Remove retired VQE runner

[timmysilv]

sadge

the CI error: qiskit_ibm_runtime.exceptions.IBMInputValueError: 'The vqe and qaoa programs have been retired in the Qiskit Runtime service. Please visit https://qiskit.org/ecosystem/ibm-runtime for an introduction on Sessions and Primitives, and to access tutorials on how to execute VQE and QAOA using Qiskit Runtime Primitives.'

too much effort to do any more than this.

PS: I pinned the maximum Qiskit version because this test fails with Qiskit 0.43, saying qiskit.exceptions.QiskitError: "Error decomposing node of instruction 'rx': 'NoneType' object has no attribute 'global_phase'. Unable to define instruction 'u' in the basis."

^ EDIT: just changed the test to transpile with other options instead

[codecov]

Codecov Report

Merging #313 (1768781) into master (76be1e1) will not change coverage.
The diff coverage is n/a.

@@            Coverage Diff            @@
##            master      #313   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            7         7           
  Lines          315       314    -1     
=========================================
- Hits           315       314    -1     

Impacted Files	Coverage Δ
pennylane_qiskit/__init__.py	100.00% <ø> (ø)

[rmoyard]

Suggested change

	"qiskit>=0.32,<=0.42.1",
	"qiskit>=0.42.1",

Should it not be that?

[trbromley]

It's to prevent this issue in a test:

qiskit.exceptions.QiskitError: "Error decomposing node of instruction 'rx': 'NoneType' object has no attribute 'global_phase'. Unable to define instruction 'u' in the basis."

Though @timmysilv, have you tried fixing the issue?

[timmysilv]

i have not, but I figured this week isn't the best time to do it. i'll give it a shot today

[timmysilv]

turns out Qiskit/qiskit#9343 changed the behaviour of the "unroller" transpilation method option to account for target ops, and this broke RX badly. the intention of the test is only to show that we can pass custom transpilation options, so I've changed it to just use the "translator" option! those should pass with the latest qiskit (confirmed locally)

[trbromley]

Thanks @timmysilv! Overall looks good to me. We can come back to deciding what we want to do here during the 0.33 cycle.

[trbromley]

It's to prevent this issue in a test:

qiskit.exceptions.QiskitError: "Error decomposing node of instruction 'rx': 'NoneType' object has no attribute 'global_phase'. Unable to define instruction 'u' in the basis."

Though @timmysilv, have you tried fixing the issue?

[rmoyard]

Can you run the IBMQ integration test from this PR?

[timmysilv]

passed: https://github.com/PennyLaneAI/pennylane-qiskit/actions/runs/5338887808

[rmoyard]

Thanks 👍