BUGFIX
in recent python versions, 1E9 (which is used as the number of rows in the sparse matrix converted to biom) is treated as float rather than int, causing the allocation to fail
Also, unit testing expects in the output fasta to have "size=XXX;" whereas deblur now gives "size=XXX" (without the trailing ";"). Fixed the unit test.
Coverage remained the same at 88.672% when pulling 628a3de45b1f00df9eb65b00fba3e847a8b3201d on fixfloat into 2f66636dd6c928c0dc192f1c390b0246851775dd on master.
BUGFIX in recent python versions, 1E9 (which is used as the number of rows in the sparse matrix converted to biom) is treated as float rather than int, causing the allocation to fail Also, unit testing expects in the output fasta to have "size=XXX;" whereas deblur now gives "size=XXX" (without the trailing ";"). Fixed the unit test.