🐛 Installing via pip in clean environment gives numpy error on np.float?

[MaximilianJanetschek]

Environment information

• mqt bench 1.1.3
• python 3.11

Description

When installing and trying to use mqt.bench, the import functions throw an attribute error:
AttributeError: np.float_ was removed in the NumPy 2.0 release. Use np.float64 instead.

Referenced code:
from mqt.bench import get_benchmark

"get a benchmark circuit on algorithmic level representing the GHZ state with 5 qubits
qc_algorithmic_level = get_benchmark(benchmark_name="dj", level="alg", circuit_size=5)

"draw the circuit
print(qc_algorithmic_level.draw())

Expected behavior

Successful import

How to Reproduce

1. Clean Environment
2. pip install mqt.bench
3. execute reference code

[burgholzer]

Hey 👋🏻
Thanks for raising this issue. This is just another case of incompatibility with numpy 2.0.
You can work around this for now by installing any numpy version < 2.

pip install --force-reinstall "numpy<2"

I'll keep this issue open until we put a temporary upper cap on numpy in MQT Bench itself.