so that print.css can
start every chapter – including cover, title page and imprint – on a new page.
-For other formats it only removes the placeholders (EPUB, DOCX, ODT: use --toc there).
+For LaTeX (PDF route md -> html -> pdf, filter in BOTH steps, no --toc) the table of contents becomes
+\tableofcontents at the same place, every chapter starts with \clearpage and there is no \maketitle
+before the cover. For other formats it only removes the placeholders (EPUB, DOCX, ODT: use --toc there).
Usage:
pandoc buch.md -s --lua-filter /pfad/zu/epub2md/pandoc/book.lua \
@@ -73,7 +75,123 @@ local function chapter_anchor(block)
return id
end
+-- Characters in running text that even DejaVu lacks: typeset as math symbols ("↩︎": footnote back-links).
+local LATEX_SYMBOLS = { ["↩"] = "$\\hookleftarrow$", ["ℓ"] = "$\\ell$", ["\u{2329}"] = "$\\langle$",
+ ["\u{232A}"] = "$\\rangle$", ["\u{3008}"] = "$\\langle$", ["\u{3009}"] = "$\\rangle$",
+ ["∖"] = "$\\setminus$", ["⋯"] = "$\\cdots$", ["\u{FE0E}"] = "" }
+local LATEX_FONTS = { mainfont = "DejaVu Serif", monofont = "DejaVu Sans Mono", mathfont = "DejaVu Math TeX Gyre" }
+
+local function font_installed(name)
+ local pipe = io.popen("fc-list -q " .. string.format("%q", name) .. " && echo yes")
+ local out = pipe and pipe:read("a") or ""
+ if pipe then
+ pipe:close()
+ end
+ return out:match("yes") ~= nil
+end
+
+local function latex_symbols(str)
+ local found = false
+ for _, c in utf8.codes(str.text) do
+ if LATEX_SYMBOLS[utf8.char(c)] then
+ found = true
+ break
+ end
+ end
+ if not found then
+ return nil
+ end
+ local out, buf = pandoc.Inlines({}), {}
+ for _, c in utf8.codes(str.text) do
+ local ch = utf8.char(c)
+ local tex = LATEX_SYMBOLS[ch]
+ if tex then
+ if #buf > 0 then
+ out:insert(pandoc.Str(table.concat(buf)))
+ buf = {}
+ end
+ if tex ~= "" then
+ out:insert(pandoc.RawInline("latex", tex))
+ end
+ else
+ buf[#buf + 1] = ch
+ end
+ end
+ if #buf > 0 then
+ out:insert(pandoc.Str(table.concat(buf)))
+ end
+ return out
+end
+
+-- LaTeX (2nd step of the PDF route md -> html -> pdf): the HTML from step 1 already holds the table of
+-- contents (div#TOC.toc), the chapter starts (div.chapter-start) and, if moved, Pandoc's title block.
+local function latex_book(doc)
+ local found = false
+ doc.blocks:walk({ Div = function(d)
+ if d.identifier == "TOC" and d.classes:includes("toc") or d.classes:includes("chapter-start") then
+ found = true
+ end
+ end })
+ if not found then
+ return nil
+ end
+ local blocks = pandoc.Blocks({})
+ for _, block in ipairs(doc.blocks) do
+ if block.t == "Div" and block.identifier == "TOC" and block.classes:includes("toc") then
+ blocks:insert(pandoc.RawBlock("latex", "\\clearpage\n\\tableofcontents\n\\clearpage"))
+ elseif block.t == "Div" and block.classes:includes("chapter-start") then
+ blocks:insert(pandoc.RawBlock("latex", "\\clearpage"))
+ blocks:insert(block)
+ elseif block.t == "Div" and block.classes:includes("cover") then
+ blocks:insert(block:walk({ Image = function(img) -- cover as large as the page allows
+ img.attributes.width = nil
+ img.attributes.height = "20cm"
+ return img
+ end }))
+ elseif block.t == "Div" and block.identifier == "title-block-header" then
+ blocks:insert(pandoc.RawBlock("latex", "\\clearpage\n\\begin{center}\n\\vspace*{3cm}"))
+ for _, b in ipairs(block.content) do
+ if b.t == "Header" then
+ local title = pandoc.Inlines({ pandoc.RawInline("latex", "{\\Huge\\bfseries ") })
+ title:extend(b.content)
+ title:insert(pandoc.RawInline("latex", "}"))
+ blocks:insert(pandoc.Para(title))
+ else
+ blocks:insert(b)
+ end
+ end
+ blocks:insert(pandoc.RawBlock("latex", "\\end{center}"))
+ else
+ blocks:insert(block)
+ end
+ end
+ doc.blocks = blocks:walk({ Str = latex_symbols })
+ -- fonts with a wide Unicode coverage (∈ λ ≠ ≥ □, thin spaces …), unless the user chose fonts himself
+ -- A4 with moderate margins, like print.css (Pandoc's LaTeX default: US letter, wide margins)
+ doc.meta.papersize = doc.meta.papersize or "a4"
+ doc.meta.geometry = doc.meta.geometry or "margin=2cm"
+ for var, font in pairs(LATEX_FONTS) do
+ if doc.meta[var] == nil and font_installed(font) then
+ doc.meta[var] = font
+ end
+ end
+ -- no \maketitle at the very top (before the cover); keep the PDF metadata
+ if doc.meta.title then
+ doc.meta["title-meta"] = pandoc.utils.stringify(doc.meta.title)
+ doc.meta.title = nil
+ end
+ if doc.meta.author then
+ doc.meta["author-meta"] = pandoc.utils.stringify(doc.meta.author)
+ doc.meta.author = nil
+ end
+ doc.meta.date = nil
+ return doc
+end
+
function Pandoc(doc)
+ if FORMAT:match("latex") then
+ return latex_book(doc)
+ end
local html = FORMAT:match("html") and not FORMAT:match("epub")
local toc_at, title_at = nil, nil
for i, block in ipairs(doc.blocks) do
@@ -98,6 +216,7 @@ function Pandoc(doc)
end
local blocks = pandoc.Blocks({})
+ local seen_chapter = false
for i, block in ipairs(doc.blocks) do
local anchor = chapter_anchor(block)
if i == toc_at then
@@ -117,15 +236,26 @@ function Pandoc(doc)
end
elseif anchor then
blocks:insert(pandoc.Div({}, pandoc.Attr(anchor, { "chapter-start" })))
- local nxt = doc.blocks[i + 1]
- if anchor:match("^%d%d%d%-cover") and nxt and nxt.t == "Figure" then
- -- cover chapter: the picture without the caption "Cover", centred on its own page
- local imgs = {}
+ -- cover: the first chapter, if it holds nothing but one picture ('Cover', '[Titelinfo]', 'Deckblatt' …);
+ -- shown without the caption "Cover", centred on its own page
+ local nxt, after = doc.blocks[i + 1], doc.blocks[i + 2]
+ local ends = after == nil or chapter_anchor(after) ~= nil or is_placeholder(after, TOC_ID)
+ or is_placeholder(after, TITLE_ID)
+ if not seen_chapter and nxt and (nxt.t == "Figure" or nxt.t == "Para") and ends then
+ local imgs, text = {}, false
nxt:walk({ Image = function(img) imgs[#imgs + 1] = img end })
- if #imgs == 1 then
+ if nxt.t == "Para" then
+ for _, inl in ipairs(nxt.content) do
+ if inl.t ~= "Image" and inl.t ~= "Space" and inl.t ~= "SoftBreak" then
+ text = true
+ end
+ end
+ end
+ if #imgs == 1 and not text then
doc.blocks[i + 1] = pandoc.Div({ pandoc.Plain({ imgs[1] }) }, pandoc.Attr("", { "cover" }))
end
end
+ seen_chapter = true
else
blocks:insert(block)
end
diff --git a/tests/test_chapters.py b/tests/test_chapters.py
index c98dd87..75d71fd 100644
--- a/tests/test_chapters.py
+++ b/tests/test_chapters.py
@@ -185,3 +185,18 @@ def test_special_chapter_is_not_split_further(convert):
"
Register
A
Apfel
Index
x
")
run = convert(Book(docs=[Doc("a.xhtml", body)]))
assert [p.name for p in run.chapters()] == ["001-literatur.md", "002-register.md"]
+
+
+def test_first_image_only_chapter_is_cover_and_titel_is_title_page(convert):
+ from conftest import PNG
+ docs = [Doc("info.xhtml", '
', "[Titelinfo]"),
+ Doc("t.xhtml", "
Erika Musterfrau
Testbuch
", "Titel"),
+ Doc("d.xhtml", '
', "Deckblatt"),
+ Doc("c1.xhtml", "
Eins
Text
")]
+ run = convert(Book(docs=docs, nav=[("Eins", "c1.xhtml")], media={"c.png": PNG}))
+ types = {c["file"]: c["types"] for c in run.report["chapters"]}
+ first, second, third = sorted(types)[:3]
+ assert "cover" in types[first]
+ assert "titlepage" in types[second]
+ assert "cover" in types[third] # 'Deckblatt' by title
+ assert "cover" not in types[sorted(types)[3]]
diff --git a/tests/test_merge.py b/tests/test_merge.py
index 2aa702a..1a2d122 100644
--- a/tests/test_merge.py
+++ b/tests/test_merge.py
@@ -118,3 +118,21 @@ def test_merge_own_toc_replaces_placeholder(convert):
assert merge_main([str(run.out), "--toc"]) == 0
text = (run.out / f"{run.report['input']['name'][:-5]}.md").read_text(encoding="utf-8")
assert "epub2md-toc" not in text and text.index("© 2024") < text.index("## Inhalt") < text.index("# Kapitel A")
+
+
+def test_merge_front_matter_region_with_gaps(convert):
+ """Krooss/Blake-like: short untyped pages between title page and imprint; main matter starts at '1 …'."""
+ docs = [Doc("cover.xhtml", '
'),
+ Doc("titel.xhtml", "
Erika Musterfrau
Vorspann
Roman
"),
+ Doc("auflage.xhtml", "
Vorspann
4. Auflage 2025
"),
+ Doc("impressum.xhtml", "
© 2024 Testverlag
Das Werk ist urheberrechtlich geschützt.
"),
+ Doc("brief.xhtml", "
Liebe Leserin, lieber Leser,
Viel Spaß.
"),
+ Doc("k1.xhtml", "
1 Der Einstieg
a
"), Doc("k2.xhtml", "
Zwischenspiel
b
")]
+ book = Book(title="Vorspann", docs=docs, nav=[("1 Der Einstieg", "k1.xhtml"), ("Zwischenspiel", "k2.xhtml")],
+ media={"img/a.png": PNG})
+ run = convert(book)
+ assert merge_main([str(run.out)]) == 0
+ text = (run.out / f"{run.report['input']['name'][:-5]}.md").read_text(encoding="utf-8")
+ toc = text.index('
')
+ assert text.index("© 2024") < text.index("Viel Spaß.") < toc < text.index("# 1 Der Einstieg")
+ assert text.count("epub2md-toc") == 1 and "epub2md-titlepage" not in text
diff --git a/tests/test_qa2_open.py b/tests/test_qa2_open.py
index 35f3e7b..e9b1e41 100644
--- a/tests/test_qa2_open.py
+++ b/tests/test_qa2_open.py
@@ -32,7 +32,9 @@ def test_qa2_10_book_readme_recommends_pdf_via_html(convert):
"""Pandoc drops raw HTML (tables,
) when going Markdown -> LaTeX; the HTML route keeps both."""
run = convert(Book(docs=[Doc("a.xhtml", "A
x
")]))
readme = (run.out / "README.md").read_text(encoding="utf-8")
- assert re.search(r"pandoc \S+\.md -s -o \S+\.html && pandoc \S+\.html --toc --pdf-engine=xelatex -o \S+\.pdf", readme)
+ book = r"--lua-filter \S+/pandoc/book\.lua" # table of contents after the front matter (both steps)
+ assert re.search(rf"pandoc \S+\.md -s {book} -o \S+\.html && pandoc \S+\.html {book} --pdf-engine=xelatex -o \S+\.pdf",
+ readme)
def test_qa2_11_code_language_hint(convert):
