fix: random circuit codes are non-CSS so return iscss(...) = false instead of nothing

[Fe-r-oz]

Hi, @royess, The random_circuit codes are non-CSS so, I am wondering why the iscss(::Type{...}) returns nothing. Shouldn't it return false? Hopefully, that is not a bad proposition?

Could you explain if it was a typo or there is some reason behind it? Thanks! I agree with you that codes are non-CSS.

BP and BP-OSD in PyQDecoders.jl (PyMatchingDecoder ) are not applicable to random circuit codes because they are non-CSS.

QuantumClifford.jl/src/ecc/codes/random_circuit.jl

Line 19 in 0d13791

iscss(::Type{CircuitCode}) = nothing

Bravyi introduces a Mixed Integer Programming for CSS/QLDPC codes so his approach don't work for non-CSS codes. Hence, I want to exclude non-CSS codes so user does not end up calculating distance for non-CSS codes which throw solver error.

iscss(::Type{...}) = false would be helpful here because in a later PR, I want to do something like

function distance(c::AbstractECC; kwargs...)
    if !iscss(c)
        throw(ArgumentError("Computing the minimum distance of QLDPC using Mixed Integer Programming works only for CSS codes."))
    end
    return distance(parity_checks(c); kwargs...)
end

@Krastanov, Please help review this PR, Thank you!

• The code is properly formatted and commented.
• Substantial new functionality is documented within the docs.
• All new functionality is tested.
• All of the automated tests on github pass.
• We recently started enforcing formatting checks. If formatting issues are reported in the new code you have written, please correct them.

[royess]

@Krastanov and I discussed this during PR. The reason is that random circuit codes are not necessarily CSS but could still possibly be CSS.

[Fe-r-oz]

There are a couple of instances, such as in bullet point 2 of the notes in the introductory message and in the line I quoted above, where you mention that these codes are non-CSS. I also reviewed the comments and all your messages posted in the PR.

I will let Stefan decide the fate of this PR.

[royess]

Let me clarify more: CircuitCode's are non-CSS in general. (That is what I meant in the sentences you quoted.) But some instances of them could be CSS.

Therefore, iscss(::Type{CircuitCode}) is undetermined, for which we assign nothing. If we return false all the time, we will make mistakes in these CSS instances.

Regarding your later PR: Does it make sense that you also exclude nothing when trying to calculate the code distance?

[Fe-r-oz]

Thanks! Yeah, I can try checking !iscss(...) and also check if nothing which can also work and also does not change anything here. The PR is #439, it's a previous PR, but I ended up saying it's a later PR.

I will close this PR to be safe.