auto_arima is giving segmentation fault on mac m1

[fkhanbitcap]

Describe the bug

I am using this as an example auto_arima use case but getting segmentation fault error on Mac M1 Max, can somebody guide me what I am doing wrong. The same code is running on Mac M1 Pro.

To Reproduce

I am using latest Mac M1 Max (64 gb) bought a week ago.

Install python3 pip install pmdarima # along with its dependencies

And use with the following code

import numpy as np
import pmdarima as pm
from pmdarima.datasets import load_wineind

wineind = load_wineind().astype(np.float64)

stepwise_fit = pm.auto_arima(wineind, start_p=1, start_q=1,
                             max_p=3, max_q=3, m=12,
                             start_P=0, seasonal=True,
                             d=1, D=1, trace=True,
                             error_action='trace',  # don't want to know if an order does not work
                             suppress_warnings=False,  # don't want convergence warnings
                             stepwise=False)  # set to stepwise

Versions

Python: 3.10.8
macOS: 13.0.1 (22A400)

Expected Behavior

1. segmentation fault # when running directly in terminal
2. finished with exit code 139 (interrupted by signal 11: SIGSEGV) # when running in django using debugger

Actual Behavior

It should return results in stepwise_fit

Additional Context

No response

[fkhanbitcap]

I was having trouble with pyenv based python 3.10 versions. Seems like after installing mambaforge version of python it gets resolved.