A .ptt file is an NCBI Protein Table file, which is a tab delimited file containing a list of all the proteins for their genomes (ftp://ftp.ncbi.nih.gov/genomes/). It corresponds with the CDS annotations from the GenBank file and can be created by parsing the GenBank files and writing the appropriate output.
The columns are:
Location Strand Length PID Gene Synonym Code COG Product
For example ftp://ftp.ncbi.nih.gov/genomes/Fungi/Aspergillus_niger_CBS_513_88_uid19263/NC_007445.ptt. These correspond to the feature start and end coordinates, the strand (+/-), the length of the amino acid (minus stop codon), and items analogous to various other GenBank format fields.
You can use BioPerl, as aaron suggests, but it appears as though the BioPython functionality has been dropped? It shouldn't be hard to write your own parser to be honest, although it doesn't store any sequence information!
Just note that you should run the script as "perl gb2ptt.pl --infile filename.gbk" instead of what they said.