Project containing scripts to build and access a local database containing UniprotKB protein info via a Python interface using an LMDB or MongoDB backend.
This is used for scripts that need to perform many random queries to fetch protein sequences without hitting a web API constantly (slow, probably annoys the server). Essentially this is a fancy cache that lets you quickly fetch by multiple types of accessions/IDs.
The LMDB backend is now the default since it is fully embedded and does not require a server.