torben
f63a72cb24
Files changed: - CHANGES.md - VERSION - tools/chemenu/tests/conftest.py - tools/chemenu/tests/test_corpus_diff.py - tools/chemenu/tests/test_index_build.py - tools/chemenu/tests/test_kb_collections.py - tools/chemenu/tests/test_lint.py - tools/chemenu/tests/test_migrate_cmd.py - tools/chemenu/tests/test_new_page.py - tools/chemenu/tests/test_page_ops.py - tools/chemenu/tests/test_provenance.py - tools/chemenu/tests/test_search.py - tools/chemenu/tests/test_touch.py
550 lines
20 KiB
Python
550 lines
20 KiB
Python
from pathlib import Path
|
|
|
|
from typer.testing import CliRunner
|
|
|
|
from chemenu.frontmatter_io import read_page, write_page
|
|
from chemenu.provenance import (
|
|
broken_raw_refs,
|
|
cite_id,
|
|
citing_pages,
|
|
duplicate_raw_file_owners,
|
|
extract_inline_cites,
|
|
legacy_citation_markers,
|
|
legacy_source_pages,
|
|
page_raw_files,
|
|
render_cite_block,
|
|
render_page_body,
|
|
source_pages_by_raw_file,
|
|
source_raw_files,
|
|
split_cite_block,
|
|
unique_cite_id,
|
|
uncovered_raw_files,
|
|
)
|
|
from chemenu.kb_scan import load_kb_pages
|
|
|
|
runner = CliRunner()
|
|
|
|
|
|
def _footnote_block(*cites: tuple[str, str | None]) -> str:
|
|
"""Build a `[^id]` reference for each (title, qualifier) plus the
|
|
trailing Footnotes block defining it, in one string - the fixture form
|
|
tests use in place of the old `^[[Title]]` marker."""
|
|
ids = [cite_id(title, qualifier) for title, qualifier in cites]
|
|
refs = "".join(f"[^{cid}]" for cid in ids)
|
|
block = render_cite_block({cid: cite for cid, cite in zip(ids, cites)})
|
|
return refs, block
|
|
|
|
|
|
def _source_page(title: str, raw_files: list[str]):
|
|
from chemenu.page import Page
|
|
|
|
return Page(
|
|
path=Path(f"/tmp/{title}.md"),
|
|
frontmatter={"type": "types/source.md", "raw_files": raw_files},
|
|
body="",
|
|
)
|
|
|
|
|
|
def test_duplicate_raw_file_owners_reports_a_file_claimed_twice():
|
|
"""The Almanac 10-bootstrap-manual case: an umbrella page and a per-step page
|
|
both claiming the same raw file, which uncovered_raw_files() cannot see."""
|
|
pages = {
|
|
"Source - Umbrella": _source_page("Source - Umbrella", ["raw/a.md", "raw/b.md"]),
|
|
"Source - Step": _source_page("Source - Step", ["raw/b.md"]),
|
|
}
|
|
assert duplicate_raw_file_owners(pages) == [
|
|
{"raw_file": "raw/b.md", "owners": ["Source - Step", "Source - Umbrella"]}
|
|
]
|
|
|
|
|
|
def test_duplicate_raw_file_owners_is_empty_when_ownership_is_unique():
|
|
pages = {
|
|
"Source - A": _source_page("Source - A", ["raw/a.md"]),
|
|
"Source - B": _source_page("Source - B", ["raw/b.md"]),
|
|
}
|
|
assert duplicate_raw_file_owners(pages) == []
|
|
|
|
|
|
def test_duplicate_raw_file_owners_ignores_repeats_within_one_page():
|
|
"""A file listed twice in the same page's raw_files: is untidy, not a
|
|
contested claim - there is still exactly one owner."""
|
|
pages = {"Source - A": _source_page("Source - A", ["raw/a.md", "raw/a.md"])}
|
|
assert duplicate_raw_file_owners(pages) == []
|
|
|
|
|
|
def test_extract_inline_cites_plain_and_qualified():
|
|
refs, block = _footnote_block(
|
|
("Source - Aurora", None), ("Source - Almanac Architecture", "storage-model.md")
|
|
)
|
|
body = f"Fact one {refs}. Fact two {refs}.\n\n{block}"
|
|
cites = extract_inline_cites(body)
|
|
assert ("Source - Aurora", None) in cites
|
|
assert ("Source - Almanac Architecture", "storage-model.md") in cites
|
|
|
|
|
|
def test_extract_inline_cites_ignores_plain_wikilinks():
|
|
body = "See [[Source - Aurora]] for background, but this line has no hard fact."
|
|
assert extract_inline_cites(body) == set()
|
|
|
|
|
|
def test_extract_inline_cites_ignores_undefined_ref():
|
|
"""A `[^id]` with no matching definition resolves to nothing here - lint's
|
|
undefined_footnote_refs is what flags that, not extract_inline_cites."""
|
|
assert extract_inline_cites("Fact one [^s-ghost].") == set()
|
|
|
|
|
|
def test_cite_id_strips_prefix_and_slugifies():
|
|
assert cite_id("Source - Aurora") == "s-aurora"
|
|
assert cite_id("Source - Almanac Architecture", "storage-model.md") == "s-almanac-architecture--storage-model-md"
|
|
|
|
|
|
def test_cite_id_is_deterministic_and_ascii():
|
|
assert cite_id("Source - Würfelspiel für Fortgeschrittene") == cite_id(
|
|
"Source - Würfelspiel für Fortgeschrittene"
|
|
)
|
|
assert cite_id("Source - Würfelspiel für Fortgeschrittene").isascii()
|
|
|
|
|
|
def test_unique_cite_id_suffixes_on_collision():
|
|
base = cite_id("Source - Aurora")
|
|
assert unique_cite_id(set(), "Source - Aurora") == base
|
|
assert unique_cite_id({base}, "Source - Aurora") == f"{base}-2"
|
|
assert unique_cite_id({base, f"{base}-2"}, "Source - Aurora") == f"{base}-3"
|
|
|
|
|
|
def test_split_and_render_cite_block_round_trip():
|
|
definitions = {"s-aurora": ("Source - Aurora", None), "s-almanac--x-md": ("Source - Almanac", "x.md")}
|
|
body = "# Page\n\nSome prose [^s-aurora].\n\n" + render_cite_block(definitions)
|
|
head, parsed = split_cite_block(body)
|
|
assert parsed == definitions
|
|
assert head == "# Page\n\nSome prose [^s-aurora]."
|
|
|
|
|
|
def test_split_cite_block_empty_when_no_footnotes_heading():
|
|
head, definitions = split_cite_block("# Page\n\nNo citations here.\n")
|
|
assert definitions == {}
|
|
assert head == "# Page\n\nNo citations here."
|
|
|
|
|
|
# --- content after the Footnotes block ---------------------------------------
|
|
#
|
|
# The block used to run to the end of the file, so a section sitting after it
|
|
# was deleted on the next cite add / cite sync / rename. `xref add` appends its
|
|
# sections at the end of the file, so which command ran last decided whether a
|
|
# page kept its cross-references.
|
|
|
|
|
|
def test_a_section_after_the_block_survives_the_round_trip():
|
|
definitions = {"s-aurora": ("Source - Aurora", None)}
|
|
relationships = "## Beziehungen\n\n- **umgesetzt von:** [[wikitool]]\n"
|
|
body = (
|
|
"# Page\n\nSome prose [^s-aurora].\n\n"
|
|
+ render_cite_block(definitions)
|
|
+ "\n"
|
|
+ relationships
|
|
)
|
|
head, parsed = split_cite_block(body)
|
|
assert parsed == definitions
|
|
assert "## Beziehungen" in head
|
|
assert "[[wikitool]]" in head
|
|
|
|
rebuilt = render_page_body(head, parsed)
|
|
assert "- **umgesetzt von:** [[wikitool]]" in rebuilt
|
|
assert "[^s-aurora]: [[Source - Aurora]]" in rebuilt
|
|
|
|
|
|
def test_the_block_is_re_emitted_last_so_the_layout_self_heals():
|
|
"""`xref add` appends at the end of the file. Folding the rescued tail into
|
|
the head means the next cite operation puts the block back at the bottom
|
|
instead of preserving the broken order forever."""
|
|
definitions = {"s-aurora": ("Source - Aurora", None)}
|
|
body = (
|
|
"# Page\n\nProse [^s-aurora].\n\n"
|
|
+ render_cite_block(definitions)
|
|
+ "\n## Siehe auch\n\n- [[Borealis]]\n"
|
|
)
|
|
rebuilt = render_page_body(*split_cite_block(body))
|
|
assert rebuilt.index("## Siehe auch") < rebuilt.index("[^s-aurora]:")
|
|
|
|
|
|
def test_repeated_round_trips_are_stable():
|
|
"""Rescuing content must not move it a little further on every run."""
|
|
definitions = {"s-aurora": ("Source - Aurora", None)}
|
|
body = (
|
|
"# Page\n\nProse [^s-aurora].\n\n"
|
|
+ render_cite_block(definitions)
|
|
+ "\n## Siehe auch\n\n- [[Borealis]]\n"
|
|
)
|
|
once = render_page_body(*split_cite_block(body))
|
|
twice = render_page_body(*split_cite_block(once))
|
|
assert once == twice
|
|
|
|
|
|
def test_stray_prose_inside_the_block_is_kept_not_dropped():
|
|
"""Not a definition and not a section - rescued rather than rejected,
|
|
because this runs under lint too, where raising would refuse to read a
|
|
page instead of reporting it."""
|
|
body = (
|
|
"# Page\n\nProse [^s-aurora].\n\n"
|
|
"## Fußnoten\n\n"
|
|
"[^s-aurora]: [[Source - Aurora]]\n"
|
|
"TODO: check this one\n"
|
|
)
|
|
head, definitions = split_cite_block(body)
|
|
assert definitions == {"s-aurora": ("Source - Aurora", None)}
|
|
assert "TODO: check this one" in head
|
|
|
|
|
|
def test_a_citation_used_in_a_rescued_section_still_resolves():
|
|
body = (
|
|
"# Page\n\nProse.\n\n"
|
|
"## Fußnoten\n\n"
|
|
"[^s-aurora]: [[Source - Aurora]]\n\n"
|
|
"## Beziehungen\n\n- **belegt durch:** [[Borealis]] [^s-aurora]\n"
|
|
)
|
|
assert extract_inline_cites(body) == {("Source - Aurora", None)}
|
|
|
|
|
|
def test_source_raw_files_prefers_raw_files_over_legacy_source():
|
|
from chemenu.page import Page
|
|
|
|
page = Page(
|
|
path=Path("/tmp/Source - X.md"),
|
|
frontmatter={"type": "source", "raw_files": ["raw/notes/A.md"], "source": "https://example.com"},
|
|
body="",
|
|
)
|
|
assert source_raw_files(page) == ["raw/notes/A.md"]
|
|
|
|
|
|
def test_source_raw_files_falls_back_to_legacy_repo_path():
|
|
from chemenu.page import Page
|
|
|
|
page = Page(
|
|
path=Path("/tmp/Source - Y.md"),
|
|
frontmatter={"type": "source", "source": "raw/notes/B.md"},
|
|
body="",
|
|
)
|
|
assert source_raw_files(page) == ["raw/notes/B.md"]
|
|
|
|
|
|
def test_source_raw_files_ignores_url_only_legacy_source():
|
|
from chemenu.page import Page
|
|
|
|
page = Page(
|
|
path=Path("/tmp/Source - Z.md"),
|
|
frontmatter={"type": "source", "source": "https://example.com/article"},
|
|
body="",
|
|
)
|
|
assert source_raw_files(page) == []
|
|
|
|
|
|
def test_source_pages_by_raw_file_uses_legacy_source_fallback(kb_dir, raw_dir):
|
|
pages = load_kb_pages(kb_dir)
|
|
by_raw = source_pages_by_raw_file(pages)
|
|
assert by_raw.get("raw/notes/Aurora.md") == ["Source - Aurora"]
|
|
|
|
|
|
def test_uncovered_raw_files_detects_ingested_gap(kb_dir, raw_dir, monkeypatch):
|
|
import chemenu.config as config
|
|
|
|
monkeypatch.setattr(config, "ROOT", raw_dir.parent)
|
|
pages = load_kb_pages(kb_dir)
|
|
uncovered = uncovered_raw_files(raw_dir, pages)
|
|
assert "raw/notes/Uningested.md" in uncovered
|
|
assert "raw/notes/Aurora.md" not in uncovered
|
|
|
|
|
|
def test_broken_raw_refs_detects_dangling_path(kb_dir, raw_dir, monkeypatch):
|
|
import chemenu.config as config
|
|
|
|
monkeypatch.setattr(config, "ROOT", raw_dir.parent)
|
|
write_page(
|
|
kb_dir / "sources/Source - Ghost.md",
|
|
{
|
|
"type": "source", "source_type": "notes", "author": "Torben",
|
|
"raw_files": ["raw/notes/Does Not Exist.md"], "date": "2026-08-02",
|
|
"tags": [], "entities": [], "concepts": [],
|
|
},
|
|
"\n# Source: Ghost\n\n## Summary\n\nGhost.\n",
|
|
)
|
|
pages = load_kb_pages(kb_dir)
|
|
issues = broken_raw_refs(pages)
|
|
assert {"page": "Source - Ghost", "raw_path": "raw/notes/Does Not Exist.md"} in issues
|
|
|
|
|
|
def test_legacy_source_pages_flags_url_and_directory(kb_dir, raw_dir):
|
|
write_page(
|
|
kb_dir / "sources/Source - External.md",
|
|
{
|
|
"type": "source", "source_type": "article", "author": "someone",
|
|
"source": "https://example.com/article", "date": "2026-08-02",
|
|
"tags": [], "entities": [], "concepts": [],
|
|
},
|
|
"\n# Source: External\n\n## Summary\n\nExternal.\n",
|
|
)
|
|
(raw_dir / "documents").mkdir()
|
|
write_page(
|
|
kb_dir / "sources/Source - DirBacked.md",
|
|
{
|
|
"type": "source", "source_type": "document", "author": "Torben",
|
|
"source": "raw/documents", "date": "2026-08-02",
|
|
"tags": [], "entities": [], "concepts": [],
|
|
},
|
|
"\n# Source: DirBacked\n\n## Summary\n\nDir.\n",
|
|
)
|
|
pages = load_kb_pages(kb_dir)
|
|
issues = legacy_source_pages(pages)
|
|
reasons = {i["page"]: i["reason"] for i in issues}
|
|
assert reasons["Source - External"] == "url-only, no raw_files"
|
|
assert reasons["Source - DirBacked"] == "directory, not a file"
|
|
|
|
|
|
def test_citing_pages_via_frontmatter_and_inline(kb_dir, raw_dir):
|
|
write_page(
|
|
kb_dir / "entities/tools/gdeploy.md",
|
|
{
|
|
"type": "entity", "entity_type": "tool", "tags": [], "created": "2026-07-25",
|
|
"modified": "2026-07-25", "related": [], "sources": ["Source - Aurora"], "confidence": 0.8,
|
|
},
|
|
"\n# gdeploy\n\n## Description\n\nDeploy tool.\n",
|
|
)
|
|
refs, block = _footnote_block(("Source - Aurora", None))
|
|
write_page(
|
|
kb_dir / "concepts/Modbus.md",
|
|
{
|
|
"type": "concept", "concept_type": "protocol", "tags": [], "created": "2026-07-25",
|
|
"modified": "2026-07-25", "related": [], "sources": [], "confidence": 0.7,
|
|
},
|
|
f"\n# Modbus\n\n## Definition\n\nUses port 502 {refs}.\n\n{block}",
|
|
)
|
|
pages = load_kb_pages(kb_dir)
|
|
citers = citing_pages(pages, "Source - Aurora")
|
|
assert "gdeploy" in citers
|
|
assert "Modbus" in citers
|
|
|
|
|
|
def test_page_raw_files_resolves_through_sources_and_inline(kb_dir, raw_dir):
|
|
write_page(
|
|
kb_dir / "concepts/Modbus.md",
|
|
{
|
|
"type": "concept", "concept_type": "protocol", "tags": [], "created": "2026-07-25",
|
|
"modified": "2026-07-25", "related": [], "sources": ["Source - Aurora"], "confidence": 0.7,
|
|
},
|
|
"\n# Modbus\n\n## Definition\n\nIndustrial protocol.\n",
|
|
)
|
|
pages = load_kb_pages(kb_dir)
|
|
raw_files = page_raw_files(pages, pages["Modbus"])
|
|
assert raw_files == ["raw/notes/Aurora.md"]
|
|
|
|
|
|
def test_provenance_md_is_not_a_wiki_page(kb_dir):
|
|
(kb_dir / "provenance.md").write_text("# Provenance Index\n", encoding="utf-8")
|
|
pages = load_kb_pages(kb_dir)
|
|
assert "provenance" not in pages
|
|
|
|
|
|
def test_sources_coverage_command_reports_uncovered_file(kb_dir, raw_dir, monkeypatch):
|
|
import chemenu.config as config
|
|
from chemenu.cli import app
|
|
|
|
monkeypatch.setattr(config, "ROOT", raw_dir.parent)
|
|
monkeypatch.setattr(config, "KB_DIR", kb_dir)
|
|
monkeypatch.setattr(config, "RAW_DIR", raw_dir)
|
|
|
|
result = runner.invoke(app, ["sources", "coverage", "--json"])
|
|
assert result.exit_code == 0, result.output
|
|
assert "raw/notes/Uningested.md" in result.output
|
|
|
|
|
|
def test_sources_trace_by_raw_and_by_page(kb_dir, raw_dir, monkeypatch):
|
|
import chemenu.config as config
|
|
from chemenu.cli import app
|
|
|
|
monkeypatch.setattr(config, "ROOT", raw_dir.parent)
|
|
monkeypatch.setattr(config, "KB_DIR", kb_dir)
|
|
monkeypatch.setattr(config, "RAW_DIR", raw_dir)
|
|
|
|
write_page(
|
|
kb_dir / "entities/systems/aurora.md",
|
|
{
|
|
"type": "entity", "entity_type": "system", "tags": ["server"],
|
|
"created": "2026-07-31", "modified": "2026-07-31", "related": ["Borealis"],
|
|
"sources": ["Source - Aurora"], "confidence": 0.9,
|
|
},
|
|
"\n# aurora\n\n## Description\n\nHosts things.\n",
|
|
)
|
|
|
|
result = runner.invoke(app, ["sources", "trace", "--raw", "raw/notes/Aurora.md"])
|
|
assert result.exit_code == 0, result.output
|
|
assert "Source - Aurora" in result.output
|
|
assert "aurora" in result.output
|
|
|
|
result2 = runner.invoke(app, ["sources", "trace", "--page", "aurora"])
|
|
assert result2.exit_code == 0, result2.output
|
|
assert "Source - Aurora" in result2.output
|
|
assert "raw/notes/Aurora.md" in result2.output
|
|
|
|
|
|
def test_new_source_with_multiple_raw_files(kb_dir, raw_dir, monkeypatch):
|
|
import chemenu.config as config
|
|
from chemenu.cli import app
|
|
|
|
monkeypatch.setattr(config, "ROOT", raw_dir.parent)
|
|
monkeypatch.setattr(config, "KB_DIR", kb_dir)
|
|
# This test is about raw_files:, not about authorship - but `new source`
|
|
# refuses to stamp a placeholder author, and the fixture root has no git
|
|
# identity. Naming one keeps the test off the runner's global git config.
|
|
monkeypatch.setenv("WIKI_AUTHOR", "Fixture Author")
|
|
|
|
(raw_dir / "notes" / "Second.md").write_text("# Second\n", encoding="utf-8")
|
|
|
|
result = runner.invoke(app, [
|
|
"new", "source", "--name", "Multi",
|
|
"--set", "raw_files=raw/notes/Aurora.md,raw/notes/Second.md",
|
|
])
|
|
assert result.exit_code == 0, result.output
|
|
fm, _body = read_page(kb_dir / "sources/Source - Multi.md")
|
|
assert fm["raw_files"] == ["raw/notes/Aurora.md", "raw/notes/Second.md"]
|
|
|
|
|
|
def test_new_source_rejects_nonexistent_raw_path(kb_dir, raw_dir, monkeypatch):
|
|
import chemenu.config as config
|
|
from chemenu.cli import app
|
|
|
|
monkeypatch.setattr(config, "ROOT", raw_dir.parent)
|
|
monkeypatch.setattr(config, "KB_DIR", kb_dir)
|
|
|
|
result = runner.invoke(app, [
|
|
"new", "source", "--name", "Bad", "--set", "raw_files=raw/notes/Nope.md",
|
|
])
|
|
assert result.exit_code != 0
|
|
|
|
|
|
def test_lint_flags_citation_not_in_frontmatter_sources(kb_dir, raw_dir, monkeypatch):
|
|
import chemenu.config as config
|
|
from chemenu.commands.lint import run_lint
|
|
|
|
monkeypatch.setattr(config, "ROOT", raw_dir.parent)
|
|
monkeypatch.setattr(config, "RAW_DIR", raw_dir)
|
|
|
|
refs, block = _footnote_block(("Source - Aurora", None))
|
|
write_page(
|
|
kb_dir / "concepts/Modbus.md",
|
|
{
|
|
"type": "concept", "concept_type": "protocol", "tags": [], "created": "2026-07-25",
|
|
"modified": "2026-07-25", "related": [], "sources": [], "confidence": 0.7,
|
|
"provenance": "sourced",
|
|
},
|
|
f"\n# Modbus\n\n## Definition\n\nUses port 502 {refs}.\n\n{block}",
|
|
)
|
|
report = run_lint(kb_dir)
|
|
drift = report["citation_frontmatter_drift"]
|
|
assert {"page": "Modbus", "cited_but_not_in_sources": "Source - Aurora"} in drift
|
|
|
|
|
|
def test_lint_no_drift_when_source_declared(kb_dir, raw_dir, monkeypatch):
|
|
import chemenu.config as config
|
|
from chemenu.commands.lint import run_lint
|
|
|
|
monkeypatch.setattr(config, "ROOT", raw_dir.parent)
|
|
monkeypatch.setattr(config, "RAW_DIR", raw_dir)
|
|
|
|
refs, block = _footnote_block(("Source - Aurora", None))
|
|
write_page(
|
|
kb_dir / "concepts/Modbus.md",
|
|
{
|
|
"type": "concept", "concept_type": "protocol", "tags": [], "created": "2026-07-25",
|
|
"modified": "2026-07-25", "related": [], "sources": ["Source - Aurora"], "confidence": 0.7,
|
|
"provenance": "sourced",
|
|
},
|
|
f"\n# Modbus\n\n## Definition\n\nUses port 502 {refs}.\n\n{block}",
|
|
)
|
|
report = run_lint(kb_dir)
|
|
assert report["citation_frontmatter_drift"] == []
|
|
|
|
|
|
def test_lint_unmarked_provenance_flags_empty_sources_without_general(kb_dir, raw_dir, monkeypatch):
|
|
import chemenu.config as config
|
|
from chemenu.commands.lint import run_lint
|
|
|
|
monkeypatch.setattr(config, "ROOT", raw_dir.parent)
|
|
monkeypatch.setattr(config, "RAW_DIR", raw_dir)
|
|
report = run_lint(kb_dir)
|
|
# gdeploy in the base fixture has sources: [] and no provenance field at all.
|
|
assert "gdeploy" in report["unmarked_provenance"]
|
|
|
|
|
|
def test_lint_does_not_flag_source_page_self_citation(kb_dir, raw_dir, monkeypatch):
|
|
import chemenu.config as config
|
|
from chemenu.commands.lint import run_lint
|
|
|
|
monkeypatch.setattr(config, "ROOT", raw_dir.parent)
|
|
monkeypatch.setattr(config, "RAW_DIR", raw_dir)
|
|
|
|
refs, block = _footnote_block(("Source - Aurora", "Aurora.md"))
|
|
write_page(
|
|
kb_dir / "sources/Source - Aurora.md",
|
|
{
|
|
"type": "source", "source_type": "notes", "author": "Torben",
|
|
"raw_files": ["raw/notes/Aurora.md"], "date": "2026-08-02",
|
|
"tags": [], "entities": ["aurora"], "concepts": [],
|
|
},
|
|
f"\n# Source: Aurora\n\n## Summary\n\nCovers ZFS setup {refs}.\n\n{block}",
|
|
)
|
|
report = run_lint(kb_dir)
|
|
assert not any(i["page"] == "Source - Aurora" for i in report["citation_frontmatter_drift"])
|
|
|
|
|
|
def _page(body: str):
|
|
from chemenu.page import Page
|
|
|
|
return Page(path=Path("/tmp/Notation.md"), frontmatter={"type": "types/concept.md"}, body=body)
|
|
|
|
|
|
def test_extract_inline_cites_ignores_notation_shown_as_code():
|
|
"""A page describing the citation mechanism resolves nothing: the markers
|
|
it shows are examples. This also decides what counts as sourced, so a
|
|
mention used to be able to raise a page's apparent provenance."""
|
|
block = render_cite_block({
|
|
"s-real": ("Source - Real", None),
|
|
"s-shown": ("Source - Shown", None),
|
|
})
|
|
body = (
|
|
"The marker `[^s-shown]` is pasted at the fact [^s-real].\n\n"
|
|
"```markdown\n[^s-shown]: [[Source - Shown]]\n```\n\n" + block
|
|
)
|
|
# `s-shown` is defined and appears twice - in backticks and in a fence -
|
|
# and is cited by neither. Only the reference in prose resolves.
|
|
assert extract_inline_cites(body) == {("Source - Real", None)}
|
|
|
|
|
|
def test_split_cite_block_ignores_a_fenced_definition_inside_the_block():
|
|
"""The counter-direction: a definition line shown as an example does not
|
|
become a definition, which would then read as an orphan."""
|
|
body = (
|
|
"# Notation\n\nProse [^s-real].\n\n"
|
|
"## Fußnoten\n\n"
|
|
"[^s-real]: [[Source - Real]]\n\n"
|
|
"```markdown\n[^s-example]: [[Source - Example]]\n```\n"
|
|
)
|
|
head, definitions = split_cite_block(body)
|
|
assert definitions == {"s-real": ("Source - Real", None)}
|
|
assert "[^s-example]" in head # rescued, not discarded
|
|
|
|
|
|
def test_legacy_citation_markers_ignore_notation_shown_as_code():
|
|
"""The pre-migration marker is exactly the thing a page about the
|
|
migration has to be able to quote."""
|
|
pages = {
|
|
"Notation": _page(
|
|
"The old form was `^[[Source - X]]`, replaced by a footnote:\n\n"
|
|
"```markdown\n^[[Source - Y]]\n```\n"
|
|
)
|
|
}
|
|
assert legacy_citation_markers(pages) == []
|
|
|
|
|
|
def test_legacy_citation_markers_still_flag_a_real_one():
|
|
pages = {"Stale": _page("A fact ^[[Source - X]] that was never migrated.\n")}
|
|
assert legacy_citation_markers(pages) == [{"page": "Stale", "marker": "^[[Source - X]]"}]
|