Fix rules that combine sample outputs failing when dataset has a large number of samples

[zjnolen]

Fixes #72. I think most of these errors are from cat receiving more arguments than it can handle. Trying to fix it by using echo and passing the output to xargs, which seems to work in the current shell container.

[zjnolen]

Updated the following rules to use python scripts instead of shell commands:

angsd_makeBamlist now uses make_bamlist.py

compile_endo_cont, merge_ind_depth, merge_ibs_ref_bias, compile_kinship_stats_sfs, aggregate_fst_global, and aggregate_fst_window now all use concat_files.py

In all cases, outputs should be identical to those of previous pipeline versions, so reruns can be safely ignored if previous versions ran successfully.