get_structures.py & cross_correlation.py both use parameters that are now deprecated in scikit-image and scikit-learn. This results in failure when running chess extract and chess cross-correlate.
Specifically in get_structures.py lines 130 & 137 use the deprecated parameter multichannel=False. This is now channel_axis=None in later versions of scikit-image. And in cross_correlation.py, line 97 uses precomputes_distances=True with Kmeans from scikit-learn. This has now simply been removed from Kmeans and is no longer used.
If you install chess using pip, then it automatically installs later versions of scikit-image and scikit-learn creating a lot of headaches. Can the chess scripts be updated for later versions or can the pip installation be modified to require the older versions?
get_structures.py & cross_correlation.py both use parameters that are now deprecated in scikit-image and scikit-learn. This results in failure when running chess extract and chess cross-correlate.
Specifically in get_structures.py lines 130 & 137 use the deprecated parameter multichannel=False. This is now channel_axis=None in later versions of scikit-image. And in cross_correlation.py, line 97 uses precomputes_distances=True with Kmeans from scikit-learn. This has now simply been removed from Kmeans and is no longer used.
If you install chess using pip, then it automatically installs later versions of scikit-image and scikit-learn creating a lot of headaches. Can the chess scripts be updated for later versions or can the pip installation be modified to require the older versions?