Introduce configuration file support

[vinistock]

While Rubydex is a framework that can be used to build all sorts of tools, in most cases we want consistency in the base analysis.

For example, imagine you have an LSP and a linter based on Rubydex, but because they each configure the framework differently, you get different results in the editor vs in the CLI. That would be a pretty poor experience.

I believe we can provide an easy way to ensure consistency while still maintaining flexibility so that the odd case can load custom configuration files for a random tool.

Implementation

I recommend reviewing per commit:

1. Allow loading configuration from a file. This adds the infrastructure to read, parse and apply a configuration file, hooking it up to the graph. See the example configuration file below for the format
2. Exposes load_config in the Ruby API
3. Adds .rubydex default configuration for this repo itself

# .rubydex
# Comments
single_value = true

multi_value_one_line = one, two
multi_value_multi_line = one, two,
  three,
  four

trailing_comma = hello,

I went with a simple text/ini format to avoid extra dependencies. It's super simple to parse and entirely based on new lines and commas.

[st0012]

The implementation looks good but I'm not sure about the format we use for .rubydex.
I personally have NEVER used the INI format in any tools before, nor have seen it being used in the Ruby ecosystem. So I think using it would add additional friction to properly configure Rubydex for users.
I think yml is still the best choice for user adoption. toml will be fine for me too.