feat!: Allow array arguments to load_pytket

[tatiana-s]

Closes #671

Allows you to load pytket circuits as function definitions that use arrays as both inputs and outputs:

circ = Circuit(2)
circ.measure_all()

guppy.load_pytket("guppy_circ", circ)

@guppy
def foo(q: array[qubit, 2]) -> array[qubit, 2]:
    return guppy_circ(q)

Note that:

• If using more registers than just the default register, this relies on the user sticking to lexicographic order for registers based on their names, with the default qubit and bit registers being q and c respectively. This is done because the tket2 conversion to HUGR relies on this order for qubit / bit ordering. It may be worth treating default registers differently in the future but that depends on user preferences
• I think the original idea was to use arrays when the user specifies them as such in the explicit function signature - I've decided to go with the implicit version as this a) seems to be used more often by users and b) would have been more complicated to implement with more questions with regards to how much flexibility you want to give / would be expected in the signature definition, but this should be a follow up issue (Support arrays in explicit pytket circuit signatures #859)

BREAKING CHANGE: load_pytket takes arrays by default (pass use_arrays=False for qubit arguments)

[ss2165]

is it feasible to just have one guppy.load method? If the behaviour can't be inferred automatically I suggest guppy.load_pytket does arrays by default and you add a kwarg to have individual wires.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 93.22%. Comparing base (75eb441) to head (b1e8ef2).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #858      +/-   ##
==========================================
+ Coverage   93.19%   93.22%   +0.03%     
==========================================
  Files          74       74              
  Lines        8711     8755      +44     
==========================================
+ Hits         8118     8162      +44     
  Misses        593      593              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[tatiana-s]

is it feasible to just have one guppy.load method? If the behaviour can't be inferred automatically I suggest guppy.load_pytket does arrays by default and you add a kwarg to have individual wires.

I don't see how you would infer it at the point of loading (you have no idea whether the user wants to use arrays or values, and if I am not mistaken you can't create a guppy function that takes either arrays or values) but I've added an optional argument to load_pytket - this does make this a breaking change though

[aborgna-q]

Positional bool arguments are confusing. Could we have use_arrays be kwargs-only?
That would also avoid having a breaking change.

[tatiana-s]

I think it's still a breaking change since if the default is true, previous code that uses load_pytket with individual qubit inputs will stop working

[aborgna-q]

Opposite order to what?

The tket2 conversion should respect pytket's input qubit order, and have qubits before bits.

[tatiana-s]

The guppy function has regular bit returns first before inout qubit returns

[NathanCQC]

NIIIIIIICCCCEEEE ONNNNEEE