make retrieval of PDB/other sequence records faster

Retrieving accession records for large numbers of IDs can be slow, particularly with PDB files which seem to take several seconds per accession to retrieve and parse. Ways should be explored to retrieve records in parallel, and/or bulk where possible.

for PDB records, JAL-801 could be implemented for those who have PDB files as local mirrors, since it is trivial to search for a file in a directory that has a particular name.