Add autograph support for for loops

[lillian542]

Context:
We added initial autograph support with custom PennyLane conversion for if/else statements in #6406, and while loops in #6413. This PR adds support for for loops as well.

Description of the Change:
We add a custom PennyLane implementation for for_stmt for autograph, based on the existing Catalyst implementation.

Benefits:

We can have logic like this in a captured function:

@qml.qnode(qml.device("default.qubit", wires=1))
def f(params):
    for x in params:
        qml.RY(x, wires=0)
    return qml.expval(qml.PauliZ(0))

ag_circuit = run_autograph(f)
jaxpr = jax.make_jaxpr(ag_circuit)(jnp.array([1.0, 2.0, 3.0]))

and the resulting JAXPR will contain a for-loop instead of unrolling the entire loop.

Possible Drawbacks:
Not a drawback, per se, but this is the part of the implementation that is most "behind" catalyst in terms of feature parity. Catalyst has an option to fall back on a Python implementation that is based on queueing instead of raising an error in some circumstances. A few test are xfailed with a note that they are implemented in Catalyst.

Relevant Shortcut Stories:
[sc-71821]

[codecov]

Codecov Report

Attention: Patch coverage is 89.89899% with 10 lines in your changes missing coverage. Please review.

Please upload report for BASE (autograph_while_loop@84ca81b). Learn more about missing BASE report.
Report is 4 commits behind head on autograph_while_loop.

Files with missing lines	Patch %	Lines
pennylane/capture/autograph/ag_primitives.py	89.89%	10 Missing ⚠️

Additional details and impacted files

@@                   Coverage Diff                   @@
##             autograph_while_loop    #6426   +/-   ##
=======================================================
  Coverage                        ?   99.43%           
=======================================================
  Files                           ?      458           
  Lines                           ?    42976           
  Branches                        ?        0           
=======================================================
  Hits                            ?    42734           
  Misses                          ?      242           
  Partials                        ?        0           

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

[PietropaoloFrisoni]

Looks pretty good to me!

[astralcai]

LGTM! 🎸