docs: improve structure and add modules

[gcroci2]

You can look at the new docs rendered here.

I suggest to not look at the commits for this PR, but directly to the files changes. Main changes:

• The modules rendered are now dataset, query, trainer, neuralnets, molstructu, tools (before only dataset was successfully rendered, see here). I removed features since we didn't add any docs there, so it's useless to render those submodules on readthedocs.
• I fixed various cross sections refs and links in the installation page.
• Code of conduct and contributing files are now .md files, a preferred file format (easier to handle).
• Contributing page in the docs is now linked to the one we have in the repo's main root (less duplication).
• Uniform query.py to dataset.py in terms of doc strings (both have a parent class which is used by two children classes; the args are explained in detail only in the children, not in the parents - otherwise it's visually very verbose).
• Add from __future__ import annotations for making possible the rendering of modules that were not correctly rendered before (or not rendered at all). To fix the issue, also TYPE_CHECKING is used. Briefly, there were problems with type hints that made sphinx build fail on some modules (an error in the type hints import caused the modules to not be rendered).
• Now the docs/ folder has a source/ subfolder. This is useful for building the docs locally (where you can have a build/ folder as well, not uploaded to the remote of course).

Note for myself:

• Hide the branch on readthedocs when the PR is merged.

[DaniBodor]

Great updates, the docs are definitely much more complete now.
Were you planning on adding descriptions for the NNs (as per #608) in this PR or in a separate one?

I added a commit that adds empty lines at the end of a couple of files were that wasn't the case

[gcroci2]

Great updates, the docs are definitely much more complete now. Were you planning on adding descriptions for the NNs (as per #608) in this PR or in a separate one?

Separate PR, but I've already added the neural networks submodule so it's ready for issue #608 :)