@taloswork/ose-shared 0.1.11 → 0.1.13

package/skills/document/bin/pptx-slide.cjs

@@ -10007,7 +10007,7 @@ 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))
@@ -10036,7 +10036,7 @@ function replaceTextAcrossRuns(xml, find, replacement, opts) {
       return pieces.join("");
     });
     let out2 = 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]);
@@ -10044,7 +10044,7 @@ 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"`;
       }
-      out2 = out2.replace(segs[i].el, `<${textTag}${attrs}>${escaped}</${textTag}>`);
+      out2 = out2.slice(0, segs[i].at) + `<${textTag}${attrs}>${escaped}</${textTag}>` + out2.slice(segs[i].at + segs[i].el.length);
     }
     return out2;
   });

package/skills/document/bin/pptx-tokenize.cjs

@@ -9910,14 +9910,27 @@ function paraText(para) {
     t += unescXml(m[1]);
   return t;
 }
-function txBodyAsToken(txBody, tokenText) {
+function leadRuns(para) {
+  const runs = [...para.matchAll(/<a:r\b[^>]*>[\s\S]*?<\/a:r>/g)].map((m) => m[0]);
+  if (runs.length < 2)
+    return [];
+  const rp = (r) => RE.rPr.exec(r)?.[0]?.replace(/\s(?:lang|altLang|dirty|err|smtClean|noProof|kern|spc)="[^"]*"/g, "");
+  const a = rp(runs[0]);
+  const b = rp(runs[1]);
+  return a && b && a !== b ? runs : [];
+}
+function txBodyAsToken(txBody, tokenText, restToken) {
   const bodyPr = RE.bodyPr.exec(txBody)?.[0] ?? "<a:bodyPr/>";
   const lstStyle = RE.lstStyle.exec(txBody)?.[0] ?? "";
   const firstPara = RE.para1.exec(txBody)?.[0] ?? "";
   const pPr = firstPara ? RE.pPr.exec(firstPara)?.[0] ?? "" : "";
   const firstRun = firstPara ? RE.firstRun.exec(firstPara)?.[0] ?? "" : "";
   const rPr = firstRun ? RE.rPr.exec(firstRun)?.[0] ?? "" : "";
-  const p = `<a:p>${pPr}<a:r>${rPr}<a:t>${tokenText}</a:t></a:r></a:p>`;
+  let p = `<a:p>${pPr}<a:r>${rPr}<a:t>${tokenText}</a:t></a:r></a:p>`;
+  if (restToken) {
+    const rPr2 = RE.rPr.exec(leadRuns(firstPara)[1])?.[0] ?? "";
+    p += `<a:p>${pPr}<a:r>${rPr2}<a:t>${restToken}</a:t></a:r></a:p>`;
+  }
   return `<p:txBody>${bodyPr}${lstStyle}${p}</p:txBody>`;
 }
 function tokenizeSlide(xml, slideIndex, warn) {
@@ -9947,7 +9960,8 @@ function tokenizeSlide(xml, slideIndex, warn) {
     const original = paras.join("\n");
     if (!phM && !original.trim())
       return sp;
-    const multiline = paras.filter((p) => p.trim().length > 0).length > 1;
+    const lead = leadRuns(RE.para1.exec(txBody)?.[0] ?? "").length > 0;
+    const multiline = lead || paras.filter((p) => p.trim().length > 0).length > 1;
     const token = mkToken(role);
     if (phM && !original.trim())
       warn(`slide ${slideIndex}: ${role} placeholder was empty \u2014 tokenized anyway`);
@@ -9956,12 +9970,13 @@ function tokenizeSlide(xml, slideIndex, warn) {
       role,
       type: "text",
       ...multiline ? { multiline: true } : {},
+      ...lead ? { leadStyle: true } : {},
       ...phType ? { phType } : {},
       ...phIdx ? { phIdx } : {},
       capacityChars: Math.max(1, Math.round(visualWidth(original) || 20)),
       ...original ? { original: original.slice(0, 120) } : {}
     });
-    return sp.replace(txBody, txBodyAsToken(txBody, `{{${token}}}`));
+    return sp.replace(txBody, txBodyAsToken(txBody, `{{${token}}}`, lead ? `{{${token}_rest}}` : void 0));
   });
   for (const m of out.matchAll(RE.pic)) {
     const rel = RE.blip.exec(m[0])?.[1];