Broken install dependencies (numpy) on PyPy #17214
Description
Describe the bug
scikit-learn installation on pypy fails because of conflicting numpy dependencies
project.toml:
"numpy==1.13.3; python_version=='3.6' and platform_system!='AIX'",
vs.
setup.py:
if platform.python_implementation() == 'PyPy':
SCIPY_MIN_VERSION = '1.1.0'
NUMPY_MIN_VERSION = '1.14.0'
else:
SCIPY_MIN_VERSION = '0.19.1'
NUMPY_MIN_VERSION = '1.13.3'
this leads to inevitable:
ImportError: Your installation of numpy 1.13.3 is out-of-date.
scikit-learn requires numpy >= 1.14.0.
Installation instructions are available on the scikit-learn website: http://scikit-learn.org/stable/install.html
----------------------------------------
ERROR: Failed building wheel for scikit-learn
Failed to build scikit-learn
Steps/Code to Reproduce
pypy3 -m pip install scikit-learn
Expected Results
package installed successfully
Actual Results
ImportError: Your installation of numpy 1.13.3 is out-of-date.
scikit-learn requires numpy >= 1.14.0.
Installation instructions are available on the scikit-learn website: http://scikit-learn.org/stable/install.html
----------------------------------------
ERROR: Failed building wheel for scikit-learn
Failed to build scikit-learn
Versions
scikit-learn-0.23.0