Updates for fixing the six versus theano issue

[mathemaphysics]

I avoided all issues with theano.compat.six versus six with these changes. I left a comment with all the details on what I did in the issue related to this. I'll attach it here when I find it. Just ran a couple of bash scripts.

[mathemaphysics]

This is in response to issue #1591. I gave details of the automated changes I made to all sources. I just updated importation of six automatically using a bash script included in the link.

I've quoted it below.

I ran two bash lines which took care of it. So far at least...

files="$(egrep -r 'theano\.compat\.six' ./*)"
found="$(for x in ${files}; do echo $(sed -nre 's/([-\_a-zA-Z0-9]+\.py)(.*)/\1/gp' <<<"${x}"); done | uniq)"
for xx in ${found}; do sed -i -re 's/theano\.compat\.six/six/g' ${xx}; done

and then finally

files="$(egrep -r 'from theano\.compat import six' ./*)"
found="$(for x in ${files}; do echo $(sed -nre 's/([-\_a-zA-Z0-9]+\.py)(.*)/\1/gp' <<<"${x}"); done | uniq)"
for xx in ${found}; do sed -i -re 's/from theano\.compat import six/import six/g' ${xx}; done

This appears to deal with all of the issues making trouble for six versus theano.compat.six.