Adding Tutorials for Circuit Renderers

[gadhvirushiraj]

New Renderer Tutorials

GSOC'24 Project - QuTiP: Enhanced Quantum Circuit Diagrams

✅ TextRenderer Tutorial
✅ MatRenderer Tutorial

[BoxiLi]

Why does this not trigger any tests... I remember there were tests waiting for approval when it was created. Now they disappear.

[BoxiLi]

@gadhvirushiraj Could you maybe push something, e.g. add a qutip.about() as the last cell just to try triggering the tests again?

[gadhvirushiraj]

Is qc.add_1q_gate() still supported to add gates across all present qubits in circuit? It used to work earlier, though showing some error now. I have removed it from current tutorial examples.

[BoxiLi]

It should be supported, but I did remember making some changes on that a few weeks ago. Maybe something went wrong. What was the error?

[nwlambert]

@gadhvirushiraj if you update your branch the tests should now run. some of the plotting stuff in old qip plotting examples now fail and need updating to the new draw() method. I tried in the other PR but these produced additional issues, so maybe better you check here (one had trouble parsing a latex label, and the other failed completely).

[gadhvirushiraj]

This should work after qutip/qutip-qip#257 merge.

[nwlambert]

these look good now. thanks rushi. for the v4 failing test, maybe we need to fix the numpy test environment to 1.20 or lower?

[nwlambert]

Sorry rushi, that seemed to make things worse. can you revert that change to the numpy version, and I will see if I can produce the original problem locally.

[gadhvirushiraj]

Sorry rushi, that seemed to make things worse. can you revert that change to the numpy version, and I will see if I can produce the original problem locally.

@nwlambert I have reverted the changes, but I previously verified that np.complex (used internally) has been deprecated starting with version 1.20, as you mentioned. I believe we also need to use python 3.9 to ensure compatibility with numpy. https://numpy.org/doc/2.0/release/1.21.2-notes.html

[nwlambert]

I think we can just fix it by using the right dtype in the notebook, mat = np.zeros((4, 4), dtype=complex) on line 421 in quantum-gates.md

[nwlambert]

latest jax release (0.4.36) breaks something in qutip-jax, but v4 tests are finally running, so will merge