W node

[RazinShaikh]

No description provided.

[jvdwetering]

I don't really see where you are changing the VertexType and EdgeType classes to allow for W-like things. Is this a change you are making in a local copy of pyzx?

[RazinShaikh]

It's in my fork here: https://github.com/RazinShaikh/pyzx/tree/non-flexsymmetric

[RazinShaikh]

@mark-koch Thanks a lot for your comments. I have gone through each of them and incorporated your suggestions. Let me know if it looks good to you.

[jvdwetering]

So just checking that this is desired behaviour:
I made the following diagram

And then when I 'fuse W nodes' I get the following:

Is that correct? I would have expected no self-loop. Is this because a W must have an output? If so, did I make something that is not well-typed?

[jvdwetering]

Also, when you add new W-spiders, depending on where you click the orientation of the W-spider changes, sometimes in weird ways:

[jvdwetering]

I don't understand the bug I outlined above, I tried with two different configurations:

And the left one rewrites correctly with fuse W nodes, but the right one doesn't:

[RazinShaikh]

Hi John, that is indeed not the right behaviour. For instance, if you drag the nodes, they will fuse correctly. I will check later today and see what's going wrong when you press the fuse button.

[RazinShaikh]

@jvdwetering the fusion bug should be fixed by this PR zxcalc/pyzx#129 in PyZX.

[RazinShaikh]

Also, when you add new W-spiders, depending on where you click the orientation of the W-spider changes, sometimes in weird ways:

Fixed in 306754b. The fix for this is to call set_vitem_rotation() when a graph is updated. It should always have the correct orientation now.

[RazinShaikh]

@mark-koch @jvdwetering are you guys happy for this to be merged?

[jvdwetering]

It looks like you got a new conflict because Boldi fixed some mypy bugs, but once that's fixed I think this can be merged in.