feat: add parsers support

[cmdoret]

This PR adds a new "Parser" interface that can parse file contents into triples. It also migrates the existing license detection code into a parser.

This required semantic changes throughout the code, to replace "source" by "git provider" to avoid confusion.

The PR does the following:

• Move gimie.source -> gimie.extractors
• Define Parser interface
• Add LicenseParser
• Simplify logic in Project and make it compatible with Parsers
• Update all tests and docstrings
• Update CLI to:
  • Allow the user to include / exclude specific parsers in gimie data
  • Add a command to list available parsers.

The important changes are in gimie/project.py, gimie/parsers and gimie/cli.py.

Implementing a new Parser is as simple as defining a subclass in gimie.parsers (and adding it to PARSERS):

from gimie.graph.namespaces import SDO
from gimie.parsers import Parser
from rdflib import Graph, Literal, URIRef

class ExampleParser(Parser):
    """A dummy parser for demonstration.
    It will parse files named 'example' and generate a
    graph with a single triple: 
    <URI> <schema:description <file contents>
    """

    def __init__(self, uri: URIRef):
        super().__init__(uri)

    def can_parse(resource: Resource) -> bool:
        # Match based on filename (or content)
        return Resource.name == "example"

    def _parse(resource: Resource) -> Graph:
        # Parsing logic
        contents = resource.open().read()
        return Graph().add(
            (self.uri, SDO.description, Literal(contents))
        )

Example usage of parsers on the command line :

# Use all parsers (default)
➜ gimie data <url>
# Only use license parser
➜ gimie data --include-parser license <url>
# Use all parsers except license
➜ gimie data --exclude-parser license <url>
# The include/exclude flags have short forms and can be specified multiple times
# Enable only cff and pyproject parsers (NOTE: these don't exist yet)
➜ gimie data -I cff -I pyproject <url> 

# list available parsers and their functionalities
➜ gimie parsers --verbose
license: Parse LICENSE file(s) into schema:license.

[vancauwe]

A very nice change of architecture towards having extractors and parsers. Clear naming and flow.

My suggestions would be:

• to decouple the parser from a file finder
• decouple the graph making from the parsing (as we decoupled the extraction from the graph making)

[cmdoret]

Here is a summary of new changes:

• CLI: print parser name in bold green 🍀
• More flexible PARSERS using NamedTuple
• Helper functions to list and use parsers (thus decoupling the file matching from parsers)
• Parsers Output Property instead of Graph, helper function in gimie.graph to make the conversion
• Moved code related to parsers and extractors out of project.py into their respective modules
• added tests for gimie.parsers

[vancauwe]

Nothing more to say: I really like how you chose to implement my suggestions and I find it makes the whole much more modular.
Really nice! ⭐