Position-based costs?

[sebastiansteffen]

Hi,

I was wondering if your package supported string-position based costs as well? This would be very useful for fuzzy string matching of, say, company names, which usually do not contain 'typos', but frequently differ at the end or front of the string - consider 'Microsoft' against ''Microsoft Corp.' and 'Microsoap' - we want the first to have a lower edit distance than the latter. The idea would be to allow lower costs for deletion and insertions at the end of a string.

[PinguDEV-original]

Upvote from me 👍