kordoc 3.8.2 → 3.8.3
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +7 -0
- package/dist/{-FEHSMPVO.js → -F5IXBLQQ.js} +3 -3
- package/dist/{chunk-553VTUVP.js → chunk-7J73IGMK.js} +2 -2
- package/dist/{chunk-DP37KF2X.js → chunk-B27QMVL7.js} +39 -4
- package/dist/chunk-B27QMVL7.js.map +1 -0
- package/dist/{chunk-JHZUFBUV.js → chunk-KIWKBRCJ.js} +2 -2
- package/dist/{chunk-YBPNKFJW.cjs → chunk-XL6O3VAY.cjs} +2 -2
- package/dist/{chunk-YBPNKFJW.cjs.map → chunk-XL6O3VAY.cjs.map} +1 -1
- package/dist/cli.js +4 -4
- package/dist/index.cjs +251 -216
- package/dist/index.cjs.map +1 -1
- package/dist/index.js +38 -3
- package/dist/index.js.map +1 -1
- package/dist/mcp.js +3 -3
- package/dist/{parser-KNQDRLZQ.js → parser-6GP535ZB.js} +153 -7
- package/dist/parser-6GP535ZB.js.map +1 -0
- package/dist/{parser-NR2TYGO3.js → parser-TM3AS25T.js} +153 -7
- package/dist/parser-TM3AS25T.js.map +1 -0
- package/dist/{parser-NS4ZPD7B.cjs → parser-WWKYMRGJ.cjs} +177 -31
- package/dist/parser-WWKYMRGJ.cjs.map +1 -0
- package/dist/{watch-XCWADLPU.js → watch-HETTZ7BO.js} +3 -3
- package/package.json +2 -1
- package/dist/chunk-DP37KF2X.js.map +0 -1
- package/dist/parser-KNQDRLZQ.js.map +0 -1
- package/dist/parser-NR2TYGO3.js.map +0 -1
- package/dist/parser-NS4ZPD7B.cjs.map +0 -1
- /package/dist/{-FEHSMPVO.js.map → -F5IXBLQQ.js.map} +0 -0
- /package/dist/{chunk-553VTUVP.js.map → chunk-7J73IGMK.js.map} +0 -0
- /package/dist/{chunk-JHZUFBUV.js.map → chunk-KIWKBRCJ.js.map} +0 -0
- /package/dist/{watch-XCWADLPU.js.map → watch-HETTZ7BO.js.map} +0 -0
package/dist/index.js
CHANGED
|
@@ -19,7 +19,7 @@ import {
|
|
|
19
19
|
sanitizeHref,
|
|
20
20
|
stripDtd,
|
|
21
21
|
toArrayBuffer
|
|
22
|
-
} from "./chunk-
|
|
22
|
+
} from "./chunk-7J73IGMK.js";
|
|
23
23
|
import {
|
|
24
24
|
parsePageRange
|
|
25
25
|
} from "./chunk-GE43BE46.js";
|
|
@@ -16799,6 +16799,9 @@ function getElements(parent, tagName) {
|
|
|
16799
16799
|
const nodes = parent.getElementsByTagName(tagName);
|
|
16800
16800
|
const result = [];
|
|
16801
16801
|
for (let i = 0; i < nodes.length; i++) result.push(nodes[i]);
|
|
16802
|
+
if (result.length > 0) return result;
|
|
16803
|
+
const nsNodes = parent.getElementsByTagNameNS?.("*", tagName);
|
|
16804
|
+
if (nsNodes) for (let i = 0; i < nsNodes.length; i++) result.push(nsNodes[i]);
|
|
16802
16805
|
return result;
|
|
16803
16806
|
}
|
|
16804
16807
|
function getTextContent(el) {
|
|
@@ -18622,6 +18625,7 @@ function walkContent(node, blocks, paraShapeMap, sectionNum, warnings, inHeaderF
|
|
|
18622
18625
|
if (tag === "P") {
|
|
18623
18626
|
if (!inHeaderFooter) {
|
|
18624
18627
|
parseParagraph3(el, blocks, paraShapeMap, sectionNum);
|
|
18628
|
+
walkTablesInP(el, blocks, paraShapeMap, sectionNum, warnings);
|
|
18625
18629
|
}
|
|
18626
18630
|
continue;
|
|
18627
18631
|
}
|
|
@@ -18638,6 +18642,21 @@ function walkContent(node, blocks, paraShapeMap, sectionNum, warnings, inHeaderF
|
|
|
18638
18642
|
walkContent(el, blocks, paraShapeMap, sectionNum, warnings, inHeaderFooter, depth + 1);
|
|
18639
18643
|
}
|
|
18640
18644
|
}
|
|
18645
|
+
function walkTablesInP(node, blocks, paraShapeMap, sectionNum, warnings, depth = 0) {
|
|
18646
|
+
if (depth > MAX_XML_DEPTH2) return;
|
|
18647
|
+
const children = node.childNodes;
|
|
18648
|
+
for (let i = 0; i < children.length; i++) {
|
|
18649
|
+
const el = children[i];
|
|
18650
|
+
if (el.nodeType !== 1) continue;
|
|
18651
|
+
const tag = localName(el);
|
|
18652
|
+
if (tag === "TABLE") {
|
|
18653
|
+
parseTable2(el, blocks, paraShapeMap, sectionNum, warnings);
|
|
18654
|
+
continue;
|
|
18655
|
+
}
|
|
18656
|
+
if (tag === "FOOTNOTE" || tag === "ENDNOTE" || tag === "HEADER" || tag === "FOOTER") continue;
|
|
18657
|
+
walkTablesInP(el, blocks, paraShapeMap, sectionNum, warnings, depth + 1);
|
|
18658
|
+
}
|
|
18659
|
+
}
|
|
18641
18660
|
function parseParagraph3(el, blocks, paraShapeMap, sectionNum) {
|
|
18642
18661
|
const paraShapeId = el.getAttribute("ParaShape") ?? "";
|
|
18643
18662
|
const shapeInfo = paraShapeMap.get(paraShapeId);
|
|
@@ -18733,13 +18752,29 @@ function collectCellText(node, parts, depth) {
|
|
|
18733
18752
|
if (tag === "P") {
|
|
18734
18753
|
const t = extractParagraphText(el);
|
|
18735
18754
|
if (t) parts.push(t);
|
|
18755
|
+
collectNestedTableText(el, parts, depth + 1);
|
|
18736
18756
|
} else if (tag === "TABLE") {
|
|
18737
|
-
|
|
18757
|
+
collectCellText(el, parts, depth + 1);
|
|
18738
18758
|
} else {
|
|
18739
18759
|
collectCellText(el, parts, depth + 1);
|
|
18740
18760
|
}
|
|
18741
18761
|
}
|
|
18742
18762
|
}
|
|
18763
|
+
function collectNestedTableText(node, parts, depth) {
|
|
18764
|
+
if (depth > 20) return;
|
|
18765
|
+
const children = node.childNodes;
|
|
18766
|
+
for (let i = 0; i < children.length; i++) {
|
|
18767
|
+
const el = children[i];
|
|
18768
|
+
if (el.nodeType !== 1) continue;
|
|
18769
|
+
const tag = localName(el);
|
|
18770
|
+
if (tag === "TABLE") {
|
|
18771
|
+
collectCellText(el, parts, depth + 1);
|
|
18772
|
+
continue;
|
|
18773
|
+
}
|
|
18774
|
+
if (tag === "FOOTNOTE" || tag === "ENDNOTE" || tag === "HEADER" || tag === "FOOTER") continue;
|
|
18775
|
+
collectNestedTableText(el, parts, depth + 1);
|
|
18776
|
+
}
|
|
18777
|
+
}
|
|
18743
18778
|
function localName(el) {
|
|
18744
18779
|
return (el.tagName || el.localName || "").replace(/^[^:]+:/, "");
|
|
18745
18780
|
}
|
|
@@ -24515,7 +24550,7 @@ async function parseHwp(buffer, options) {
|
|
|
24515
24550
|
async function parsePdf(buffer, options) {
|
|
24516
24551
|
let parsePdfDocument;
|
|
24517
24552
|
try {
|
|
24518
|
-
const mod = await import("./parser-
|
|
24553
|
+
const mod = await import("./parser-TM3AS25T.js");
|
|
24519
24554
|
parsePdfDocument = mod.parsePdfDocument;
|
|
24520
24555
|
} catch {
|
|
24521
24556
|
return {
|