spaczz: Plural is not chosen over similar word

@brunobg thank you for the examples in #47. I will pull that soon and start experimenting with the examples and using them as benchmarks.

I will also try to spend some time on algorithmic complexity to hopefully further improve spaczz’s speed.

I am about to release v0.4.1 which incorporates some of the optimization I discussed above.

1. I eliminated the unnecessary extra comparison.
2. If flex > 0 and the new thresh value is met (default 100) optimization will not be attempted.
3. Optimization when flex > 0 could be further improved but for now it will terminate if a round of optimization does not produce a better ratio. I still want to have it only pursue the best single flexing direction after the initial round but that will require more code changes.

See the performance section in the readme for parameter changes that will further increase speed.

Short answer for now as I am trying to wrap up v0.4.0 and the potential changes you’re asking for could be patch version changes after that. A combination of ideas you’ve presented and settings tweaking could definitely optimize things considerably. I’ll give a more detailed update later today or tomorrow. Attempting some of these optimizations will be a priority for me after the next release.