RecursionError: maximum recursion depth exceeded

[pooryousefshahrooz]

Environment

I am running this command:
from qiskit_algorithms.utils import algorithm_globals

and getting error:

RecursionError                            Traceback (most recent call last)
Cell In[8], line 1
----> 1 from qiskit_machine_learning.neural_networks import EstimatorQNN
      3 estimator_qnn = EstimatorQNN(
      4     circuit=qc1, observables=observable1, input_params=[params1[0]], weight_params=[params1[1]]
      5 )
      6 estimator_qnn

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/qiskit_machine_learning/neural_networks/__init__.py:60
      1 # This code is part of a Qiskit project.
      2 #
      3 # (C) Copyright IBM 2019, 2023.
   (...)
     10 # copyright notice, and modified files need to carry a notice indicating
     11 # that they have been altered from the originals.
     13 """
     14 Quantum neural networks (:mod:`qiskit_machine_learning.neural_networks`)
     15 ========================================================================
   (...)
     57    LocalEffectiveDimension
     58 """
---> 60 from .effective_dimension import EffectiveDimension, LocalEffectiveDimension
     61 from .estimator_qnn import EstimatorQNN
     62 from .neural_network import NeuralNetwork

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/qiskit_machine_learning/neural_networks/effective_dimension.py:23
     21 from qiskit_algorithms.utils import algorithm_globals
     22 from qiskit_machine_learning import QiskitMachineLearningError
---> 23 from .estimator_qnn import EstimatorQNN
     24 from .neural_network import NeuralNetwork
     26 logger = logging.getLogger(__name__)

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/qiskit_machine_learning/neural_networks/estimator_qnn.py:26
     24 from qiskit.quantum_info import SparsePauliOp
     25 from qiskit.quantum_info.operators.base_operator import BaseOperator
---> 26 from qiskit_algorithms.gradients import (
     27     BaseEstimatorGradient,
     28     EstimatorGradientResult,
     29     ParamShiftEstimatorGradient,
     30 )
     32 from qiskit_machine_learning.circuit.library import QNNCircuit
     33 from qiskit_machine_learning.exceptions import QiskitMachineLearningError

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/qiskit_algorithms/__init__.py:254
    252 from .algorithm_job import AlgorithmJob
    253 from .algorithm_result import AlgorithmResult
--> 254 from .variational_algorithm import VariationalAlgorithm, VariationalResult
    255 from .amplitude_amplifiers import Grover, GroverResult, AmplificationProblem, AmplitudeAmplifier
    256 from .amplitude_estimators import (
    257     AmplitudeEstimator,
    258     AmplitudeEstimatorResult,
   (...)
    267     EstimationProblem,
    268 )

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/qiskit_algorithms/variational_algorithm.py:36
     33 from qiskit.circuit import QuantumCircuit
     35 from .algorithm_result import AlgorithmResult
---> 36 from .optimizers import OptimizerResult
     39 class VariationalAlgorithm(ABC):
     40     """The Variational Algorithm Base Class."""

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/qiskit_algorithms/optimizers/__init__.py:145
    143 from .spsa import SPSA
    144 from .tnc import TNC
--> 145 from .umda import UMDA
    147 __all__ = [
    148     "Optimizer",
    149     "OptimizerSupportLevel",
   (...)
    181     "UMDA",
    182 ]

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/qiskit_algorithms/optimizers/umda.py:21
     18 from typing import Any
     20 import numpy as np
---> 21 from scipy.stats import norm
     23 from qiskit_algorithms.utils import algorithm_globals
     24 from .optimizer import OptimizerResult, POINT

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/scipy/stats/__init__.py:610
      1 """
      2 .. _statsrefmanual:
      3 
   (...)
    605 
    606 """  # noqa: E501
    608 from ._warnings_errors import (ConstantInputWarning, NearConstantInputWarning,
    609                                DegenerateDataWarning, FitError)
--> 610 from ._stats_py import *
    611 from ._variation import variation
    612 from .distributions import *

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/scipy/stats/_stats_py.py:49
     46 # Import unused here but needs to stay until end of deprecation periode
     47 # See https://github.com/scipy/scipy/issues/15765#issuecomment-1875564522
     48 from scipy import linalg  # noqa: F401
---> 49 from . import distributions
     50 from . import _mstats_basic as mstats_basic
     52 from ._stats_mstats_common import _find_repeats, theilslopes, siegelslopes

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/scipy/stats/distributions.py:10
      1 #
      2 # Author:  Travis Oliphant  2002-2011 with contributions from
      3 #          SciPy Developers 2004-2011
   (...)
      6 #       instead of `git blame -Lxxx,+x`.
      7 #
      8 from ._distn_infrastructure import (rv_discrete, rv_continuous, rv_frozen)  # noqa: F401
---> 10 from . import _continuous_distns
     11 from . import _discrete_distns
     13 from ._continuous_distns import *  # noqa: F403

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/scipy/stats/_continuous_distns.py:12
     10 import numpy as np
     11 from numpy.polynomial import Polynomial
---> 12 from scipy.interpolate import BSpline
     13 from scipy._lib.doccer import (extend_notes_in_docstring,
     14                                replace_notes_in_docstring,
     15                                inherit_docstring_from)
     16 from scipy._lib._ccallback import LowLevelCallable

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/scipy/interpolate/__init__.py:167
      1 """
      2 ========================================
      3 Interpolation (:mod:`scipy.interpolate`)
   (...)
    165 (should not be used in new code).
    166 """
--> 167 from ._interpolate import *
    168 from ._fitpack_py import *
    170 # New interface to fitpack library:

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/scipy/interpolate/_interpolate.py:12
      9 from scipy._lib._util import copy_if_needed
     10 from scipy.special import comb
---> 12 from . import _fitpack_py
     13 from ._polyint import _Interpolator1D
     14 from . import _ppoly

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/scipy/interpolate/_fitpack_py.py:8
      5 import numpy as np
      7 # These are in the API for fitpack even if not used in fitpack.py itself.
----> 8 from ._fitpack_impl import bisplrep, bisplev, dblint  # noqa: F401
      9 from . import _fitpack_impl as _impl
     10 from ._bsplines import BSpline

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/scipy/interpolate/_fitpack_impl.py:103
     52 _iermess = {
     53     0: ["The spline has a residual sum of squares fp such that "
     54         "abs(fp-s)/s<=0.001", None],
   (...)
     68     'unknown': ["An error occurred", TypeError]
     69 }
     71 _iermess2 = {
     72     0: ["The spline has a residual sum of squares fp such that "
     73         "abs(fp-s)/s<=0.001", None],
   (...)
     99     'unknown': ["An error occurred", TypeError]
    100 }
    102 _parcur_cache = {'t': array([], float), 'wrk': array([], float),
--> 103                  'iwrk': array([], dfitpack_int), 'u': array([], float),
    104                  'ub': 0, 'ue': 1}
    107 def splprep(x, w=None, u=None, ub=None, ue=None, k=3, task=0, s=None, t=None,
    108             full_output=0, nest=None, per=0, quiet=1):
    109     # see the docstring of `_fitpack_py/splprep`
    110     if task <= 0:

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/numpy/core/_dtype.py:46, in __repr__(dtype)
     45 def __repr__(dtype):
---> 46     arg_str = _construction_repr(dtype, include_align=False)
     47     if dtype.isalignedstruct:
     48         arg_str = arg_str + ", align=True"

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/numpy/core/_dtype.py:100, in _construction_repr(dtype, include_align, short)
     98     return _subarray_str(dtype)
     99 else:
--> 100     return _scalar_str(dtype, short=short)

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/numpy/core/_dtype.py:143, in _scalar_str(dtype, short)
    140 elif dtype.type == np.timedelta64:
    141     return "'%sm8%s'" % (byteorder, _datetime_metadata_str(dtype))
--> 143 elif np.issubdtype(dtype, np.number):
    144     # Short repr with endianness, like '<f8'
    145     if short or dtype.byteorder not in ('=', '|'):
    146         return "'%s%c%d'" % (byteorder, dtype.kind, dtype.itemsize)

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/numpy/core/numerictypes.py:417, in issubdtype(arg1, arg2)
    359 r"""
    360 Returns True if first argument is a typecode lower/equal in type hierarchy.
    361 
   (...)
    414 
    415 """
    416 if not issubclass_(arg1, generic):
--> 417     arg1 = dtype(arg1).type
    418 if not issubclass_(arg2, generic):
    419     arg2 = dtype(arg2).type

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/numpy/core/_dtype.py:46, in __repr__(dtype)
     45 def __repr__(dtype):
---> 46     arg_str = _construction_repr(dtype, include_align=False)
     47     if dtype.isalignedstruct:
     48         arg_str = arg_str + ", align=True"

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/numpy/core/_dtype.py:100, in _construction_repr(dtype, include_align, short)
     98     return _subarray_str(dtype)
     99 else:
--> 100     return _scalar_str(dtype, short=short)

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/numpy/core/_dtype.py:143, in _scalar_str(dtype, short)
    140 elif dtype.type == np.timedelta64:
    141     return "'%sm8%s'" % (byteorder, _datetime_metadata_str(dtype))
--> 143 elif np.issubdtype(dtype, np.number):
    144     # Short repr with endianness, like '<f8'
    145     if short or dtype.byteorder not in ('=', '|'):
    146         return "'%s%c%d'" % (byteorder, dtype.kind, dtype.itemsize)

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/numpy/core/numerictypes.py:417, in issubdtype(arg1, arg2)
    359 r"""
    360 Returns True if first argument is a typecode lower/equal in type hierarchy.
    361 
   (...)
    414 
    415 """
    416 if not issubclass_(arg1, generic):
--> 417     arg1 = dtype(arg1).type
    418 if not issubclass_(arg2, generic):
    419     arg2 = dtype(arg2).type

    [... skipping similar frames: __repr__ at line 46 (717 times), _construction_repr at line 100 (717 times), _scalar_str at line 143 (716 times), issubdtype at line 417 (716 times)]

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/numpy/core/_dtype.py:143, in _scalar_str(dtype, short)
    140 elif dtype.type == np.timedelta64:
    141     return "'%sm8%s'" % (byteorder, _datetime_metadata_str(dtype))
--> 143 elif np.issubdtype(dtype, np.number):
    144     # Short repr with endianness, like '<f8'
    145     if short or dtype.byteorder not in ('=', '|'):
    146         return "'%s%c%d'" % (byteorder, dtype.kind, dtype.itemsize)

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/numpy/core/numerictypes.py:417, in issubdtype(arg1, arg2)
    359 r"""
    360 Returns True if first argument is a typecode lower/equal in type hierarchy.
    361 
   (...)
    414 
    415 """
    416 if not issubclass_(arg1, generic):
--> 417     arg1 = dtype(arg1).type
    418 if not issubclass_(arg2, generic):
    419     arg2 = dtype(arg2).type

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/numpy/core/_dtype.py:46, in __repr__(dtype)
     45 def __repr__(dtype):
---> 46     arg_str = _construction_repr(dtype, include_align=False)
     47     if dtype.isalignedstruct:
     48         arg_str = arg_str + ", align=True"

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/numpy/core/_dtype.py:100, in _construction_repr(dtype, include_align, short)
     98     return _subarray_str(dtype)
     99 else:
--> 100     return _scalar_str(dtype, short=short)

File ~/miniconda3/envs/conda_env/lib/python3.12/site-packages/numpy/core/_dtype.py:104, in _scalar_str(dtype, short)
    103 def _scalar_str(dtype, short):
--> 104     byteorder = _byte_order_str(dtype)
    106     if dtype.type == np.bool_:
    107         if short:

RecursionError: maximum recursion depth exceeded

Suggested solutions
I know another issue has been opened on this but I could not find any response for that. And that is why I opened another one. thanks

Additional Information

qiskit-ibm-runtime version: qiskit-machine-learning: 0.7.2 and qiskit: 1.1.1
Python version:3.12
Operating system: LINUX

What is happening?

I am getting RecursionError: maximum recursion depth exceeded error when running command from qiskit_algorithms.utils import algorithm_globals.

How can we reproduce the issue?

qiskit-ibm-runtime version: qiskit-machine-learning: 0.7.2 and qiskit: 1.1.1
Python version:3.12
Operating system: LINUX
Numpy version: 2.0.1

from qiskit_algorithms.utils import algorithm_globals

What should happen?

I should be able to import the library.

Any suggestions?

No response

[woodsp-ibm]

I know another issue has been opened on this but I could not find any response for that.

What other issue? Can you link it here.

from qiskit_algorithms.utils import algorithm_globals

The import fine for me locally, entered from a Python prompt

>>> from qiskit_algorithms.utils  import algorithm_globals
>>> 

and many of the unit tests import that in order to seed things so the outcome is reproducible for the test. All unit tests here are passing as expected - these are run nightly.

From your trace it seems to start with an import from machine learning and has nothing to do with algorithm_globals. algorithm_globals has nothing to do with machine learning so that cannot have been the start point.

Since it seems to fail recursion in numpy what version are you running, have you tried updating that? Maybe try a clean environment, if you already haven't, and just install the minimum needed for what you want to run there and see.

[pooryousefshahrooz]

I had opened an issue on qiskit-ibm-runtime but jyu00 suggested I open an issue here.

[woodsp-ibm]

Ok I see that there now, which has been closed.

• RecursionError: maximum recursion depth exceeded Qiskit/qiskit-ibm-runtime#1845

[pooryousefshahrooz]

my numpy library version is 2.0.1.

[woodsp-ibm]

Try installing the newest version of the 1. series of numpy i.e. 1.26.4. Qiskit ML has included a constraint to restrict numpy < 2.0 which I think was mainly due to compatibility with torch qiskit-community/qiskit-machine-learning#815 but I know 2.0 series has caused some problems for some projects so it may be worth a try with the 1.26.4

I will note that the tests run here daily across the code, with various OS and Python versions, do install 2.0.1 numpy and all the tests pass as I noted earlier with algorithms_globals being successfully imported with no such recursion error.

[pooryousefshahrooz]

Update: I just downgraded my numpy to 1.26.4 and it did work!! thanks @woodsp-ibm so very much for your help.

[woodsp-ibm]

Glad to hear that sorted out your issue - or at least avoids whatever the problem is. As I noted algorithms is tested here with 2.0.1 so at that level things work, but of course you are including/using more so hopefully whatever the overall problem you had with 2.0.1 will get sorted out over time - the 2.0 numpy series is pretty new so it may just need a bit for other projects to ensure things work with that. E.g. like Machine Learning which has pinned it to not use the 2.0 series due to some issue(s). Anyway as this seems to have sorted your problem, and with algorithms working afaik based on the tests that are run using it, I am going to close this.