snowkit 0.0.0-stage → 0.1.0

package/src/engine/docpdf-render.js

@@ -0,0 +1,106 @@
+import { layoutDocument } from './docpdf-layout.js';
+
+const FACES = ['regular', 'bold', 'italic', 'bolditalic'];
+
+function link(L, doc, page, o) {
+  const ctx = doc.context;
+  let uri = o.url;
+  try {
+    uri = encodeURI(decodeURI(o.url));
+  } catch (e) {}
+  const ref = ctx.register(ctx.obj({ Type: 'Annot', Subtype: 'Link', Rect: [o.x, o.y, o.x + o.w, o.y + o.h], Border: [0, 0, 0], F: 4, A: { Type: 'Action', S: 'URI', URI: L.PDFString.of(uri) } }));
+  const annots = page.node.lookup(L.PDFName.of('Annots'));
+  if (annots instanceof L.PDFArray) annots.push(ref);
+  else page.node.set(L.PDFName.of('Annots'), ctx.obj([ref]));
+}
+
+export async function buildDocPdf(L, blocks, opts = {}) {
+  const doc = await L.PDFDocument.create();
+  const fonts = {};
+  if (opts.fontkit && opts.fontBytes && opts.fontBytes.regular) {
+    doc.registerFontkit(opts.fontkit);
+    for (const f of FACES) fonts[f] = await doc.embedFont(opts.fontBytes[f] || opts.fontBytes.regular, { subset: true });
+  } else {
+    const std = { regular: L.StandardFonts.Helvetica, bold: L.StandardFonts.HelveticaBold, italic: L.StandardFonts.HelveticaOblique, bolditalic: L.StandardFonts.HelveticaBoldOblique };
+    for (const f of FACES) fonts[f] = await doc.embedFont(std[f]);
+  }
+  fonts.mono = await doc.embedFont(L.StandardFonts.Courier);
+  fonts.monobold = await doc.embedFont(L.StandardFonts.CourierBold);
+  const custom = !!(opts.fontkit && opts.fontBytes && opts.fontBytes.regular);
+  const encodable = new Map();
+  const canEncode = (f, text) => {
+    if (custom && !f.startsWith('mono')) return true;
+    const key = f + '\u0000' + text;
+    if (encodable.has(key)) return encodable.get(key);
+    let ok = true;
+    try {
+      fonts[f].encodeText(text);
+    } catch (e) {
+      ok = false;
+    }
+    if (encodable.size < 20000) encodable.set(key, ok);
+    return ok;
+  };
+  const measure = (text, f, size) => fonts[f].widthOfTextAtSize(text, size);
+  const images = opts.images || new Map();
+  const layout = layoutDocument(blocks, { page: opts.page, margin: opts.margin, base: opts.base, numbers: opts.numbers, measure, canEncode, image: (src) => (images.get(src) ? { w: images.get(src).w, h: images.get(src).h } : null) });
+  const embedded = new Map();
+  const color = (c) => L.rgb(c[0], c[1], c[2]);
+  for (const p of layout.pages) {
+    const page = doc.addPage([layout.width, layout.height]);
+    for (const o of p.ops) {
+      if (o.op === 'text') {
+        const t = { x: o.x, y: o.y, size: o.size, font: fonts[o.font] || fonts.regular, color: color(o.color) };
+        try {
+          page.drawText(o.text, t);
+        } catch (e) {
+          page.drawText(o.text.replace(/[^\x20-\x7e\u00a0-\u00ff]/g, '?'), t);
+        }
+      } else if (o.op === 'rect') {
+        const r = { x: o.x, y: o.y, width: o.w, height: o.h };
+        if (o.fill) r.color = color(o.fill);
+        if (o.stroke) {
+          r.borderColor = color(o.stroke);
+          r.borderWidth = o.sw || 1;
+        }
+        page.drawRectangle(r);
+      } else if (o.op === 'line') {
+        page.drawLine({ start: { x: o.x1, y: o.y1 }, end: { x: o.x2, y: o.y2 }, thickness: o.w, color: color(o.color) });
+      } else if (o.op === 'circle') {
+        const c = { x: o.x, y: o.y, size: o.r };
+        if (o.fill) c.color = color(o.fill);
+        if (o.stroke) {
+          c.borderColor = color(o.stroke);
+          c.borderWidth = o.sw || 1;
+        }
+        page.drawCircle(c);
+      } else if (o.op === 'path') {
+        for (let i = 1; i < o.points.length; i++) page.drawLine({ start: { x: o.points[i - 1][0], y: o.points[i - 1][1] }, end: { x: o.points[i][0], y: o.points[i][1] }, thickness: o.w, color: color(o.color), lineCap: L.LineCapStyle.Round });
+      } else if (o.op === 'image') {
+        const src = images.get(o.src);
+        if (!src) continue;
+        let img = embedded.get(o.src);
+        if (!img) {
+          img = src.mime === 'image/jpeg' ? await doc.embedJpg(src.bytes) : await doc.embedPng(src.bytes);
+          embedded.set(o.src, img);
+        }
+        page.drawImage(img, { x: o.x, y: o.y, width: o.w, height: o.h });
+      } else if (o.op === 'link') {
+        link(L, doc, page, o);
+      }
+    }
+  }
+  if (opts.title) doc.setTitle(opts.title);
+  doc.setCreator('Snowkit');
+  doc.setProducer('Snowkit');
+  const bytes = await doc.save();
+  return { bytes, pages: layout.pages.length, skipped: layout.skipped, links: layout.links };
+}
+
+export function imageSources(blocks, out = new Set()) {
+  for (const b of blocks) {
+    if (b.type === 'image' && b.src) out.add(b.src);
+    if (b.type === 'table') for (const r of b.rows) for (const c of r.cells) imageSources(c.blocks, out);
+  }
+  return out;
+}

package/src/engine/html.js

@@ -0,0 +1,59 @@
+const VOID = new Set(["area", "base", "br", "col", "embed", "hr", "img", "input", "link", "meta", "source", "track", "wbr"]);
+const RAW = new Set(["script", "style", "textarea", "title"]);
+const AUTO_CLOSE = { p: /^(p|div|ul|ol|table|h[1-6]|pre|blockquote|section|article|header|footer|hr|dl)$/, li: /^li$/, dt: /^(dt|dd)$/, dd: /^(dt|dd)$/, tr: /^tr$/, td: /^(td|th|tr)$/, th: /^(td|th|tr)$/, option: /^option$/ };
+const ENT = { amp: "&", lt: "<", gt: ">", quot: '"', apos: "'", nbsp: "\u00a0", copy: "\u00a9", reg: "\u00ae", hellip: "\u2026", mdash: "\u2014", ndash: "\u2013", lsquo: "\u2018", rsquo: "\u2019", ldquo: "\u201c", rdquo: "\u201d", bull: "\u2022", middot: "\u00b7", trade: "\u2122", euro: "\u20ac", pound: "\u00a3" };
+
+export function decode(s) {
+  return s.replace(/&(#x[0-9a-f]+|#\d+|[a-z]+);/gi, (m, e) => {
+    if (e[0] === "#") {
+      const code = e[1] === "x" || e[1] === "X" ? parseInt(e.slice(2), 16) : parseInt(e.slice(1), 10);
+      return Number.isFinite(code) && code > 0 && code < 0x110000 ? String.fromCodePoint(code) : m;
+    }
+    return ENT[e.toLowerCase()] ?? m;
+  });
+}
+
+function el(name, attrs, parent) {
+  return { nodeType: 1, tagName: name, attrs, childNodes: [], parentNode: parent };
+}
+
+export function parseHTML(src) {
+  const root = el("#root", {}, null);
+  let cur = root;
+  const re = /<!--[\s\S]*?-->|<!doctype[^>]*>|<\/([a-zA-Z][\w:-]*)\s*>|<([a-zA-Z][\w:-]*)((?:\s+[^\s"'>/=]+(?:\s*=\s*(?:"[^"]*"|'[^']*'|[^\s"'=<>`]+))?)*)\s*(\/?)>/gi;
+  let at = 0;
+  let m;
+  const text = (s) => {
+    if (s) cur.childNodes.push({ nodeType: 3, nodeValue: decode(s), parentNode: cur });
+  };
+  while ((m = re.exec(src))) {
+    text(src.slice(at, m.index));
+    at = re.lastIndex;
+    if (m[0].startsWith("<!")) continue;
+    if (m[1]) {
+      const name = m[1].toLowerCase();
+      let n = cur;
+      while (n && n.tagName !== name) n = n.parentNode;
+      if (n && n !== root) cur = n.parentNode;
+      continue;
+    }
+    const name = m[2].toLowerCase();
+    const attrs = {};
+    for (const a of m[3].matchAll(/([^\s"'>/=]+)(?:\s*=\s*(?:"([^"]*)"|'([^']*)'|([^\s"'=<>`]+)))?/g)) attrs[a[1].toLowerCase()] = decode(a[2] ?? a[3] ?? a[4] ?? "");
+    while (AUTO_CLOSE[cur.tagName] && AUTO_CLOSE[cur.tagName].test(name)) cur = cur.parentNode;
+    const node = el(name, attrs, cur);
+    cur.childNodes.push(node);
+    if (VOID.has(name) || m[4]) continue;
+    if (RAW.has(name)) {
+      const end = src.toLowerCase().indexOf("</" + name, at);
+      const stop = end < 0 ? src.length : end;
+      node.childNodes.push({ nodeType: 3, nodeValue: name === "textarea" || name === "title" ? decode(src.slice(at, stop)) : src.slice(at, stop), parentNode: node });
+      const close = src.indexOf(">", stop);
+      at = re.lastIndex = close < 0 ? src.length : close + 1;
+      continue;
+    }
+    cur = node;
+  }
+  text(src.slice(at));
+  return root;
+}