Make TS types consistent with the documentation terminology

[mathfox]

This PR comes as a solution to the problem that I faced when I was not able to migrate my project from the 0.3.2 version to the 0.4.1-rc.0 one.
The reason was that the latest type definitions break the whole project with an unclear errors.

I believe that the changes are based off the terminology used in the jecs/flecs documentation.

I breakdown the entity type into 3 variations:

• Id (the base type);
• Tag (extends Id, allows the entity to be used as a tag);
• Entity (extends Tag, allows the entity to be used as a component);

This PR also adds a missing 'world' type parameter in the type definition for 'pair_first' and 'pair_second' functions.

Fixes #161

[Ukendio]

I don't think this PR is necessary, have you tried installing the package via npm i git+https://github.com/ukendio/jecs.git

[mathfox]

Well, it was necessary for me to migrate from the previous version with a very few errors showing up. I don't think that the installation part is relevant in my case (either version breaks the whole project with a bunch of unclear errors). I believe I got the terminology right in this PR, correct me if I'm wrong. I think there should not exist a Pair type when at it's core it's just an entity (or id) that could be used as a component/tag only (although I did not restrict the usage of 'add'/'set' methods on this kind of entities in this PR).

[Ukendio]

I am mostly referring to the types which appear to me to be ill formed. @EncodedVenom can review it as well

[EncodedVenom]

Upon first glance I don't like some of the renaming done, but I'll come back tomorrow with a more detailed analysis on this to see what needs some work.

[EncodedVenom]

Upon further review, it seems the whole reason this PR was made was to fix the issue listed in #161 which stems from an ill-defined entity type fixed in this commit as found via several users in the jecs thread on the ROSS server. I think this is beyond the scope of the original issue and should be split off into a new one. I will not make that call and let @Ukendio handle that one.

Since this is not a huge problem right now and mostly stems from minor type issues, I think we can afford to sit on this for a while.