Add semantic block parser and PDF export

scripts/export-pdf.mjs

@@ -0,0 +1,64 @@
+import { existsSync, mkdirSync } from "node:fs";
+import path from "node:path";
+import { spawnSync } from "node:child_process";
+import { fileURLToPath } from "node:url";
+
+const repoRoot = path.resolve(fileURLToPath(new URL("..", import.meta.url)));
+const filterPath = path.join(repoRoot, "scripts", "pandoc-filters", "semantic-blocks.lua");
+const preamblePath = path.join(repoRoot, "scripts", "pandoc-templates", "semantic-preamble.tex");
+
+function usage() {
+  console.error("Usage: node scripts/export-pdf.mjs <post.md> [--out output.pdf]");
+}
+
+function parseArgs(argv) {
+  const args = [...argv];
+  const input = args.find((arg) => !arg.startsWith("--"));
+  const outIndex = args.indexOf("--out");
+  const output = outIndex >= 0 ? args[outIndex + 1] : "";
+  return { input, output };
+}
+
+const { input, output } = parseArgs(process.argv.slice(2));
+
+if (!input) {
+  usage();
+  process.exit(2);
+}
+
+const inputPath = path.resolve(input);
+if (!existsSync(inputPath)) {
+  console.error(`Input file not found: ${inputPath}`);
+  process.exit(2);
+}
+
+const outputPath = path.resolve(
+  output || path.join(path.dirname(inputPath), `${path.basename(inputPath, path.extname(inputPath))}.pdf`),
+);
+mkdirSync(path.dirname(outputPath), { recursive: true });
+
+const pandocArgs = [
+  inputPath,
+  "--from",
+  "markdown+fenced_code_attributes+tex_math_dollars+tex_math_single_backslash+footnotes+raw_html",
+  "--pdf-engine=lualatex",
+  "--lua-filter",
+  filterPath,
+  "--include-in-header",
+  preamblePath,
+  "--output",
+  outputPath,
+];
+
+const result = spawnSync("pandoc", pandocArgs, { stdio: "inherit" });
+
+if (result.error?.code === "ENOENT") {
+  console.error("pandoc is not installed or is not on PATH. Install pandoc and a LuaLaTeX distribution first.");
+  process.exit(127);
+}
+
+if (result.status !== 0) {
+  process.exit(result.status ?? 1);
+}
+
+console.log(`Wrote ${outputPath}`);

scripts/pandoc-filters/semantic-blocks.lua

@@ -0,0 +1,131 @@
+local env_types = {
+  theorem = true,
+  lemma = true,
+  proposition = true,
+  corollary = true,
+  definition = true,
+  example = true,
+  remark = true,
+  proof = true,
+}
+
+local function has_class(el, class_name)
+  for _, value in ipairs(el.classes) do
+    if value == class_name then
+      return true
+    end
+  end
+  return false
+end
+
+local function first_env_class(el)
+  for _, value in ipairs(el.classes) do
+    if env_types[value] then
+      return value
+    end
+  end
+  return nil
+end
+
+local function latex_escape(value)
+  value = tostring(value or "")
+  value = value:gsub("\\", "\\textbackslash{}")
+  value = value:gsub("([{}%%$&#_])", "\\%1")
+  value = value:gsub("%^", "\\textasciicircum{}")
+  value = value:gsub("~", "\\textasciitilde{}")
+  return value
+end
+
+local function internal_link_to_cref(el)
+  if FORMAT:match("latex") and el.target:sub(1, 1) == "#" then
+    local label = el.target:sub(2)
+    if label ~= "" then
+      return pandoc.RawInline("latex", "\\cref{" .. label .. "}")
+    end
+  end
+  return nil
+end
+
+local function markdown_to_latex(markdown)
+  local doc = pandoc.read(
+    markdown or "",
+    "markdown+fenced_code_attributes+tex_math_dollars+tex_math_single_backslash+footnotes+raw_html"
+  )
+  doc = doc:walk({ Link = internal_link_to_cref })
+  return pandoc.write(doc, "latex")
+end
+
+local function label_line(label)
+  if label and label ~= "" then
+    return "\\label{" .. label .. "}\n"
+  end
+  return ""
+end
+
+local function interactive_to_latex(el)
+  local label = el.identifier
+  local caption = el.attributes.caption or el.attributes.title or ""
+  local fallback = el.attributes.pdf or el.attributes.fallback or el.text
+  local fallback_latex = markdown_to_latex(fallback)
+  local caption_line = ""
+
+  if caption ~= "" then
+    caption_line = "\\caption{" .. latex_escape(caption) .. "}\n"
+  end
+
+  return pandoc.RawBlock(
+    "latex",
+    table.concat({
+      "\\begin{figure}[htbp]\n",
+      "\\centering\n",
+      "\\fbox{\\begin{minipage}{0.86\\linewidth}\n",
+      fallback_latex,
+      "\n\\end{minipage}}\n",
+      caption_line,
+      label_line(label),
+      "\\end{figure}",
+    })
+  )
+end
+
+local function env_to_latex(el)
+  local env_type = el.attributes.type or first_env_class(el)
+  if not env_type or not env_types[env_type] then
+    return nil
+  end
+
+  local title = el.attributes.title or el.attributes.name or ""
+  local begin_env = "\\begin{" .. env_type .. "}"
+  if title ~= "" then
+    begin_env = begin_env .. "[" .. latex_escape(title) .. "]"
+  end
+
+  return pandoc.RawBlock(
+    "latex",
+    table.concat({
+      begin_env,
+      "\n",
+      label_line(el.identifier),
+      markdown_to_latex(el.text),
+      "\n\\end{",
+      env_type,
+      "}",
+    })
+  )
+end
+
+function CodeBlock(el)
+  if FORMAT:match("latex") and has_class(el, "interactive") then
+    return interactive_to_latex(el)
+  end
+
+  if FORMAT:match("latex") and (has_class(el, "env") or has_class(el, "latex-env") or first_env_class(el)) then
+    return env_to_latex(el)
+  end
+
+  return nil
+end
+
+function Link(el)
+  return internal_link_to_cref(el)
+end

scripts/pandoc-templates/semantic-preamble.tex

@@ -0,0 +1,44 @@
+\usepackage{amsmath}
+\usepackage{amssymb}
+\usepackage{amsthm}
+
+\theoremstyle{plain}
+\newtheorem{theorem}{Theorem}
+\newtheorem{lemma}[theorem]{Lemma}
+\newtheorem{proposition}[theorem]{Proposition}
+\newtheorem{corollary}[theorem]{Corollary}
+
+\theoremstyle{definition}
+\newtheorem{definition}[theorem]{Definition}
+\newtheorem{example}[theorem]{Example}
+
+\theoremstyle{remark}
+\newtheorem{remark}[theorem]{Remark}
+
+\newcommand{\SemanticLoadCleveref}{%
+  \usepackage{cleveref}%
+  \crefname{theorem}{theorem}{theorems}%
+  \Crefname{theorem}{Theorem}{Theorems}%
+  \crefname{lemma}{lemma}{lemmas}%
+  \Crefname{lemma}{Lemma}{Lemmas}%
+  \crefname{proposition}{proposition}{propositions}%
+  \Crefname{proposition}{Proposition}{Propositions}%
+  \crefname{corollary}{corollary}{corollaries}%
+  \Crefname{corollary}{Corollary}{Corollaries}%
+  \crefname{definition}{definition}{definitions}%
+  \Crefname{definition}{Definition}{Definitions}%
+  \crefname{example}{example}{examples}%
+  \Crefname{example}{Example}{Examples}%
+  \crefname{remark}{remark}{remarks}%
+  \Crefname{remark}{Remark}{Remarks}%
+  \crefname{figure}{figure}{figures}%
+  \Crefname{figure}{Figure}{Figures}%
+}
+
+\makeatletter
+\@ifpackageloaded{hyperref}{%
+  \SemanticLoadCleveref
+}{%
+  \AddToHook{package/hyperref/after}{\SemanticLoadCleveref}%
+}
+\makeatother