-
Notifications
You must be signed in to change notification settings - Fork 368
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Prepare for Qiskit 1.0 #2047
Prepare for Qiskit 1.0 #2047
Conversation
This modifies the dependency-resolution steps to do their job in a single step; Aer should have no dependencies pinning Qiskit, so we can use the preferable form more eagerly. This also adds a test run to the PR CI for Qiskit 1.0.0rc1 (or rc2, etc) that always runs, so Aer doesn't regress. Co-authored-by: Luciano Bello <bel@zurich.ibm.com>
This commit *does not* touch the primitives, which appear to be broken at a relatively basic level with Qiskit 1.0 - hopefully the primitives team can fix that separately. This modifies the test suite to work with both Qiskit 0.46 and 1.0.0rc1 simultaneously. Unfortunately, some of the new-style fake backends aren't in 0.46, so I just put in some import switches for my ease here. A few tests, particularly in the noise model sections, were asserting specific things about the properties/calibrations of Terra's fake backends, which was quite fragile already, so in those particular cases I replaced them with a manually constructed `Target` with properties entirely under Aer's control.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I deleted this file, because the only tests in it are circular logic - the old FakeBackend.run
method also did this same thing, so the test wasn't doing much.
This should pass the full suite once #2048 is merged. |
a7e80f0
to
70d8b83
Compare
This will need the branch-protection rules changing to accommodate it; the names of the tests have changed. The auto-generated ones would have changed anyway because of how the test matrix is build now, so since they had to change, I gave them slightly more descriptive names. |
* Resolve CI dependencies in a single step and add qiskit==1.0.0rc1 job This modifies the dependency-resolution steps to do their job in a single step; Aer should have no dependencies pinning Qiskit, so we can use the preferable form more eagerly. This also adds a test run to the PR CI for Qiskit 1.0.0rc1 (or rc2, etc) that always runs, so Aer doesn't regress. Co-authored-by: Luciano Bello <bel@zurich.ibm.com> * Fix test suite to work with Qiskit<1.0 and >=1.0 This commit *does not* touch the primitives, which appear to be broken at a relatively basic level with Qiskit 1.0 - hopefully the primitives team can fix that separately. This modifies the test suite to work with both Qiskit 0.46 and 1.0.0rc1 simultaneously. Unfortunately, some of the new-style fake backends aren't in 0.46, so I just put in some import switches for my ease here. A few tests, particularly in the noise model sections, were asserting specific things about the properties/calibrations of Terra's fake backends, which was quite fragile already, so in those particular cases I replaced them with a manually constructed `Target` with properties entirely under Aer's control. * Run black --------- Co-authored-by: Luciano Bello <bel@zurich.ibm.com>
Summary
This PR sorts out CI to test against the current release of Qiskit and the 1.0.0 release candidates in PR CI, and updates the Qiskit
main
branchDetails and comments
This replaces #2046 - I ended up with more changes than makes sense to tack onto the end of that PR.
I haven't touched the primitives, so they will fail against Terra 1.0.0rc1 right now. I'm not certain what the scope of changes are needed there to support Qiskit 0.45/0.46 and Qiskit 1.0+, but hopefully the primitives team can work on that.