#15 Simplify interpretation of error rate parameter

[Jason Rafe Miller]

The unittiger's -e option lets users specify the expected error rate of base
calls. The unitigger interprets the given value in one of two ways,
depending on whether the value contains a decimal. Simplify the
interface, for instance by declaring that the value shall be expressed as a
percentage so 15 means 15% and 1.5 means 1.5%.

[Michael Schatz]

Logged In: YES
user_id=1122424

It is true that floating point and integer values are
handled a little differently, but I think 15 is interpretted
as 1.5% error, not 15% error. Integer values are treated as
parts per thousand (1/1000 = .1%). The rational for this is
the error rate for an overlap is stored in .1% increments in
the ovlStore.

-Mike

[Brian Walenz]

Unitigger & BOG now use the same "fraction error" as the reset of the assembler. The default value supplied to the -e options is now 0.015 == 1.5% error. Both programs will fail to run if the value supplied to -e is above 0.25. runCA catches this error before the assembly starts.