@veryfront/ext-document-kreuzberg 0.1.1278 → 0.1.1279-rc.24504

package/esm/extensions/ext-document-kreuzberg/src/native-extraction.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"native-extraction.d.ts","sourceRoot":"","sources":["../../../../src/extensions/ext-document-kreuzberg/src/native-extraction.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAIH,OAAO,KAAK,EACV,+BAA+B,EAC/B,kBAAkB,EACnB,MAAM,yCAAyC,CAAC;AAIjD,0DAA0D;AAC1D,MAAM,MAAM,oBAAoB,GAAG,YAAY,GAAG,UAAU,CAAC;AAE7D,MAAM,WAAW,uBAAuB;IACtC,IAAI,EAAE,oBAAoB,CAAC;IAC3B,YAAY,EAAE,CAAC,KAAK,EAAE,+BAA+B,KAAK,IAAI,CAAC;IAC/D,kEAAkE;IAClE,UAAU,CAAC,EAAE,MAAM,OAAO,CAAC,kBAAkB,CAAC,CAAC;CAChD;AAsWD;;;;;GAKG;AACH,wBAAsB,qBAAqB,CACzC,MAAM,EAAE,WAAW,EACnB,QAAQ,EAAE,MAAM,EAChB,OAAO,EAAE,uBAAuB,GAC/B,OAAO,CAAC,MAAM,CAAC,CAYjB"}
+{"version":3,"file":"native-extraction.d.ts","sourceRoot":"","sources":["../../../../src/extensions/ext-document-kreuzberg/src/native-extraction.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAIH,OAAO,KAAK,EACV,+BAA+B,EAC/B,kBAAkB,EACnB,MAAM,yCAAyC,CAAC;AAIjD,0DAA0D;AAC1D,MAAM,MAAM,oBAAoB,GAAG,YAAY,GAAG,UAAU,CAAC;AAE7D,MAAM,WAAW,uBAAuB;IACtC,IAAI,EAAE,oBAAoB,CAAC;IAC3B,YAAY,EAAE,CAAC,KAAK,EAAE,+BAA+B,KAAK,IAAI,CAAC;IAC/D,kEAAkE;IAClE,UAAU,CAAC,EAAE,MAAM,OAAO,CAAC,kBAAkB,CAAC,CAAC;CAChD;AAkXD;;;;;GAKG;AACH,wBAAsB,qBAAqB,CACzC,MAAM,EAAE,WAAW,EACnB,QAAQ,EAAE,MAAM,EAChB,OAAO,EAAE,uBAAuB,GAC/B,OAAO,CAAC,MAAM,CAAC,CAYjB"}

package/esm/extensions/ext-document-kreuzberg/src/native-extraction.js

@@ -11,11 +11,26 @@ function isPdfMimeType(mimeType) {
 function isPptxMimeType(mimeType) {
   return normalizeMimeType(mimeType) === "application/vnd.openxmlformats-officedocument.presentationml.presentation";
 }
+const capturedTextReplace = String.prototype.replace;
+const applyTextReplace = Reflect.apply;
+function decodeNamedEntities(value) {
+  let decoded = value;
+  decoded = applyTextReplace(capturedTextReplace, decoded, [/"/g, '"']);
+  decoded = applyTextReplace(capturedTextReplace, decoded, [/'/g, "'"]);
+  decoded = applyTextReplace(capturedTextReplace, decoded, [/>/g, ">"]);
+  decoded = applyTextReplace(capturedTextReplace, decoded, [/&lt;/g, "<"]);
+  decoded = applyTextReplace(capturedTextReplace, decoded, [/&amp;/g, "&"]);
+  return decoded;
+}
 function decodeXmlText(value) {
-  return value.replace(
+  const decoded = value.replace(
     /&#x([0-9a-f]+);/gi,
     (_, hex) => String.fromCodePoint(Number.parseInt(hex, 16))
-  ).replace(/&#(\d+);/g, (_, decimal) => String.fromCodePoint(Number.parseInt(decimal, 10))).replace(/&quot;/g, '"').replace(/&apos;/g, "'").replace(/&gt;/g, ">").replace(/&lt;/g, "<").replace(/&amp;/g, "&");
+  ).replace(
+    /&#(\d+);/g,
+    (_, decimal) => String.fromCodePoint(Number.parseInt(decimal, 10))
+  );
+  return decodeNamedEntities(decoded);
 }
 function extractPptxSlideText(xml) {
   return Array.from(xml.matchAll(/<a:t(?:\s[^>]*)?>([\s\S]*?)<\/a:t>/g)).map((match) => decodeXmlText(match[1] ?? "").trim()).filter(Boolean).join(" ").trim();
@@ -24,7 +39,7 @@ function normalizePptxText(value) {
   return value.replace(/\s+/g, " ").trim();
 }
 function cleanExtractedMarkdown(value) {
-  return value.replace(/!\[[^\]\n]*\]\(\s*\)/g, "").replace(/[ \t]+\n/g, "\n").replace(/\n{3,}/g, "\n\n").trim();
+  return value.replace(/!\[[^\]\n]*\]\(\s*\)/g, "").replace(/(?<![ \t])[ \t]+\n/g, "\n").replace(/\n{3,}/g, "\n\n").trim();
 }
 function slideNumber(path) {
   return Number(path.match(/\/slide(\d+)\.xml$/)?.[1] ?? 0);
@@ -76,7 +91,7 @@ function normalizePptxMarkdownHeadings(markdown, xml) {
   if (roles.size === 0) return markdown;
   return cleanExtractedMarkdown(
     markdown.split("\n").map((line) => {
-      const heading = line.match(/^#{1,6}\s+(.+?)\s*$/);
+      const heading = line.trimEnd().match(/^#{1,6}\s+(\S.*)$/);
       if (!heading) return line;
       const text = heading[1] ?? "";
       const queue = roles.get(normalizePptxText(text));

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@veryfront/ext-document-kreuzberg",
-  "version": "0.1.1278",
+  "version": "0.1.1279-rc.24504",
   "description": "Veryfront first-party extension package for ext-document-kreuzberg",
   "keywords": [
     "veryfront",
@@ -69,7 +69,7 @@
     "pdf-lib": "1.17.1"
   },
   "peerDependencies": {
-    "veryfront": "^0.1.1278"
+    "veryfront": "^0.1.1279-rc.24504"
   },
   "type": "module",
   "types": "./esm/extensions/ext-document-kreuzberg/src/index.d.ts",
@@ -79,5 +79,5 @@
     "NOTICE",
     "README.md"
   ],
-  "gitHead": "bfdfe6942891bb8715360ec75ced35e28a4632e3"
+  "gitHead": "1b06e0208046774cd83290c8a3ab9306c2f73b7b"
 }