@taloswork/ose-shared 0.1.12 → 0.1.13

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

@@ -29833,9 +29833,9 @@ var require_load2 = __commonJS({
 var require_lib4 = __commonJS({
   "node_modules/jszip/lib/index.js"(exports2, module2) {
     "use strict";
-    function JSZip() {
-      if (!(this instanceof JSZip)) {
-        return new JSZip();
+    function JSZip2() {
+      if (!(this instanceof JSZip2)) {
+        return new JSZip2();
       }
       if (arguments.length) {
         throw new Error("The constructor with parameters has been removed in JSZip 3.0, please check the upgrade guide.");
@@ -29844,7 +29844,7 @@ var require_lib4 = __commonJS({
       this.comment = null;
       this.root = "";
       this.clone = function() {
-        var newObj = new JSZip();
+        var newObj = new JSZip2();
         for (var i in this) {
           if (typeof this[i] !== "function") {
             newObj[i] = this[i];
@@ -29853,16 +29853,16 @@ var require_lib4 = __commonJS({
         return newObj;
       };
     }
-    JSZip.prototype = require_object2();
-    JSZip.prototype.loadAsync = require_load2();
-    JSZip.support = require_support2();
-    JSZip.defaults = require_defaults2();
-    JSZip.version = "3.10.1";
-    JSZip.loadAsync = function(content, options) {
-      return new JSZip().loadAsync(content, options);
+    JSZip2.prototype = require_object2();
+    JSZip2.prototype.loadAsync = require_load2();
+    JSZip2.support = require_support2();
+    JSZip2.defaults = require_defaults2();
+    JSZip2.version = "3.10.1";
+    JSZip2.loadAsync = function(content, options) {
+      return new JSZip2().loadAsync(content, options);
     };
-    JSZip.external = require_external();
-    module2.exports = JSZip;
+    JSZip2.external = require_external();
+    module2.exports = JSZip2;
   }
 });
 
@@ -32776,7 +32776,7 @@ var require_browser_buffer_encode = __commonJS({
 var require_zip_stream = __commonJS({
   "node_modules/exceljs/lib/utils/zip-stream.js"(exports2, module2) {
     var events = require("events");
-    var JSZip = require_lib4();
+    var JSZip2 = require_lib4();
     var StreamBuf = require_stream_buf();
     var { stringToBuffer } = require_browser_buffer_encode();
     var ZipWriter = class extends events.EventEmitter {
@@ -32789,7 +32789,7 @@ var require_zip_stream = __commonJS({
           },
           options
         );
-        this.zip = new JSZip();
+        this.zip = new JSZip2();
         this.stream = new StreamBuf();
       }
       append(data, options) {
@@ -44469,7 +44469,7 @@ var require_theme1 = __commonJS({
 var require_xlsx = __commonJS({
   "node_modules/exceljs/lib/xlsx/xlsx.js"(exports2, module2) {
     var fs = require("fs");
-    var JSZip = require_lib4();
+    var JSZip2 = require_lib4();
     var { PassThrough } = require_readable2();
     var ZipStream = require_zip_stream();
     var StreamBuf = require_stream_buf();
@@ -44708,7 +44708,7 @@ var require_xlsx = __commonJS({
           tables: {},
           vmlDrawings: {}
         };
-        const zip = await JSZip.loadAsync(buffer);
+        const zip = await JSZip2.loadAsync(buffer);
         for (const entry of Object.values(zip.files)) {
           if (!entry.dir) {
             let entryName = entry.name;
@@ -88983,6 +88983,8 @@ function fillOfficeTemplate(bytes, data, opts = {}) {
     throw new Error(`not a readable OOXML file: ${e.message}`);
   }
   const unfilled = /* @__PURE__ */ new Set();
+  const restNames = /* @__PURE__ */ new Set();
+  data = expandPptxLines(zip, data, restNames);
   let doc;
   try {
     doc = new import_docxtemplater.default(zip, {
@@ -89005,7 +89007,7 @@ function fillOfficeTemplate(bytes, data, opts = {}) {
       // record it so the caller can report what was left blank.
       nullGetter(part) {
         if (part.value)
-          unfilled.add(part.value);
+          unfilled.add(restNames.has(part.value) ? part.value.slice(0, -"_rest".length) : part.value);
         return "";
       }
     });
@@ -89016,6 +89018,50 @@ function fillOfficeTemplate(bytes, data, opts = {}) {
   const out = doc.getZip().generate({ type: "nodebuffer", compression: "DEFLATE" });
   return { bytes: new Uint8Array(out), unfilled: [...unfilled] };
 }
+var PPTX_SLIDE = /^ppt\/slides\/slide\d+\.xml$/;
+var SOLE_TOKEN_PARA = /^<a:p>((?:<a:pPr\b[^>]*(?:\/>|>[\s\S]*?<\/a:pPr>))?)<a:r>((?:<a:rPr\b[^>]*(?:\/>|>[\s\S]*?<\/a:rPr>))?)<a:t>\{\{([\w.]+)\}\}<\/a:t><\/a:r><\/a:p>$/;
+function expandPptxLines(zip, data, restNames) {
+  const out = { ...data };
+  for (const name of Object.keys(zip.files)) {
+    if (!PPTX_SLIDE.test(name))
+      continue;
+    const xml = zip.file(name).asText();
+    const paras = xml.match(/<a:p>[\s\S]*?<\/a:p>/g) ?? [];
+    let next = xml;
+    for (let i = 0; i < paras.length; i++) {
+      const m = SOLE_TOKEN_PARA.exec(paras[i]);
+      if (!m)
+        continue;
+      const tok = m[3];
+      const restM = SOLE_TOKEN_PARA.exec(paras[i + 1] ?? "");
+      const hasRest = !!restM && restM[3] === `${tok}_rest`;
+      const v = data[tok];
+      if (hasRest)
+        restNames.add(`${tok}_rest`);
+      if (v === void 0 || v === null)
+        continue;
+      const lines = typeof v === "string" && v.includes("\n") ? v.split(/\r?\n/) : null;
+      if (!lines && !hasRest)
+        continue;
+      const make = (para, from, key, text) => {
+        out[key] = text;
+        return para.replace(`{{${from}}}`, `{{${key}}}`);
+      };
+      const parts = [];
+      const all = lines ?? [typeof v === "string" ? v : v == null ? "" : String(v)];
+      all.forEach((line, k) => {
+        const useRest = hasRest && k > 0;
+        parts.push(make(useRest ? paras[i + 1] : paras[i], useRest ? `${tok}_rest` : tok, `${tok}__${k}`, line));
+      });
+      next = next.replace(paras[i] + (hasRest ? paras[i + 1] : ""), parts.join(""));
+      if (hasRest)
+        i++;
+    }
+    if (next !== xml)
+      zip.file(name, next);
+  }
+  return out;
+}
 function templateError(e) {
   const err = e;
   const inner = err.properties?.errors;
@@ -89073,6 +89119,65 @@ async function fillXlsxTemplate(bytes, data) {
   return { bytes: new Uint8Array(out), unfilled: [...unfilled] };
 }
 
+// packages/core/dist/document/template-residue.js
+var import_jszip = __toESM(require_lib4(), 1);
+
+// packages/core/dist/document/ooxml-util.js
+function unescXml(s) {
+  return s.replace(/&(amp|lt|gt|quot|apos|#x?[0-9a-fA-F]+);/g, (whole, code) => {
+    switch (code) {
+      case "amp":
+        return "&";
+      case "lt":
+        return "<";
+      case "gt":
+        return ">";
+      case "quot":
+        return '"';
+      case "apos":
+        return "'";
+      default: {
+        const n = code.startsWith("#x") || code.startsWith("#X") ? parseInt(code.slice(2), 16) : parseInt(code.slice(1), 10);
+        return Number.isInteger(n) && n >= 0 && n <= 1114111 ? String.fromCodePoint(n) : whole;
+      }
+    }
+  });
+}
+
+// packages/core/dist/document/template-residue.js
+function residueOf(templateTexts, outputTexts, fixed) {
+  const out = new Set(outputTexts.map((t) => t.trim()));
+  return [...new Set(templateTexts.map((t) => t.trim()))].filter((t) => [...t].length >= 8 && !fixed.has(t) && out.has(t));
+}
+async function pptxSlideParagraphs(bytes) {
+  const zip = await import_jszip.default.loadAsync(Buffer.from(bytes));
+  const names = Object.keys(zip.files).map((p) => ({ p, n: /^ppt\/slides\/slide(\d+)\.xml$/.exec(p)?.[1] })).filter((x) => Boolean(x.n)).sort((a, b) => Number(a.n) - Number(b.n));
+  const slides = [];
+  for (const { p } of names) {
+    const xml = await zip.file(p).async("string");
+    const paras = [];
+    for (const m of xml.matchAll(/<a:p\b[^>]*>[\s\S]*?<\/a:p>/g)) {
+      let t = "";
+      for (const r of m[0].matchAll(/<a:t\b[^>]*>([\s\S]*?)<\/a:t>/g))
+        t += unescXml(r[1]);
+      if (t.trim())
+        paras.push(t.trim());
+    }
+    slides.push(paras);
+  }
+  return slides;
+}
+var MIN_SLIDES = 3;
+function fixedTexts(slides, ratio = 0.8) {
+  if (slides.length < MIN_SLIDES)
+    return /* @__PURE__ */ new Set();
+  const count = /* @__PURE__ */ new Map();
+  for (const s of slides)
+    for (const t of new Set(s))
+      count.set(t, (count.get(t) ?? 0) + 1);
+  return new Set([...count].filter(([, c]) => slides.length > 0 && c / slides.length >= ratio).map(([t]) => t));
+}
+
 // apps/admin/data/skills.lib/document/bin-src/doc-fill.mjs
 function die(msg, code = 1) {
   process.stderr.write(`doc-fill: ${msg}
@@ -89100,7 +89205,23 @@ async function main() {
     die(e.message);
   }
   await (0, import_promises.writeFile)(out, Buffer.from(result.bytes)).catch((e) => die(`cannot write ${out}: ${e.message}`));
-  process.stdout.write(JSON.stringify({ ok: true, out, unfilled: result.unfilled }) + "\n");
+  let residue;
+  if (ext === ".pptx") {
+    try {
+      const tplSlides = await pptxSlideParagraphs(bytes);
+      const outSlides = await pptxSlideParagraphs(result.bytes);
+      residue = residueOf(tplSlides.flat(), outSlides.flat(), fixedTexts(tplSlides));
+    } catch (e) {
+      residue = null;
+      process.stderr.write(`doc-fill: residue check skipped: ${e.message}
+`);
+    }
+  }
+  process.stdout.write(JSON.stringify({ ok: true, out, unfilled: result.unfilled, residue }) + "\n");
+  if (residue && residue.length) {
+    process.stderr.write(`doc-fill: ${residue.length} sentence(s) of the template's own content are still in the deck \u2014 replace or delete them before delivering
+`);
+  }
   if (result.unfilled.length) {
     process.stderr.write(`doc-fill: left ${result.unfilled.length} placeholder(s) blank: ${result.unfilled.join(", ")}
 `);