chaospy.Distribution.sample : Issue with seed argument in conda install

[mrudd210]

Describe the bug
I get an error when I run the chaospy.Distribution.sample() function and specific a seed, if I install chaospy using the conda install. It does not recognized the "seed" argument. I reinstalled chaopy using the pip install and the error when away.

To Reproduce
t_shell = chaospy.Uniform(1.27, 2.54) #mm
rho= chaospy.Uniform(0, 800) #mm
alpha= chaospy.Uniform(10, 25) #degrees

uniform_joint_distribution = chaospy.J(alpha,rho, t_shell)
count = 20

analyses_var = uniform_joint_distribution.sample(count, rule="latin_hypercube", seed=1234)

Expected behavior

The analyses_var array should be created with 3 rows and 20 columns with values based on the latin hypercube sampling technique and the joint distribution of t_shell, rho, and alpha.

Desktop (please complete the following information):

• OS: Windows 11
• Python version: 3.12
• Chaospy version 4.3.17 for pip install. This version does NOT have an error.

[jonathf]

I can see that it is quite outdated on the Anaconda side.

I am not responsible for the package deployment there. You could propose an update of the package from the feed stock page:
https://github.com/conda-forge/chaospy-feedstock

[davide-f]

Probably linked to #438, if #438 gets solved, maybe this is automatically solved?