fix docs * Fix doc whitespace * Update dynamical decoupling release note * Copy docstrings from pre-refactor to fix docs build * Add back test coverage for legacy scheduling and alignment passes * Update releasenotes/notes/update-instruction-alignment-passes-ef0f20d4f89f95f3.yaml Co-authored-by: Naoki Kanazawa <nkanazawa1989@gmail.com> * Fix release notes * Rename base scheduler class in legacy path This commit renames the BaseScheduler class in the legacy scheduling path to BaseSchedulerTransform to differentiate it from the new path scheduling paths. It also documents the pending deprecation in all the legacy scheduling passes to point people to the new scheduling workflow. * Rename DynamicalDecouplingPadding to PadDynamicalDecoupling Co-authored-by: Naoki Kanazawa <nkanazawa1989@gmail.com> Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

To prepare for the upcoming 0.20.0 release this commit moves all the release notes into a 0.20 subdirectory to keep them separate from future development notes post-release. It also adds a release note with the prelude section providing the high level overview of the release. When this merges it should be the commit used to tag the 0.20.0 release.

Co-authored-by: Ali Javadi-Abhari <ajavadia@users.noreply.github.com>

jakelishman

Looks good to me. After the merging of the PR I'd made, I was happy with all the release notes already.

kevinhartman · 2022-03-31T19:03:38Z

...senotes/notes/0.20/add-v2-backend-support-in-transpiler-parse-inst-map-a617801850178d05.yaml

+fixes:
+  - |
+    Fixed support in :func:`~qiskit.compiler.transpile` for passing a
+    :class:`~.InstructionScheduleMap` to object to the underlying


Suggested change

:class:`~.InstructionScheduleMap` to object to the underlying

:class:`~.InstructionScheduleMap` object to the underlying

kevinhartman · 2022-03-31T19:05:23Z

releasenotes/notes/0.20/add-xxminusyy-gate-63e6530c23500de9.yaml

+---
+features:
+  - |
+    Added a new gate gate class :class:`~qiskit.circuit.library.XXMinusYYGate`


Suggested change

Added a new gate gate class :class:`~qiskit.circuit.library.XXMinusYYGate`

Added a new gate class :class:`~qiskit.circuit.library.XXMinusYYGate`

kevinhartman · 2022-03-31T19:30:39Z

releasenotes/notes/0.20/update-instruction-alignment-passes-ef0f20d4f89f95f3.yaml

+    only updates the property set of the pass manager, in which new property set
+    ``node_start_time`` is created there to perform these operation on nodes
+    scheduled on the absolute time. This new workflow is both more efficient and can correct
+    for additional timing constraints exposed by a backend.


Suggested change

only updates the property set of the pass manager, in which new property set

``node_start_time`` is created there to perform these operation on nodes

scheduled on the absolute time. This new workflow is both more efficient and can correct

for additional timing constraints exposed by a backend.

only update the property set of the pass manager, specifically new property set item

``node_start_time``, which holds the absolute start time of each opnode. A separate

:class:`~.TransformationPass` such as :class:`~.PadDelay` is subsequently used

to apply scheduling to the DAG. This new workflow is both more efficient and can

correct for additional timing constraints exposed by a backend.

kevinhartman · 2022-03-31T19:32:23Z

releasenotes/notes/0.20/update-instruction-alignment-passes-ef0f20d4f89f95f3.yaml

+
+    For those who are creating custom :class:`~.PassManager` objects that involve
+    circuit scheduling you will need to adjust your :class:`~.PassManager`
+    to include insert one of the :class:`~.BasePadding` passes (currently


Suggested change

to include insert one of the :class:`~.BasePadding` passes (currently

to insert one of the :class:`~.BasePadding` passes (currently

Cryoris

Just some minor typos, otherwise LGTM!

Cryoris · 2022-03-31T19:35:24Z

releasenotes/notes/0.20/add-sparsepauliop-equiv-7a8a1420117dba21.yaml

+    Added a new method :meth:`~qiskit.quantum_info.SparsePauliOp.equiv` to the
+    :class:`~.SparsePauliOp` class for testing the equivalence of a
+    :class:`~.SparsePauliOp` with another :class:`.SparsePauliOp` object.
+    Unlike the ``==`` operator which compares operartors element-wise,


Suggested change

Unlike the ``==`` operator which compares operartors element-wise,

Unlike the ``==`` operator which compares operators element-wise,

Cryoris · 2022-03-31T19:36:08Z

releasenotes/notes/0.20/add-xxminusyy-gate-63e6530c23500de9.yaml

+---
+features:
+  - |
+    Added a new gate gate class :class:`~qiskit.circuit.library.XXMinusYYGate`


Suggested change

Added a new gate gate class :class:`~qiskit.circuit.library.XXMinusYYGate`

Added a new gate class :class:`~qiskit.circuit.library.XXMinusYYGate`

Ah I just saw Kevin beat me to this correction 😄

Cryoris · 2022-03-31T19:37:27Z

releasenotes/notes/0.20/cleanup-deprecated-circuitmeth-89edb244f572b754.yaml

+    :mod:`qiskit.circuit.library` has been removed. It was originally deprecated in the
+    0.16.0 release. Instead the :class:`~qiskit.circuit.library.GMS` class should be used, as
+    this allows you to create an equivalent 2 qubit MS gate in addition to
+    an MSGate for any number of qubits.


Suggested change

an MSGate for any number of qubits.

an ``MSGate`` for any number of qubits.

Cryoris · 2022-03-31T19:38:57Z

releasenotes/notes/0.20/dense-layout-target-aware-2b330ccee948d31a.yaml

+    The :class:`~qiskit.transpiler.passes.DenseLayout` pass has a new keyword
+    argument on its constructor, ``target``. This argument is used to specify a
+    :class:`~qiskit.transpiler.Target` object representing the compilation
+    target for the pass. If it is specified it superscedes the other arguments


I'm not a 100% sure, but is there a c too much?

Suggested change

target for the pass. If it is specified it superscedes the other arguments

target for the pass. If it is specified it supersedes the other arguments

mtreinish · 2022-03-31T19:58:02Z

Thanks for the suggestions, given the CI time required to apply them I'm going to save these changes fix them manually in the hosted version when we push the release and open up a follow up PR after the release is tagged and backport them for 0.20.1

mtreinish added the on hold Can not fix yet label Mar 28, 2022

mtreinish added this to the 0.20 milestone Mar 28, 2022

mtreinish requested a review from a team as a code owner March 28, 2022 21:56

mtreinish changed the title ~~Prepare 0.20.0 release~~ [WIP] Prepare 0.20.0 release Mar 28, 2022

jakelishman self-assigned this Mar 29, 2022

mtreinish mentioned this pull request Mar 29, 2022

Revert deprecation and breaking changes of scheduling and alignment passes #7835

Merged

mtreinish requested review from ikkoham and t-imamichi as code owners March 29, 2022 21:17

mtreinish commented Mar 29, 2022

View reviewed changes

tox.ini Outdated Show resolved Hide resolved

eliarbel reviewed Mar 30, 2022

View reviewed changes

releasenotes/notes/0.20/Operation-abstract-base-class-99f46180c7ddaa6b.yaml Outdated Show resolved Hide resolved

ajavadia reviewed Mar 31, 2022

View reviewed changes

releasenotes/notes/0.20/prepare-0.20-79918ed0fc5b496e.yaml Outdated Show resolved Hide resolved

releasenotes/notes/0.20/prepare-0.20-79918ed0fc5b496e.yaml Outdated Show resolved Hide resolved

releasenotes/notes/0.20/primitives-fb4515ec0f4cbd8e.yaml Outdated Show resolved Hide resolved

jakelishman reviewed Mar 31, 2022

View reviewed changes

releasenotes/notes/0.20/Operation-abstract-base-class-99f46180c7ddaa6b.yaml Outdated Show resolved Hide resolved

mtreinish mentioned this pull request Mar 31, 2022

DNM: Test wheel jobs pre-release #7848

Closed

mtreinish added 10 commits March 31, 2022 13:22

Move and update new release notes

9c139f1

Fix typo

5cfa21f

Update more release notes

95cbcba

Move and update new release note

cac7ce2

More release note updates

d98b1e6

Drop operation release note

efe4e31

Add missing release notes

cc1584a

Fix one more release note

4e36275

Fix upgrade and deprecation notes

40d0d5b

mtreinish and others added 10 commits March 31, 2022 13:22

Remove duplicate release notes from 0.19.x

4d0e84f

Remove Operation from doc toctree

19e6419

Fix more bug fix relaese notes

fe03e6a

Finish first pass over release notes

c6fc57c

Add link to rust in prelude

33438c1

Move and update new release notes

bfecdd8

Apply suggestions from code review

f3ec376

Co-authored-by: Ali Javadi-Abhari <ajavadia@users.noreply.github.com>

Release note fixes

c7a35f4

Fix macOS arm64 wheel config

61bfb2a

Move and update new release notes

caa7acc

mtreinish force-pushed the prepare-0.20 branch from 033d057 to caa7acc Compare March 31, 2022 17:50

mtreinish added Changelog: None Do not include in changelog and removed on hold Can not fix yet labels Mar 31, 2022

mtreinish changed the title ~~[WIP] Prepare 0.20.0 release~~ Prepare 0.20.0 release Mar 31, 2022

Revert tox.ini change

0ba0592

jakelishman approved these changes Mar 31, 2022

View reviewed changes

mtreinish added the automerge label Mar 31, 2022

Merge branch 'main' into prepare-0.20

770c7dd

kevinhartman reviewed Mar 31, 2022

View reviewed changes

Cryoris reviewed Mar 31, 2022

View reviewed changes

mergify bot merged commit a2d13f5 into Qiskit:main Mar 31, 2022

kevinhartman mentioned this pull request Mar 31, 2022

Fix a few 0.20 release note typos. #7853

Merged

mtreinish deleted the prepare-0.20 branch April 25, 2022 21:42

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Prepare 0.20.0 release #7828

Prepare 0.20.0 release #7828

mtreinish commented Mar 28, 2022 •

edited

Loading

coveralls commented Mar 28, 2022 •

edited

Loading

mtreinish commented Mar 30, 2022

jakelishman commented Mar 30, 2022

jakelishman left a comment

jakelishman left a comment

kevinhartman Mar 31, 2022

kevinhartman Mar 31, 2022

kevinhartman Mar 31, 2022

kevinhartman Mar 31, 2022

Cryoris left a comment

Cryoris Mar 31, 2022

Cryoris Mar 31, 2022

Cryoris Mar 31, 2022

Cryoris Mar 31, 2022

Cryoris Mar 31, 2022

mtreinish commented Mar 31, 2022

	:class:`~.InstructionScheduleMap` to object to the underlying
	:class:`~.InstructionScheduleMap` object to the underlying

	Added a new gate gate class :class:`~qiskit.circuit.library.XXMinusYYGate`
	Added a new gate class :class:`~qiskit.circuit.library.XXMinusYYGate`

-    only updates the property set of the pass manager, in which new property set
-    ``node_start_time`` is created there to perform these operation on nodes
-    scheduled on the absolute time. This new workflow is both more efficient and can correct
-    for additional timing constraints exposed by a backend.
+    only update the property set of the pass manager, specifically new property set item
+    ``node_start_time``, which holds the absolute start time of each opnode. A separate
+    :class:`~.TransformationPass` such as :class:`~.PadDelay` is subsequently used
+    to apply scheduling to the DAG. This new workflow is both more efficient and can
+    correct for additional timing constraints exposed by a backend.

	to include insert one of the :class:`~.BasePadding` passes (currently
	to insert one of the :class:`~.BasePadding` passes (currently

	Unlike the ``==`` operator which compares operartors element-wise,
	Unlike the ``==`` operator which compares operators element-wise,

	an MSGate for any number of qubits.
	an ``MSGate`` for any number of qubits.

	target for the pass. If it is specified it superscedes the other arguments
	target for the pass. If it is specified it supersedes the other arguments

Prepare 0.20.0 release #7828

Prepare 0.20.0 release #7828

Conversation

mtreinish commented Mar 28, 2022 • edited Loading

Summary

Details and comments

coveralls commented Mar 28, 2022 • edited Loading

Pull Request Test Coverage Report for Build 2073151203

💛 - Coveralls

mtreinish commented Mar 30, 2022

jakelishman commented Mar 30, 2022

jakelishman left a comment

Choose a reason for hiding this comment

jakelishman left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Cryoris left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mtreinish commented Mar 31, 2022

mtreinish commented Mar 28, 2022 •

edited

Loading

coveralls commented Mar 28, 2022 •

edited

Loading