formatferry 1.0.25 → 1.0.27

package/dist/index.js

@@ -22549,6 +22549,58 @@ function dedupAgainstWidgets(textItems, widgets) {
   }
   return kept;
 }
+function associateLabelsToWidgets(widgets, textItems) {
+  if (widgets.length === 0 || textItems.length === 0) return [widgets, textItems];
+  const consumedIndices = /* @__PURE__ */ new Set();
+  const updatedWidgets = widgets.map((w) => {
+    if (/^\[([ X]|Signed|Signature)\]$/.test(w.str)) return w;
+    const refSize = Math.max(w.height, 10);
+    const leftTolerance = refSize * 3;
+    const aboveTolerance = refSize * 1;
+    let bestCandidate = null;
+    for (let i = 0; i < textItems.length; i++) {
+      if (consumedIndices.has(i)) continue;
+      const txt = textItems[i];
+      const isToLeft = txt.right <= w.x + 2;
+      if (isToLeft) {
+        const hGap = w.x - txt.right;
+        if (hGap <= leftTolerance) {
+          const yOverlap = Math.min(txt.y + txt.height, w.y + w.height) - Math.max(txt.y, w.y);
+          const minH = Math.min(txt.height, w.height);
+          if (yOverlap / minH >= 0.5) {
+            const score = hGap;
+            if (!bestCandidate || score < bestCandidate.score) {
+              bestCandidate = { idx: i, str: txt.str, score };
+            }
+          }
+        }
+      }
+      if (!isToLeft) {
+        const widgetTop = w.y + w.height;
+        const labelBottom = txt.y;
+        const vGap = labelBottom - widgetTop;
+        if (vGap >= 0 && vGap <= aboveTolerance) {
+          const xOverlap = Math.min(txt.right, w.right) - Math.max(txt.x, w.x);
+          if (xOverlap > 0) {
+            const score = vGap;
+            if (!bestCandidate || score < bestCandidate.score) {
+              bestCandidate = { idx: i, str: txt.str, score };
+            }
+          }
+        }
+      }
+    }
+    if (bestCandidate) {
+      consumedIndices.add(bestCandidate.idx);
+      const labelStr = bestCandidate.str.trimEnd();
+      const separator = /[:：]\s*$/.test(labelStr) ? " " : ": ";
+      return { ...w, str: `${labelStr}${separator}${w.str}` };
+    }
+    return w;
+  });
+  const remainingText = textItems.filter((_, i) => !consumedIndices.has(i));
+  return [updatedWidgets, remainingText];
+}
 function itemsToLinearText(items) {
   if (items.length === 0) return "";
   const sorted = [...items].sort((a, b) => b.y + b.height - (a.y + a.height) || a.x - b.x);
@@ -23203,7 +23255,14 @@ var PdfAdapter = class {
           ...annotations.flatMap(widgetToPositionedItems),
           ...groupRadioButtons(annotations)
         ];
-        const items = widgetItems.length > 0 ? [...dedupAgainstWidgets(positionedText, widgetItems), ...widgetItems] : positionedText;
+        let items;
+        if (widgetItems.length > 0) {
+          const dedupedText = dedupAgainstWidgets(positionedText, widgetItems);
+          const [labelledWidgets, remainingText] = associateLabelsToWidgets(widgetItems, dedupedText);
+          items = [...remainingText, ...labelledWidgets];
+        } else {
+          items = positionedText;
+        }
         if (getFeatureFlags().enablePdfEnhanced) {
           applyLinkAnnotations(items, annotations);
         }
@@ -25761,8 +25820,8 @@ import { readFileSync as readFileSync4 } from "fs";
 import { resolve, dirname as dirname3 } from "path";
 import { fileURLToPath } from "url";
 var getVersion = () => {
-  if ("1.0.25") {
-    return "1.0.25";
+  if ("1.0.27") {
+    return "1.0.27";
   }
   try {
     const __dirname = dirname3(fileURLToPath(import.meta.url));

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "formatferry",
-  "version": "1.0.25",
+  "version": "1.0.27",
   "description": "FormatFerry CLI — Convert HTML, DOCX, PDF, XLSX, CSV, JSON, XML, PPTX to Markdown",
   "keywords": [
     "markdown",