wapdown/tests/test_validity.py

"""Validate generated decks against the real WML 1.3 DTD.

The DTD is vendored at `tests/dtd/wml13.dtd` (fetched from
http://www.wapforum.org/DTD/wml13.dtd) so the suite never depends on a
network round trip to a twenty-year-old URL. Validation needs lxml, which
is an optional extra: `uv run --extra dtd pytest`.
"""
from __future__ import annotations

from pathlib import Path

import pytest

lxml_etree = pytest.importorskip(
    "lxml.etree", reason="install the 'dtd' extra for DTD validation"
)

DTD_PATH = Path(__file__).resolve().parent / "dtd" / "wml13.dtd"


@pytest.fixture(scope="module")
def dtd():
    with DTD_PATH.open("rb") as handle:
        return lxml_etree.DTD(handle)


def validation_errors(dtd, markup: str) -> list[str]:
    # The document's own DOCTYPE points at a dead URL, so parse with the
    # external DTD left alone and validate against the vendored copy.
    parser = lxml_etree.XMLParser(resolve_entities=False, load_dtd=False)
    tree = lxml_etree.fromstring(markup.encode("utf-8"), parser)
    if dtd.validate(tree):
        return []
    return [str(error) for error in dtd.error_log.filter_from_errors()]


def assert_valid(dtd, markup: str) -> None:
    errors = validation_errors(dtd, markup)
    if errors:
        pytest.fail("\n".join(errors))


SOURCES = {
    "single_card": "# Title\n\nA paragraph with **bold** and a [link](http://x.test/).\n",
    "two_sections": "Intro.\n\n{.card One}\nFirst.\n\n{.card Two}\nSecond.\n",
    "table": "| A | B |\n| --- | --- |\n| 1 | 2 |\n",
    "code_block": "```\n<wml> & $HOME\n```\n",
    "blockquote_and_lists": "> quoted\n\n- a\n- b\n\n1. one\n2. two\n\n---\n",
    "image": "![Map](map.wbmp)\n",
    "entities": 'Fish & chips < > " $5\n',
}


class TestDeckValidity:
    @pytest.mark.parametrize("name", sorted(SOURCES))
    def test_sources_are_valid(self, dtd, render, name):
        assert_valid(dtd, render(SOURCES[name]))

    @pytest.mark.parametrize(
        "options",
        [
            {},
            {"menu": False},
            {"menu_style": "select"},
            {"template_nav": False},
            {"home_label": "Home"},
            {"cache_control": 0},
            {"access_domain": "example.com", "access_path": "/wap"},
            {"max_card_bytes": 80},
            {"title": "Deck"},
        ],
        ids=[
            "defaults",
            "linear",
            "select-menu",
            "no-template",
            "home-softkey",
            "cache-control",
            "access",
            "paginated",
            "titled",
        ],
    )
    def test_option_combinations_are_valid(self, dtd, render, options):
        assert_valid(dtd, render(SOURCES["two_sections"], **options))

    def test_every_deck_file_is_valid(self, dtd, render_output):
        output = render_output(SOURCES["two_sections"], deck_per_card=True, home_label="Home")
        for document in output.documents:
            assert_valid(dtd, document.text)

    def test_example_documents_are_valid(self, dtd, example_decks):
        for name, markup in example_decks.items():
            assert_valid(dtd, markup)


class TestDetectsInvalidMarkup:
    """The validator has to actually reject things, or it proves nothing."""

    def test_table_outside_a_paragraph_is_rejected(self, dtd):
        # The shape the renderer used to emit before the <p> wrapper landed.
        bad = (
            '<wml><card id="c"><table columns="1"><tr><td>x</td></tr></table>'
            "</card></wml>"
        )
        assert validation_errors(dtd, bad)

    def test_emphasis_inside_an_anchor_is_rejected(self, dtd):
        # The shape the renderer used to emit for `[**go**](url)`.
        bad = '<wml><card id="c"><p><a href="#x"><b>no</b></a></p></card></wml>'
        assert validation_errors(dtd, bad)

    def test_a_valid_deck_reports_no_errors(self, dtd):
        good = '<wml><card id="c"><p>ok</p></card></wml>'
        assert validation_errors(dtd, good) == []