chemenu/tools/chemenu/tests/test_blocks.py

"""Tests for generated regions - the delimiters that replaced heading matching.

The whole point is that a region's *identity* stops depending on its heading
text. Everything here is about the two questions the old approach answered by
guessing: where does the region start, and where does it stop.
"""
from __future__ import annotations

from chemenu import blocks

PROSE = "# Page\n\n## Beschreibung\n\nProse.\n"


def _links(heading="Beziehungen", lines=("- **uses:** [[X]]",)):
    return blocks.render(blocks.LINKS, heading, list(lines))


def test_a_region_round_trips():
    body = blocks.replace(PROSE, blocks.LINKS, _links())
    assert blocks.find(body, blocks.LINKS) == "## Beziehungen\n\n- **uses:** [[X]]"
    assert blocks.strip(body, blocks.LINKS) == PROSE


def test_replacing_does_not_append_a_second_region():
    body = blocks.replace(PROSE, blocks.LINKS, _links())
    again = blocks.replace(body, blocks.LINKS, _links(lines=["- **uses:** [[Y]]"]))
    assert again.count(blocks.open_marker(blocks.LINKS)) == 1
    assert "[[X]]" not in again and "[[Y]]" in again


def test_the_heading_inside_a_region_is_not_how_it_is_found():
    """A page whose region carries a heading the instance never declared - an
    unconverted page, a hand-edit, another language - is still located exactly.
    Under heading matching this was the case that silently created a second
    section."""
    body = blocks.replace(PROSE, blocks.LINKS, _links(heading="Something Else Entirely"))
    assert "[[X]]" in blocks.find(body, blocks.LINKS)
    assert blocks.strip(body, blocks.LINKS) == PROSE


def test_content_after_a_region_survives_a_rewrite():
    """The eight-page bug, as a test. The old block ran to the next heading -
    and before that to the end of the file - so anything sitting after it was
    deleted on the next write."""
    body = blocks.replace(PROSE, blocks.LINKS, _links()) + "\n## Afterwards\n\nKeep me.\n"
    rewritten = blocks.replace(body, blocks.LINKS, _links(lines=["- **uses:** [[Z]]"]))
    assert "Keep me." in rewritten
    assert rewritten.count("## Afterwards") == 1


def test_two_regions_coexist_without_reading_each_other():
    body = blocks.replace(PROSE, blocks.LINKS, _links())
    body = blocks.replace(
        body, blocks.FOOTNOTES,
        blocks.render(blocks.FOOTNOTES, "Fußnoten", ["[^s-x]: [[Source - X]]"]),
    )
    assert "[[X]]" in blocks.find(body, blocks.LINKS)
    assert "[^s-x]" in blocks.find(body, blocks.FOOTNOTES)
    assert blocks.marker_pairs(body) == {"links": 1, "footnotes": 1}


def test_an_empty_region_is_no_region_at_all():
    """A page that cites nothing must not carry an empty Footnotes heading."""
    assert blocks.render(blocks.LINKS, "Beziehungen", []) == ""
    body = blocks.replace(PROSE, blocks.LINKS, _links())
    assert blocks.replace(body, blocks.LINKS, "") == PROSE


def test_an_absent_region_reads_as_none_not_as_empty():
    """None and "" have to stay distinguishable: one means the page has no
    region, the other that it has one holding nothing."""
    assert blocks.find(PROSE, blocks.LINKS) is None


def test_a_dropped_marker_is_detectable():
    """An agent rewriting prose at the boundary can lose one. Silent otherwise:
    the region becomes ordinary prose and the next write appends a second one
    beside it."""
    body = blocks.replace(PROSE, blocks.LINKS, _links())
    assert blocks.unbalanced_markers(body) == []
    assert blocks.unbalanced_markers(body.replace(blocks.close_marker(blocks.LINKS), "")) == ["links"]
    assert blocks.unbalanced_markers(body.replace(blocks.open_marker(blocks.LINKS), "")) == ["links"]


def test_marker_pairs_counts_rather_than_sets():
    """A page that went from one region to two has the same set of names and a
    different count - which is why `migrate verify` compares counts."""
    body = blocks.replace(PROSE, blocks.LINKS, _links())
    doubled = body + "\n" + _links() + "\n"
    assert blocks.marker_pairs(doubled) == {"links": 2}