Fix qml.counts measurement in new device

[eddddddy]

Context:
Part of the effort to integrate the new device API. Most measurements are supported with only a few weird edge cases (including counts).

Description of the Change:
Previous behaviour:

>>> dev = DefaultQubit2()
>>> qs1 = qml.tape.QuantumScript([qml.RY(1.4, wires=0)], [qml.counts(wires=[0, 1])], shots=15)
>>> dev.execute(qs1)
{'FalseFalse': 9, 'TrueFalse': 6}
>>> qs2 = qml.tape.QuantumScript([qml.RY(1.4, wires=0)], [qml.counts(qml.PauliZ(0))], shots=15)
>>> dev.execute(qs2)
{tensor(-1, requires_grad=True): 10, tensor(1, requires_grad=True): 5}

New behaviour:

>>> dev = DefaultQubit2()
>>> qs1 = qml.tape.QuantumScript([qml.RY(1.4, wires=0)], [qml.counts(wires=[0, 1])], shots=15)
>>> dev.execute(qs1)
{'00': 9, '10': 6}
>>> qs2 = qml.tape.QuantumScript([qml.RY(1.4, wires=0)], [qml.counts(qml.PauliZ(0))], shots=15)
>>> dev.execute(qs2)
{-1,: 10, 1: 5}

[github-actions]

Hello. You may have forgotten to update the changelog!
Please edit doc/releases/changelog-dev.md with:

• A one-to-two sentence description of the change. You may include a small working example for new features.
• A link back to this PR.
• Your name (or GitHub username) in the contributors section.

[codecov]

Codecov Report

Merging #4172 (b8a3ca4) into master (5d8ee2e) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #4172   +/-   ##
=======================================
  Coverage   99.77%   99.77%           
=======================================
  Files         340      340           
  Lines       30561    30566    +5     
=======================================
+ Hits        30491    30496    +5     
  Misses         70       70           

Impacted Files	Coverage Δ
pennylane/devices/qubit/sampling.py	100.00% <100.00%> (ø)
pennylane/measurements/counts.py	100.00% <100.00%> (ø)

[mudit2812]

Looks great. Thanks @eddddddy ! 🚀

[albi3ro]

🎉

Just changelog and unit tests for CountsMP.process_samples.

Actually not seeing any unit tests for CountsMP.process_samples in test_counts.py, but this might be a good opportunity to change that. Even simple stuff like:

samples = np.zeros((4,2))
output = CountsMP(wires=(0,1)).process_samples(samples, wire_order=(0,1))
assert output == {'00': 4}

or something. probably messed up the orders, but you know what I mean anyway.

[eddddddy]

@albi3ro There are unit tests in test_counts.py, but those tests use integer arrays for the samples instead of boolean arrays. Isn't this behaviour new for DefaultQubit2? Should I change the previous unit tests to accept boolean arrays instead?

[albi3ro]

Thanks for the added tests. Looks great 🎉

[eddddddy]

[sc-38218]