[GH-142]: Parse attribute annotations for test cases

[MaxDall]

This PR implements a parsing mechanism based on eval() to get the attribute type hints used in the unit tests directly from the attribute_guidelines.md. See #142 for further info.

Closes #142

[dobbersc]

I've condensed my comments into the function below, as well as some more type hints and docstring.

# noinspection PyUnresolvedReferences
@lru_cache(maxsize=1)
def parse_attribute_annotations() -> Dict[str, object]:
    """Returns a dictionary of the parser's attribute type guidelines mapping from the attribute's name to its type."""

    # We import the attribute annotations types locally to make them accessible in the local namespace,
    # such that eval() can evaluate the type annotations.
    # Therefore, these imports are not unused and manageable more easily than defining them in the global namespace.
    from datetime import datetime
    from typing import Optional

    from src.parser.html_parser import ArticleBody

    with Path("doc/attribute_guidelines.md").open("rb") as file:
        attribute_guidelines: bytes = file.read()

    root = lxml.html.fromstring(attribute_guidelines)
    row_nodes: List[lxml.html.HtmlElement] = root.xpath("//table[@class='annotations']/tr[position() > 1]")
    rows: List[Tuple[str, ...]] = [tuple(child.text_content() for child in node.iterchildren()) for node in row_nodes]
    assert rows and len(rows[0]) == 3, (
        "The annotation guideline table is expected to have exactly three columns: " "'Name', 'Description' and 'Type'."
    )

    local_ns: Dict[str, Any] = locals()
    return {name: eval(annotation, globals(), local_ns) for name, _, annotation in rows}