Many applications require mapping strings to integers. In Python, this usually looks like:

d = {"apple": 100, "banana": 200, "cherry": 300}

If there are 1 million keys, this can consume a lot of memory — more than 100 bytes per key.
Our elephant has published a new library that uses about 9 bytes per key. Yes, only 9 bytes. Usage looks like this:

from fastconstmap import ConstMap

d = {"apple": 100, "banana": 200, "cherry": 300}
m = ConstMap(d)

m["apple"]                  # -> 100
m.get_many(["banana", "cherry"])  # -> [200, 300]

It can be significantly faster (for example, up to 2 times in some cases) than the standard dictionary. It can also be serialized and deserialized to disk or network for convenient reuse.

https://pypi.org/project/fastconstmap/

github: https://github.com/lemire/fastconstmap

👉 @PythonRe