could not convert string to float in prepare_data function in module arcgis.learn

Describe the bug Due to my Slovak locale settings my system is using comma (,) in float number values. But the code to convert a string to float using float() function requires a dot (.) value.

To Reproduce Steps to reproduce the behavior:

https://learn.arcgis.com/en/projects/use-deep-learning-to-assess-palm-tree-health/
https://downloads.esri.com/learnarcgis/use-deep-learning-to-assess-palm-tree-health/deeplearning.zip

error:

could not convert string to float: '395,31'

Screenshots

Expected behavior I believe the user should not handle this type of error simply because the code is not written well. There should be a some kind of pre-check function that determines what locale is being used in the system and based on that a numerical value representation of arcgis API should be adjusted automatically. So that a user should not dwelve for 2 weeks the root of this problem just to find out that his locale has to be changed to some friendly country code as depicted here by a blue color https://upload.wikimedia.org/wikipedia/commons/a/a8/DecimalSeparator.svg

Platform (please complete the following information):

• OS: Windows 10 64bit
• Browser Chrome latest
• Python API Version 1.6.2

Additional context Whole problem is described at the bottom of this issue: https://community.esri.com/message/890543

[priyankatuteja]

You’re Welcome! No worries @CMPeng I will do that as it gets fixed. Will also migrate it to Pro team backlogs.

[priyankatuteja]

@kuttamuwa You could try this workaround https://community.esri.com/t5/imagery-and-remote-sensing/deep-learning-to-assess-palm-tree-prepare-data-problem/m-p/411499/highlight/true#M3770