Account for padding/gap elements in input sequences

[RMichae1]

Currently it is the case that either the problem alphabet is required to have the required 'gap', '""' , etc. for encoding the sequence to an integer array correctly.

However some sequences may contain elements not listed in the alphabet and we require a fallback integer label.

I recommend using alphabet + 1 indices where zero is reserved as fallback.

This way sampling the alphabet will not yield broken sequences, as it is still restricted to the valid alphabet elements, however strings with padding can be encoded to valid integers.