Add more interesting algorithms

[BoxiLi]

qutip-qip.algorithm remains mostly empty, it has only a qft algorithm, it would be very interesting to include more common algorithms.

We should focus on algorithms that use fewer than 10 qubits, so that there might be a chance to simulate them with the pulse-level simulator, such as the repetition error correction code. The algorithms should have a unified format, all return a QubitCircuit object.

[rochisha0]

hi @BoxiLi , can i take up this issue?

[hodgestar]

@rochisha0 Instead of asking whether you can take on an issue, it would be better to think a bit about the issue and either ask a question about what needs to be done, or propose some specific work you'd like to do and ask for feedback.

An example of the first might be: "What algorithm would you suggest adding next?"

An example of the second might be: "I'd like to add a VQE algorithm, what do you think?"

In open source projects issues are not typically earmarked for any one person. People contribute as capacity allows and interest dictates. You don't need permission to start work on an issue, but do get feedback on the approach early.

[BoxiLi]

Thanks, @hodgestar for the suggestions!

Indeed those issues are all open. Ideas or proposals/implementations in more detail are surely all welcome @rochisha0 ;)

[rochisha0]

Thanks @hodgestar I understand that some open-source repositories handle things differently by assigning a person to each issue to avoid overlapping efforts.

@BoxiLi I'll take some time to brainstorm and come up with ideas.

[nwlambert]

just to add, if you are familiar with some small examples of some basic error correction codes, those would be great! but just a suggestion.

[rochisha0]

@nwlambert thanks for the suggestion.

I read the qft code, it has qft implemented in 3 ways ["qft", "qft_steps", "qft_gate_sequence"].

Do we require all 3 for future algorithms or just return QubitCircuit object?

@BoxiLi

[rochisha0]

@BoxiLi any info on the same?

[BoxiLi]

Hey, sorry, I remember I replied to this somehow. Must have missed it.

I would say it is enough to have one function that returns the full Circuit object.