Both the previous and current behavior is now possible

hard_copy to ensure reparamterized nodes don't affect those already selected

used for reparamterization against residuals

With parameters the error tolerance should be a bit lower

Fixes a minor bug

Should remove any overwrite issues

Remove lines used to check if the feature space is correct

Every computational environment will have the same order of features in it

Changes necessary for new parmeterization scheme

use selected.size()

Use "file" when it is considered standard

When passing into it have the right sized vector

Add sanity checker when accessing VALUES_ARR
Only reindex on the fly generated features

Edits to get the tests to pass with the merged code

Remove all possible overlaps

op_slots are now used throughout

Cleaning some errors up that occured after the merge

generate_node functions now use is_const of the new features

Removes potential of conflicts in feat_inds as the feature set is built

Fixes from master adapted here

int overflows are leading to issues

Need #ifdef in more places

Used scores_sel and not sent_scores for combining features in SIS

prevent is_valid from calling nullptr

0.0 is min for classification

Done to remove copy/paste from python binding functions to C++ constructors

Remove the treadprivate vectors that caused the valgrind errors
It may not have been an actual leak but don't risk it

OMP threadprivate leads to definitely loss in valgrind need to fix

Bug in the python interface led to problem

Fixed all bugs that came up becasue of the merger/removal of domain/log_reg_fit_c

Clean up tests and sources

see https://gitlab.com/sissopp_developers/sissopp
commits: up to 763e40e4deaf7a45dba3f0d8052cca053c1fa957

This works now with intel