@taloswork/ose-shared 0.1.11 → 0.1.13

package/skills/document/bin/doc-xml.cjs

@@ -9863,7 +9863,7 @@ function replaceTextAcrossRuns(xml, find, replacement, opts) {
     let flat = "";
     for (const m of para.matchAll(textRe)) {
       const text = unescXml(m[2]);
-      segs.push({ el: m[0], attrs: m[1] ?? "", text, start: flat.length });
+      segs.push({ el: m[0], at: m.index, attrs: m[1] ?? "", text, start: flat.length });
       flat += text;
     }
     if (!flat.includes(find))
@@ -9892,7 +9892,7 @@ function replaceTextAcrossRuns(xml, find, replacement, opts) {
       return pieces.join("");
     });
     let out = para;
-    for (let i = 0; i < segs.length; i++) {
+    for (let i = segs.length - 1; i >= 0; i--) {
       if (newTexts[i] === segs[i].text)
         continue;
       const escaped = escXml(newTexts[i]);
@@ -9900,7 +9900,7 @@ function replaceTextAcrossRuns(xml, find, replacement, opts) {
       if (textTag === "w:t" && /^\s|\s$/.test(newTexts[i]) && !/xml:space="preserve"/.test(attrs)) {
         attrs = `${attrs} xml:space="preserve"`;
       }
-      out = out.replace(segs[i].el, `<${textTag}${attrs}>${escaped}</${textTag}>`);
+      out = out.slice(0, segs[i].at) + `<${textTag}${attrs}>${escaped}</${textTag}>` + out.slice(segs[i].at + segs[i].el.length);
     }
     return out;
   });
@@ -10029,7 +10029,7 @@ async function replaceTextInDocx(bytes, find, replacement) {
   }
   if (!zip.file("word/document.xml"))
     throw new Error("not a Word document (word/document.xml missing)");
-  const parts = Object.keys(zip.files).filter((p) => p === "word/document.xml" || /^word\/(?:header|footer)\d+\.xml$/.test(p));
+  const parts = Object.keys(zip.files).filter((p) => p === "word/document.xml" || p === "word/footnotes.xml" || p === "word/endnotes.xml" || /^word\/(?:header|footer)\d+\.xml$/.test(p));
   let count = 0;
   for (const part of parts) {
     const xml = await zip.file(part).async("string");
@@ -10100,7 +10100,11 @@ async function main() {
       const r = await replaceTextInDocx(bytes, find, replacement);
       await (0, import_promises2.writeFile)(out, Buffer.from(r.bytes));
       process.stdout.write(JSON.stringify({ ok: true, out, replaced: r.count }) + "\n");
-      if (r.count === 0) process.stderr.write("doc-xml: no occurrence found (text may differ or sit in a table/textbox part)\n");
+      if (r.count === 0) {
+        process.stderr.write(
+          "doc-xml: no occurrence found \u2014 the text must match exactly within one paragraph (body, tables, headers/footers and footnotes/endnotes are searched; comments are not)\n"
+        );
+      }
       return;
     }
   } catch (e) {

package/skills/document/bin/docx-annotate.cjs

@@ -9830,6 +9830,7 @@ function escXml(s) {
 var import_node_crypto = require("node:crypto");
 var W_NS = "http://schemas.openxmlformats.org/wordprocessingml/2006/main";
 var W14_NS = "http://schemas.microsoft.com/office/word/2010/wordml";
+var MC_NS = "http://schemas.openxmlformats.org/markup-compatibility/2006";
 var W15_NS = "http://schemas.microsoft.com/office/word/2012/wordml";
 var W16CID_NS = "http://schemas.microsoft.com/office/word/2016/wordml/cid";
 var MODERN_PARTS = [
@@ -9926,11 +9927,34 @@ async function ensureCommentsWiring(zip) {
   }
   const existing = zip.file("word/comments.xml");
   if (!existing) {
-    zip.file("word/comments.xml", `<?xml version="1.0" encoding="UTF-8" standalone="yes"?><w:comments xmlns:w="${W_NS}" xmlns:w14="${W14_NS}"></w:comments>`);
+    zip.file("word/comments.xml", `<?xml version="1.0" encoding="UTF-8" standalone="yes"?><w:comments xmlns:w="${W_NS}" xmlns:mc="${MC_NS}" xmlns:w14="${W14_NS}" mc:Ignorable="w14"></w:comments>`);
     return 0;
   }
   const body = await existing.async("string");
-  return (body.match(/<w:comment\b/g) ?? []).length;
+  return maxId(body, /<w:comment\b[^>]*?\sw:id="(\d+)"/g) + 1;
+}
+function maxId(xml, re) {
+  let max = -1;
+  for (const m of xml.matchAll(re))
+    max = Math.max(max, Number(m[1]));
+  return max;
+}
+function prepareCommentsRoot(xml) {
+  xml = xml.replace(/<w:comments\b([^>]*?)\s*\/>/, "<w:comments$1></w:comments>");
+  return xml.replace(/<w:comments\b[^>]*>/, (open) => {
+    let tag = open.slice(0, -1);
+    if (!/\sxmlns:w14=/.test(tag))
+      tag += ` xmlns:w14="${W14_NS}"`;
+    if (!/\sxmlns:mc=/.test(tag))
+      tag += ` xmlns:mc="${MC_NS}"`;
+    const ign = /\smc:Ignorable="([^"]*)"/.exec(tag);
+    if (!ign)
+      tag += ` mc:Ignorable="w14"`;
+    else if (!ign[1].split(/\s+/).includes("w14")) {
+      tag = tag.replace(ign[0], ` mc:Ignorable="${`${ign[1]} w14`.trim()}"`);
+    }
+    return `${tag}>`;
+  });
 }
 async function addComments(bytes, comments) {
   if (!comments.length)
@@ -9943,12 +9967,13 @@ async function addComments(bytes, comments) {
   let doc = documentXml;
   let commentsBody = "";
   const skipped = [];
+  const warnings = [];
   const paraIds = comments.map(() => void 0);
   const modern = [];
   comments.forEach((c, batchIdx) => {
     const target = c.anchor ? paras.find((p) => textOf(p.full).includes(c.anchor)) ?? null : paras[0];
     if (c.anchor && !target) {
-      skipped.push(c.anchor);
+      skipped.push({ index: batchIdx, anchor: c.anchor });
       return;
     }
     const id = startId++;
@@ -9958,27 +9983,67 @@ async function addComments(bytes, comments) {
     doc = doc.replace(target.full, wrapped);
     target.full = wrapped;
     commentsBody += commentXml(id, c, paraId);
-    const parentParaId = c.replyTo != null ? paraIds[c.replyTo] : void 0;
+    let parentParaId;
+    if (c.replyTo != null) {
+      parentParaId = paraIds[c.replyTo];
+      if (!parentParaId) {
+        warnings.push(c.replyTo >= 0 && c.replyTo < comments.length && skipped.some((s) => s.index === c.replyTo) ? `comment ${batchIdx}: replyTo ${c.replyTo} was skipped (anchor not found) \u2014 written as a top-level comment` : `comment ${batchIdx}: replyTo ${c.replyTo} does not refer to an earlier written comment in this batch \u2014 written as a top-level comment`);
+      }
+    }
     modern.push({ paraId, parentParaId, author: c.author });
   });
   if (skipped.length === comments.length)
-    throw new Error(`no anchor matched: ${skipped.join(", ")}`);
+    throw new Error(`no anchor matched: ${skipped.map((s) => s.anchor).join(", ")}`);
   zip.file("word/document.xml", doc);
-  const cx = await zip.file("word/comments.xml").async("string");
-  const cxWithNs = cx.includes("xmlns:w14") ? cx : cx.replace(`xmlns:w="${W_NS}"`, `xmlns:w="${W_NS}" xmlns:w14="${W14_NS}"`);
-  zip.file("word/comments.xml", cxWithNs.replace("</w:comments>", `${commentsBody}</w:comments>`));
+  const cx = prepareCommentsRoot(await zip.file("word/comments.xml").async("string"));
+  if (!cx.includes("</w:comments>"))
+    throw new Error("word/comments.xml has no <w:comments> root");
+  const cxOut = cx.replace("</w:comments>", `${commentsBody}</w:comments>`);
+  const bodies = new Set([...cxOut.matchAll(/<w:comment\b[^>]*?\sw:id="(\d+)"/g)].map((m) => m[1]));
+  const dangling = [...new Set([...doc.matchAll(/<w:comment(?:RangeStart|RangeEnd|Reference)\b[^>]*?\sw:id="(\d+)"/g)].map((m) => m[1]))].filter((id) => !bodies.has(id));
+  if (dangling.length)
+    throw new Error(`internal: comment id(s) ${dangling.join(", ")} referenced in document.xml have no <w:comment> body`);
+  zip.file("word/comments.xml", cxOut);
   await writeModernCommentParts(zip, modern);
   const out = await zip.generateAsync({ type: "nodebuffer", compression: "DEFLATE" });
-  return new Uint8Array(out);
+  return { bytes: new Uint8Array(out), written: comments.length - skipped.length, skipped, warnings };
+}
+function bodySectPrOffset(documentXml, bodyEnd) {
+  let found = -1;
+  let pPrDepth = 0;
+  let changeDepth = 0;
+  const re = /<(\/?)w:(pPr|sectPrChange|sectPr)\b[^>]*?(\/?)>/g;
+  re.lastIndex = Math.max(0, documentXml.indexOf("<w:body"));
+  let m;
+  while ((m = re.exec(documentXml)) && m.index < bodyEnd) {
+    const [, close, name, selfClose] = m;
+    if (selfClose) {
+      if (name === "sectPr" && pPrDepth === 0 && changeDepth === 0)
+        found = m.index;
+      continue;
+    }
+    const d = close ? -1 : 1;
+    if (name === "pPr")
+      pPrDepth += d;
+    else if (name === "sectPrChange")
+      changeDepth += d;
+    else if (!close && pPrDepth === 0 && changeDepth === 0)
+      found = m.index;
+  }
+  return found;
 }
 async function addTrackedInsertion(bytes, ins) {
   const { zip, documentXml } = await loadDocx(bytes);
-  const date = ins.date ?? "1970-01-01T00:00:00Z";
-  const para = `<w:p><w:ins w:id="900001" w:author="${escXml(ins.author)}" w:date="${escXml(date)}"><w:r><w:t xml:space="preserve">${escXml(ins.text)}</w:t></w:r></w:ins></w:p>`;
-  const doc = documentXml.includes("<w:sectPr") ? documentXml.replace(/(<w:sectPr[\s>])/, `${para}$1`) : documentXml.replace("</w:body>", `${para}</w:body>`);
-  if (doc === documentXml)
+  const date = escXml(ins.date ?? "1970-01-01T00:00:00Z");
+  const author = escXml(ins.author);
+  const base = maxId(documentXml, /\sw:id="(\d+)"/g) + 1;
+  const para = `<w:p><w:pPr><w:rPr><w:ins w:id="${base}" w:author="${author}" w:date="${date}"/></w:rPr></w:pPr><w:ins w:id="${base + 1}" w:author="${author}" w:date="${date}"><w:r><w:t xml:space="preserve">${escXml(ins.text)}</w:t></w:r></w:ins></w:p>`;
+  const bodyEnd = documentXml.lastIndexOf("</w:body>");
+  if (bodyEnd < 0)
     throw new Error("could not locate document body");
-  zip.file("word/document.xml", doc);
+  const sect = bodySectPrOffset(documentXml, bodyEnd);
+  const at = sect >= 0 ? sect : bodyEnd;
+  zip.file("word/document.xml", documentXml.slice(0, at) + para + documentXml.slice(at));
   const out = await zip.generateAsync({ type: "nodebuffer", compression: "DEFLATE" });
   return new Uint8Array(out);
 }
@@ -10095,10 +10160,12 @@ function die(msg, code = 1) {
 }
 var nowIso = () => (/* @__PURE__ */ new Date()).toISOString().replace(/\.\d+Z$/, "Z");
 async function main() {
-  const [mode, inDocx, out, a, b] = process.argv.slice(2);
+  const argv = process.argv.slice(2);
+  const strictAnchors = argv.includes("--strict-anchors");
+  const [mode, inDocx, out, a, b] = argv.filter((x) => x !== "--strict-anchors");
   try {
     if (mode === "comment") {
-      if (!inDocx || !out || !a) die("usage: docx-annotate comment <in.docx> <out.docx> <comments.json>", 64);
+      if (!inDocx || !out || !a) die("usage: docx-annotate comment <in.docx> <out.docx> <comments.json> [--strict-anchors]", 64);
       let comments;
       try {
         comments = JSON.parse(await (0, import_promises.readFile)(a, "utf8"));
@@ -10107,9 +10174,18 @@ async function main() {
       }
       if (!Array.isArray(comments) || !comments.length) die("comments json must be a non-empty array", 64);
       const bytes = new Uint8Array(await (0, import_promises.readFile)(inDocx).catch((e) => die(`cannot read ${inDocx}: ${e.message}`)));
-      const result = await addComments(bytes, comments.map((c) => ({ date: nowIso(), ...c })));
-      await (0, import_promises.writeFile)(out, Buffer.from(result));
-      process.stdout.write(JSON.stringify({ ok: true, out, comments: comments.length }) + "\n");
+      const r = await addComments(bytes, comments.map((c) => ({ date: nowIso(), ...c })));
+      await (0, import_promises.writeFile)(out, Buffer.from(r.bytes));
+      for (const w of r.warnings) process.stderr.write(`docx-annotate: warning: ${w}
+`);
+      if (r.skipped.length) {
+        process.stderr.write(
+          `docx-annotate: warning: ${r.skipped.length} of ${comments.length} comment(s) NOT written \u2014 anchor matched no paragraph: ` + r.skipped.map((s) => `#${s.index} ${JSON.stringify(s.anchor)}`).join(", ") + " (anchors must match text within one paragraph)\n"
+        );
+      }
+      const ok = !(strictAnchors && r.skipped.length);
+      process.stdout.write(JSON.stringify({ ok, out, written: r.written, skipped: r.skipped, warnings: r.warnings }) + "\n");
+      if (!ok) process.exit(1);
       return;
     }
     if (mode === "suggest") {