pyo3_runtime.PanicException: no entry found for key in /project/rustworkx-core/src/token_swapper.rs:141:41

[lmaltsis]

Environment

• Qiskit version: 8acbb04
• Python version: Python 3.12.3
• Operating system: Ubuntu 24.04.4 LTS

What is happening?

$ RUST_BACKTRACE=full python /tmp/bug.py

thread '<unnamed>' panicked at /project/rustworkx-core/src/token_swapper.rs:141:41:
no entry found for key
stack backtrace:
   0:     0x747fa43c4af6 - <std::sys::backtrace::BacktraceLock::print::DisplayBacktrace as core::fmt::Display>::fmt::hf435e8e9347709a8
   1:     0x747fa4378f13 - core::fmt::write::h0a51fad3804c5e7c
   2:     0x747fa43c456f - std::io::Write::write_fmt::h9759e4151bf4a45e
   3:     0x747fa43c4843 - std::sys::backtrace::BacktraceLock::print::h1ec5ce5bb8ee285e
   4:     0x747fa43c4243 - std::panicking::rust_panic_with_hook::h8b29cbe181d50030
   5:     0x747fa43ef328 - std::panicking::begin_panic_handler::{{closure}}::h9f5b6f6dc6fde83e
   6:     0x747fa43ef289 - std::sys::backtrace::__rust_end_short_backtrace::hd7b0c344383b0b61
   7:     0x747fa43efa8c - __rustc[5224e6b81cd82a8f]::rust_begin_unwind
   8:     0x747fa405adbf - core::panicking::panic_fmt::hc49fc28484033487
   9:     0x747fa405b2ea - core::option::expect_failed::hfe7afbd436ce9c45
  10:     0x747fa435e47d - rustworkx::token_swapper::__pyfunction_graph_token_swapper::h09ea1fd0969a13bd
  11:     0x747fa40f431c - pyo3::impl_::trampoline::trampoline::h013bbb6699fae59b
  12:     0x747fa435b4d1 - rustworkx::token_swapper::<impl rustworkx::token_swapper::graph_token_swapper::MakeDef>::_PYO3_DEF::trampoline::h48cf98b4d8f569ea
  13:           0x5821ef - <unknown>
  14:           0x5492b5 - _PyObject_MakeTpCall
  15:           0x5d7389 - _PyEval_EvalFrameDefault
  16:           0x5d58bb - PyEval_EvalCode
  17:           0x608b52 - <unknown>
  18:           0x6b5263 - <unknown>
  19:           0x6b4fca - _PyRun_SimpleFileObject
  20:           0x6b4dff - _PyRun_AnyFileObject
  21:           0x6bce65 - Py_RunMain
  22:           0x6bc94d - Py_BytesMain
  23:     0x748024e2a1ca - __libc_start_call_main
                               at ./csu/../sysdeps/nptl/libc_start_call_main.h:58:16
  24:     0x748024e2a28b - __libc_start_main_impl
                               at ./csu/../csu/libc-start.c:360:3
  25:           0x657b05 - _start
  26:                0x0 - <unknown>
Traceback (most recent call last):
  File "/tmp/bug.py", line 15, in <module>
    ltpass.run(dag)
  File "/home/user/software/qiskit-venv/slot-a/src/qiskit/transpiler/passes/routing/layout_transformation.py", line 115, in run
    perm_circ = self.token_swapper.permutation_circuit(permutation, self.trials)
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/user/software/qiskit-venv/slot-a/src/qiskit/transpiler/passes/routing/algorithms/token_swapper.py", line 77, in permutation_circuit
    sequential_swaps = self.map(permutation, trials=trials)
                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/user/software/qiskit-venv/slot-a/src/qiskit/transpiler/passes/routing/algorithms/token_swapper.py", line 105, in map
    return rx.graph_token_swapper(self.graph, mapping, trials, seed, parallel_threshold)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
pyo3_runtime.PanicException: no entry found for key

How can we reproduce the issue?

from qiskit import QuantumRegister, QuantumCircuit
from qiskit.converters import circuit_to_dag
from qiskit.transpiler import CouplingMap, Layout, Target
from qiskit.transpiler.passes import LayoutTransformation

v = QuantumRegister(4, "v")
coupling = CouplingMap([[0, 1], [1, 2], [2, 3]])
from_layout = Layout({v[0]: 0, v[1]: 4, v[2]: 2, v[3]: 3})
to_layout = Layout({v[0]: 3, v[1]: 0, v[2]: 1, v[3]: 2})
ltpass = LayoutTransformation(
    coupling_map=coupling, from_layout=from_layout, to_layout=to_layout, seed=42
)
qc = QuantumCircuit(4)
dag = circuit_to_dag(qc)
ltpass.run(dag)

What should happen?

input validation of layout

Any suggestions?

No response

[jakelishman]

These are true bugs, but in general this feels like you are just spamming us with the results of fuzzing things with random clearly broken inputs.

LayoutTransformation should not run without error here, the layout clearly is entirely incompatible with the coupling map.

These will all be very low priority fixes - they're obviously broken input, and there is a client-side error appearing.

[lmaltsis]

Edited + Apologies for the burst of several bug reports, this was a backlog and potential future reports will be made gradually

[jakelishman]

No I'm sorry, I shouldn't have replied last night while I wasn't at work and was all grumpy from the heatwave (I'm one city over from you, by the looks of our profiles).

I do appreciate the bug reports coming in, and the reproducers have been done very well (especially so, assuming they're coming from a fuzzing tool - usually those are a mess of unrelated stuff on top). I stand by that they'll be lower priority to fix, and there will be places where we potentially choose not to add additional input validation in cases where the cost/benefit trade-off would be too high (the CommutationChecker one is on my mind here: we generally have to allocate to check duplication, and the vast majority of calls to the commutation checker are derived from circuits that can't have duplicate qargs), but I do appreciate having them.

[lmaltsis]

I appreciate the feedback. I take great care reducing reproducers (although sometimes lack the domain-specific knowledge to get fully semantic reductions), but I think in the case of these last 3 reports the other parts of the report could've taken more care (it was late and hot for me as well).

What some maintainers have asked before is for several reports of a similar nature (e.g. in your case runtime panics) to be consolidated into a single issue where we can keep adding comments for new crashes, this helps to keep down the noise in the issue tracker.