Setup data-retrieval script and GitHub action

[tcompa]

• This should be based on https://github.com/tcompa/test-table-github-page/blob/main/src/compile_html.py, in terms of how/where to define packages to be included.
• For the initial tests/development, we won't include any package which is not on PyPI
• Output data must be a JSON object with the same structure as the GET /api/v2/task/ response in fractal-server
• All this procedure must be available through a GitHub action, which runs regularly (once a week) or on demand

We can setup a lightweight model here, based on TaskReadV2 (https://github.com/fractal-analytics-platform/fractal-server/blob/main/fractal_server/app/schemas/v2/task.py#L107-L132), e.g. like

class TaskReadV2(BaseModel):
    name: str
    type: Literal["parallel", "non_parallel", "compound"]
    source: Optional[str] = None
    version: Optional[str] = None
    docs_info: Optional[str] = None
    docs_link: Optional[HttpUrl] = None
    input_types: dict[str, bool]
    output_types: dict[str, bool]
    category: Optional[str] = None
    modality: Optional[str] = None
    authors: Optional[str] = None
    tags: list[str]

[tcompa]

Update about supported sources:

Existing:

• A PyPI project (already there)

To add

• A GitHub repository where the latest release has a wheel asset attached (to be added)

To deprecate

• The URL of a specific wheel file
• The URL of a specific zip file