Hi

I am currently encountering below issues when trying to install any version of Scikit Learn (0.19.0 or 0.19.1 or 0.19.2) on Linux - RHEL / Centos 7 OS. 

Few months back I could successfully install scikit-learn under python 2.7 without issues. When I re-run the installation of scikit-learn package under python 2.7 its failing with the same below error.

I have installed all the required dependencies of OS and Python packages for scikit-learn

Python Version - 3.7

/usr/local/bsb-python37 - Custom Location in which I configured and installed Python 3.7

I have installed all the prerequisite - OS packages - bias-devel , lapack-devel , atlas-devel. 

Earlier when I installed Python 2.7 I didn't run into any issues. Now when I re-try to install the scikit learn under Python 2.7 I am running into the same issue and the earlier successful installed version of scikit learn is corrupted.

Error
# /usr/local/bsb-python37/bin/python3.7 setup.py install --prefix=/usr/local/bsb-python37
Partial import of sklearn during the build process.
blas_opt_info:
blas_mkl_info:
customize UnixCCompiler
  libraries mkl_rt not found in ['/usr/local/bsb-python37/lib', '/usr/local/lib64', '/usr/local/lib', '/usr/lib64', '/usr/lib', '/usr/lib/']
  NOT AVAILABLE

blis_info:
customize UnixCCompiler
  libraries blis not found in ['/usr/local/bsb-python37/lib', '/usr/local/lib64', '/usr/local/lib', '/usr/lib64', '/usr/lib', '/usr/lib/']
  NOT AVAILABLE

  error: Command "g++ -pthread -Wno-unused-result -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -fPIC -I/usr/local/bsb-python37/lib/python3.7/site-packages/numpy-1.14.5-py3.7-linux-x86_64.egg/numpy/core/include -I/usr/local/bsb-python37/lib/python3.7/site-packages/numpy-1.14.5-py3.7-linux-x86_64.egg/numpy/core/include -I/usr/local/bsb-python37/include/python3.7m -c sklearn/cluster/_dbscan_inner.cpp -o build/temp.linux-x86_64-3.7/sklearn/cluster/_dbscan_inner.o -MMD -MF build/temp.linux-x86_64-3.7/sklearn/cluster/_dbscan_inner.o.d" failed with exit status 1

Any help is greatly appreciated. I tried to google around and did all that I could try. No luck.

Thanks,
Abhishek