P4D compatability with NumPy

[UncleWael]

When trying the Webinar II tutorials , PyVizSVG and PyChartHTML until now , this error keeps popping :

" A module that was compiled using NumPy 1.x cannot be run in NumPy 2.2.4 as it may crash. To support both 1.x and 2.x versions of NumPy, modules must be compiled with
NumPy 2.0. Some module may need to rebuild instead e.g. with 'pybind11>=2.12'. "

and this prevents me from using the package since it is not compatable with the new versions of Numpy as the error says. is there a solution from the Delphi side ? because
obviously I cannot downgrade the Numpy.

Context : Delphi 12.1 , Python 3.12.3 , Numpy 2.2.4

[pyscripter]

Have you googled for the error message?

e.g. https://stackoverflow.com/questions/78641150/a-module-that-was-compiled-using-numpy-1-x-cannot-be-run-in-numpy-2-0-0

This is not a P4D issue.

[UncleWael]

yes ofcourse i googled it, and as i mentioned above , downgrading the Numpy is not an option (fix 1 in the link you sent) for obvious reasons, fix 2 is what i'm asking for : "install the latest version of the module which is failing" in our case , it is P4D

[pyscripter]

"install the latest version of the module which is failing" in our case , it is P4D

P4D is not a python module. These demos make use of a number of python modules:

• pandas
• matplotlib
• seaborn
• mpld3
• bokeh
• altair
• scikit-learn

and it appears that some of them are not yet compatible with numpy v2. You can easily find which one by importing them from the python CLI.

running pip install -U package_name

with each of the above will likely fix the error.

[pyscripter]

I have checked the Webinar II demos with python 3.13 and updated versions of the python modules and they work (no import errors).
VizDemo was updated to make it compatible with recent versions of python.
The AnalyticsDemo needs some updating to make it compatible.

It will run if you replace in line 66:

bayesian_grid = {'tol': tol, 'alpha_1': alpha_1, 'alpha_2' : alpha_2, 'lambda_1': lambda_1, 'lambda_2' : lambda_2,
                 'normalize' : normalize}

with
bayesian_grid = {'tol': tol, 'alpha_1': alpha_1, 'alpha_2' : alpha_2, 'lambda_1': lambda_1, 'lambda_2' : lambda_2}

but the results are not correct.

[UncleWael]

Ok , I will do and I will follow up. Thank you for help.