Add a script to reindex files or change how we id about 5e-database

[ogregoire]

I just hope that doesn't break anything for anyone.

That will definitely break something for someone. But keeping changing the ids because we remove or add stuff will also definitely break something for someone.

The change will break everything for everyone, but after that nothing will break for no one.

[bagelbits]

This is coming off of a short discussion here: #119

[ogregoire]

For id, I'd just use the name and normalize it with regexes, like this:

1. Use the language's "to lowercase" function.
2. Delete unwanted characters such as initial and final whitespaces, as well as apostrophes inside words: (?:(?<=\bo)'|'(?=s\b)|^\W+|\W+$)
3. Replace sequences of non letters with a single dash: \W+ to -.

This has the advantage to be deterministic and usually unique because, in the SRD, names are unique.

I think regexes are a good fit because they can be used in any language so no matter the scripting language, it will just work. Also, those regex can be simplified or aggregated, but I haven't found a single regex to do all that at once.

[bagelbits]

I think that's a reasonable ID pattern.

[bagelbits]

So that would convert Arcanist's Magic Aura into arcanists-magic-aura. Yeah. I definitely like that.

[bagelbits]

I just hope that doesn't break anything for anyone.

[bagelbits]

If we're going to do this, we may as well drop the domain from the URL as well.

[bagelbits]

Actually, that would be a pretty good use for a repository tag. Since essentially that would be a major version change.

[bagelbits]

@ogregoire If you want to pull the trigger on this, feel free to. I think I'm about to have my hands full.

[bagelbits]

I added a PR to address #113 which drops the domain from the URLs.

[bagelbits]

I also started throwing together a script to do the ID conversion.

[bagelbits]

Well. That might do it but I probably messed something up.

[bagelbits]

Okay. I think this PR is in a good state for adding string ids.