Fixes for BEAST loggers and STEMMA state encodings
This minor release adds some missing attributes to state/ancestral logger elements in BEAST outputs to ensure that root frequencies (corresponding to intrinsic probability judgments) are incorporated into probability calculations for state sampling. It also fixes a previous bug in mapping variant reading indices to state codes in STEMMA outputs, so that reading indices (up to a maximum of 24 per variation unit) are now mapped to single-character state codes.