Add Harmony backend for IonQ

[Cynocracy]

No description provided.

[codecov]

Codecov Report

Merging #49 (53ea8cb) into master (f69916a) will decrease coverage by 3.25%.
The diff coverage is 84.61%.

@@            Coverage Diff             @@
##           master      #49      +/-   ##
==========================================
- Coverage   88.77%   85.51%   -3.26%     
==========================================
  Files           5        5              
  Lines         285      290       +5     
==========================================
- Hits          253      248       -5     
- Misses         32       42      +10     

Impacted Files	Coverage Δ
pennylane_ionq/device.py	89.90% <75.00%> (-3.37%)	⬇️
pennylane_ionq/__init__.py	100.00% <100.00%> (ø)
pennylane_ionq/api_client.py	91.19% <100.00%> (-3.78%)	⬇️

---

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 f69916a...53ea8cb. Read the comment docs.

[albi3ro]

Hi @Cynocracy and thanks for opening a PR.

Could you provide some information about what you are wanting to achieve with this PR?

You may also need to upgrade your version of black and use the -l 100 line length specification. In the most recent version, the formatting of a ** b changed to a**b.

[Cynocracy]

Thank you! Will update and do another formatting pass

[Cynocracy]

Also Hi :) Nice to meet you!

[albi3ro]

Was something about qml.device('ionq.qpu', target='qpu.harmony', wires=wires) not working?

Does this PR just make it slightly easier to create, fixing something that was broken, or add a new feature?

[Cynocracy]

That method works fine for now, but we're working on new QPU backend targets, and thought it would map more cleanly into pennylane by adding support here.

Are there any reasons to prefer having support in Pennylane itself for different sized QPU backends?

If not, I can scrap this PR and keep with the target="" approach

[Cynocracy]

Tested out that arg, it doesn't seem like we're able to send it through the qml device like you mentioned.

I would think that would work for us though, perhaps it's just some missing wiring?

[Cynocracy]

Cut #50 as an alternative to this approach

[josh146]

@Cynocracy let us know when this PR (and also #50) are ready for review :)

[Cynocracy]

Aye aye! I don't mind closing this in favor of #50 will comment over there when ready for review, am not sure why the tests are failing.