[WIP] Create a runtime OQD device that generates openapl

[paul0403]

Context:
After lowering the ion dialect to llvm dialect, and then llvm ir, with calls to the runtime capi stubs (#1466), the final llvm ir needs to be executed. The stubs will need to generate openapl, which is the common entry point for oqd backends.

Description of the Change:

• Note that since the lowering from ion dialect to llvm dialect ([WIP] Lowering Ion dialect to llvm dialect #1466 ) is not finished yet, we need to manually write an IR (and use replace_ir) for testing. Undo this and use the actual lowering passes to generate llvm once they are in place. As of now, simply run python oqd_run.py to see the effect of the stubs.
• Create a compilation pipeline for the OQD device. The pipeline needs to diverge from the standard pipeline after reaching the quantum dialect, and needs to go into quantum-to-ion, then ion-to-llvm.
• Fill the OQD capi stubs with openapl program generation.
• Isolate the oqd capi stubs into their own .so, instead of throwing them into the existing runtime capi's .so

Benefits:
Catalyst can now generate programs that can be accepted by OQD backends!

[sc-73710]

[github-actions]

Hello. You may have forgotten to update the changelog!
Please edit doc/releases/changelog-dev.md on your branch 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.

[paul0403]

changelog