Speed up parseMod by pre-filtering patterns

[vaisest]

Description of the problem being solved:

This is an idea PR of a way to speed up parseMod() a lot. This brings the time to generate a modcache containing the item db mods, and all item mods data.itemMods from around 18 sec to around 2 sec for the first working version of this. It's definitely possible to be many times faster.

This is more of a solution looking for a problem and was mostly done out of curiosity

I think this is somewhat similar to e.g. what the Rust regex crate or Intel hyperscan etc do. The idea is to:

1. Pre-calculate the longest literal part of every pattern
2. Use aho-corasick with those literals to find what patterns can match the line
3. Continue with old parsing code, just with fewer candidates

The used aho-corasick implementation was https://gist.github.com/Validark/d493cfd1b3425c2e3073f5ccd08fbeb9, which is the first pure lua implementation I found. I have zero clue if it's slow or not.

Steps taken to verify a working solution:

Link to a build that showcases this PR:

Before screenshot:

After screenshot: