FR: mo-doc improvements

[crusso]

It looks like mo-doc describes let-bound function values as just values, not functions, which maybe should be changed. It also seems to partially document private functions in objects (maybe slightly broken support for nested modules causing this?

While going through the Motoko base library docs: https://internetcomputer.org/docs/current/motoko/main/base/ I've noticed a few things that seem to all be caused by the auto-generation of those pages:
1 - Lots of the pages include references to 'values' that hold functions. While they are defined as values in the source code, and are thus technically correct, you might want to consider if it makes more sense to developers to just use the 'value' label for those values that hold static primitives, while calling those values that hold functions as functions. e.g. https://internetcomputer.org/docs/current/motoko/main/base/Int8#value-toint
2 - The 'Edit this page' links at the bottom point to missing pages on Github e.g. The link at the bottom of: https://internetcomputer.org/docs/current/motoko/main/base/Int8 points to: https://github.com/dfinity/portal/edit/master/docs/motoko/main/base/Int8.md I'm not sure if there are other 'Edit this page' links that are inaccurate outside the base library docs, but you might want to double-check them where the docs have been moved on Github.
3 - Where public functions are defined within another function (e.g. as an object method), the internal function is also listed in the links on the top-right. e.g. On this page: https://internetcomputer.org/docs/current/motoko/main/base/Iter there's a next() function defined twice for both class range and revRange, when really they probably don't want to be there at all, as well as the auto-generated docs.

See https://forum.dfinity.org/t/edits-in-motoko-base-library-docs/23433