Introduce default values for embed-code attributes

[dmytro-kashcheiev]

In most cases, embed-code is used with a limited set of languages (Kotlin, Groovy, Java, etc.).
It should be possible to specify default values for some attributes.

This is important because specifying the project root as a code source currently results in processing directories such as build, generated, node_modules, etc. Those directories can contain files, links, sockets, and other entries that cannot be copied or opened, resulting in embed-code crashes.

Review the list of settings and define default values where possible. See ripgrep(for example) as a reference for how this could be handled. Different operating systems may require different default settings (TBD).

[alexander-yevsyukov]

The main point here is having sensible defaults for the code-includes attribute so that the user does not have to bother with entering the obvioius.

Here's what fixed my problem with Git socket on the file system:

code-includes:
  - ".gitignore"
  - "**/*.kts"
  - "**/*.md"
  - "**/*.proto"
  - "**/*.java"
  - "**/*.kt"
  - "**/*.yml"

[alexander-yevsyukov]

We can probably take the .gitattributes file from the root of the repository to find out which text files are used in a project.

[alexander-yevsyukov]

The main idea behind this request is to avoid doing stupid things by default. E.g., we don't need to copy binary files of the well-known formats (.class, .exe, image files etc.) even if the user forgot to specify the custom-includes parameter.

We should not mess with the .git directory either. I got the issue with the git repository when I set the project root as one of the source routes for embedding. Here I am explaining the ripgrep theme in the issue details above.

[alexander-yevsyukov]

Ignoring files that are ignored via .gitignore similarly to ripgrep is also a good idea for the sensible defaults behavior.