confusing behavoiur when pass non boolean to stats/binary-measures

[behrica]

Passing seq such as
[0 1 0 1 0 0]
[1 1 1 1 1 1]

to stats/binary-measures

is handled as seq being equal, so for example giving accuracy of 1

It took me very long to discover that it requires true/false.

I suggest to fail fast in this case.

[genmeblog]

Maybe it's not clear in the documentation but you can pass a third argument which will be used to decode your encoding.
It can be a list of truths or predicate, so in your case it's a [1] or #{1} or even a map {1 true 0 false}.

I think I can add reasoning about numerical encoding where 1 is always true and other values will be treated as false.

[genmeblog]

Will be fixed in the next alpha.