Prevent Hamiltonians that share wires from being multiplied together

[trbromley]

Raises a ValueError if Hamiltonians sharing wires are multiplied together:

h1 = qml.Hamiltonian([1, 1], [qml.PauliZ(0), qml.PauliZ(1)])
h1 @ h1

This will raise a ValueError.

Fixes #1272 although this is just to help prevent users getting caught-out, we still need a better solution for manipulation of Hamiltonians.

[github-actions]

Hello. You may have forgotten to update the changelog!
Please edit .github/CHANGELOG.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 #1273 (0bc2bc4) into master (8cf8475) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #1273   +/-   ##
=======================================
  Coverage   98.13%   98.13%           
=======================================
  Files         147      147           
  Lines       11256    11260    +4     
=======================================
+ Hits        11046    11050    +4     
  Misses        210      210           

Impacted Files	Coverage Δ
pennylane/vqe/vqe.py	92.97% <100.00%> (+0.15%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 8cf8475...0bc2bc4. Read the comment docs.

[glassnotes]

Looks good! 👍 In the long term, maybe @ can call grouping.pauli_mult for cases like this (though I think there would be some things to iron out in terms of wire maps -- perhaps @josh146 can speak further to this, he pointed it out while reviewing #1181).

[josh146]

Looks good! 👍 In the long term, maybe @ can call grouping.pauli_mult for cases like this (though I think there would be some things to iron out in terms of wire maps -- perhaps @josh146 can speak further to this, he pointed it out while reviewing #1181).

Alternatively, is there anything stopping us from having __matmul__ call grouping.pauli_mult in the short term? 🙂 It might solve a lot of open issues we have on operator multiplication

[anthayes92]

Nice short term solution! A long term fix for this would render 1213 obselete.

[glassnotes]

Looks good! +1 In the long term, maybe @ can call grouping.pauli_mult for cases like this (though I think there would be some things to iron out in terms of wire maps -- perhaps @josh146 can speak further to this, he pointed it out while reviewing #1181).

Alternatively, is there anything stopping us from having __matmul__ call grouping.pauli_mult in the short term? slightly_smiling_face It might solve a lot of open issues we have on operator multiplication

Which __matmul__, the one in the Hamiltonian class? I think as long as a consistent wire map is chosen it should be fine (and we could make a like a cumulative wire map using the shared wires of all the terms, or even make it an argument to the Hamiltonian itself.).