Create in-memory option for sqlite based operations #14
Description
On some file systems sqlite performs badly (e.g. smb mounts, lustre). This hurts especially in bigger workflows with lots of input.
Example: large number of peptides to make a decoy of, each tryptic-reverse peptide needs looking up in DB to make sure it is no a target peptide. Or anything with an expensive select (e.g. lots of joins, or lots of individual selects)
- decoy lookup
- large set-merge
- psm table
Solution: Make an in memory option for higher speed. If you're running GBs of peptide input, you are likely already on some kind of HPC with enough memory.