It would be good to extend preprocessor with the following:
1. Case insensitive regexp
2. Optional regexp: the regexp that does not make item unsupported if pattern does not match
3. Allow ${x} syntax for capture patterns

This way i.e. conversions from '300M'/'100G'/'...' to certain bps values can be performed like in the following row of transformations:

OPTIONAL CASE-INSENSITIVE ^(\d+)G$ ${1}000000000
OPTIONAL CASE-INSENSITIVE ^(\d+)M$ ${1}000000
OPTIONAL CASE-INSENSITIVE ^(\d+)K$ ${1}000