Added support for indexing sequence IDs as counts

[karasikov]

Usage: same as indexing counts, except one needs to add --enum-headers instead of --count-kmers:

metagraph annotate ... --enum-headers --count-width <floor(log2(num_sequences)) + 1>

The next transform steps (row-diff, brwt) are the same as with counts -- pass --count-kmers.