reamkit 1.15.0 → 1.15.1

package/dist/esm/core/ole/cfb.js

@@ -30,6 +30,7 @@ function openCfb(bytes) {
 	const sectorSize = 1 << sectorShift;
 	const miniSectorSize = 1 << miniSectorShift;
 	const miniCutoff = view.getUint32(56, true);
+	const majorVersion = view.getUint16(26, true);
 	const totalSectors = Math.floor((bytes.length - HEADER_SIZE) / sectorSize);
 	const sectorOffset = (sector) => {
 		if (sector < 0 || sector >= totalSectors) throw new CfbError(`sector ${sector} out of range`);
@@ -86,19 +87,29 @@ function openCfb(bytes) {
 	const dirView = new DataView(dirBytes.buffer, dirBytes.byteOffset, dirBytes.byteLength);
 	const entries = [];
 	const entryCount = Math.floor(dirBytes.length / DIR_ENTRY_SIZE);
+	const rawByIndex = new Array(entryCount);
 	for (let i = 0; i < entryCount; i++) {
 		const base = i * DIR_ENTRY_SIZE;
 		const objType = dirView.getUint8(base + 66);
 		if (objType !== 1 && objType !== 2 && objType !== 5) continue;
 		const name = decodeName(dirBytes, base, dirView.getUint16(base + 64, true));
 		const sizeLow = dirView.getUint32(base + 120, true);
-		const size = dirView.getUint32(base + 124, true) * 4294967296 + sizeLow;
-		entries.push({
+		const sizeHigh = dirView.getUint32(base + 124, true);
+		const size = majorVersion === 3 ? sizeLow : sizeHigh * 4294967296 + sizeLow;
+		const entry = {
 			name,
 			type: objType === 5 ? "root" : objType === 1 ? "storage" : "stream",
 			startSector: dirView.getUint32(base + 116, true),
 			size
-		});
+		};
+		entries.push(entry);
+		rawByIndex[i] = {
+			entry,
+			left: dirView.getUint32(base + 68, true),
+			right: dirView.getUint32(base + 72, true),
+			child: dirView.getUint32(base + 76, true),
+			isRoot: objType === 5
+		};
 	}
 	const root = entries.find((e) => e.type === "root");
 	if (!root) throw new CfbError("compound file has no root entry");
@@ -137,8 +148,27 @@ function openCfb(bytes) {
 		if (entry.size >= miniCutoff) return readChainBytes(entry.startSector, MAX_TOTAL_BYTES).subarray(0, entry.size);
 		return readMiniStream(entry);
 	};
+	const topLevel = /* @__PURE__ */ new Set();
+	const rootRaw = rawByIndex.find((r) => r?.isRoot);
+	if (rootRaw) {
+		const stack = [rootRaw.child];
+		const seen = /* @__PURE__ */ new Set();
+		while (stack.length > 0) {
+			const idx = stack.pop();
+			if (idx >= entryCount || seen.has(idx)) continue;
+			seen.add(idx);
+			const node = rawByIndex[idx];
+			if (!node) continue;
+			topLevel.add(node.entry);
+			stack.push(node.left, node.right);
+		}
+	}
 	const byName = /* @__PURE__ */ new Map();
-	for (const e of entries) if (e.type === "stream" && !byName.has(e.name.toLowerCase())) byName.set(e.name.toLowerCase(), e);
+	const addStreams = (list) => {
+		for (const e of list) if (e.type === "stream" && !byName.has(e.name.toLowerCase())) byName.set(e.name.toLowerCase(), e);
+	};
+	addStreams(topLevel);
+	addStreams(entries);
 	return {
 		entries,
 		hasStream: (name) => byName.has(name.toLowerCase()),

package/dist/esm/pdf-reader/flow-build.d.ts

@@ -1,7 +1,8 @@
-import { BodyElement } from '../core/document-model/index.js';
+import { BodyElement, SectionProperties } from '../core/document-model/index.js';
 import { FlowDoc } from '../core/ir/flow.js';
 import { Loss, ResourceStore } from '../core/ir/index.js';
 import { PdfImage } from './images.js';
+import { PdfPage } from './document.js';
 import { PdfVector } from './vector.js';
 export interface Reconstruction {
     readonly doc: FlowDoc;
@@ -16,4 +17,5 @@ export declare function paragraphFromRuns(spans: ReadonlyArray<TextSpan>, outlin
 export declare function imageBlock(image: PdfImage, resources: ResourceStore, alt?: string): BodyElement;
 export declare function dedupeLosses(losses: ReadonlyArray<Loss>): Array<Loss>;
 export declare function shapeBlock(v: PdfVector): BodyElement;
-export declare function buildFlowDoc(body: ReadonlyArray<BodyElement>, resources?: ResourceStore): FlowDoc;
+export declare function sectionFromPdfPages(pages: ReadonlyArray<PdfPage>): SectionProperties | undefined;
+export declare function buildFlowDoc(body: ReadonlyArray<BodyElement>, resources?: ResourceStore, section?: SectionProperties): FlowDoc;

package/dist/esm/pdf-reader/flow-build.js

@@ -124,14 +124,37 @@ function shapeBlock(v) {
 		}
 	};
 }
-function buildFlowDoc(body, resources = new ResourceStore()) {
+function sectionFromPdfPages(pages) {
+	const box = pages[0]?.mediaBox;
+	if (!box) return void 0;
+	const width = Math.abs(box[2] - box[0]);
+	const height = Math.abs(box[3] - box[1]);
+	if (!(width > 0 && height > 0)) return void 0;
+	return {
+		pageSize: {
+			width: pt(width),
+			height: pt(height),
+			orientation: width > height ? "landscape" : "portrait"
+		},
+		margins: {
+			top: pt(0),
+			right: pt(0),
+			bottom: pt(0),
+			left: pt(0)
+		},
+		headers: [],
+		footers: []
+	};
+}
+function buildFlowDoc(body, resources = new ResourceStore(), section) {
 	return {
 		kind: "flow",
 		body: resolveBodyStyles([...body], EMPTY_STYLE_SHEET),
 		sections: [],
+		...section ? { section } : {},
 		styles: EMPTY_STYLE_SHEET,
 		resources
 	};
 }
 //#endregion
-export { buildFlowDoc, dedupeLosses, imageBlock, paragraphBlock, paragraphFromRuns, shapeBlock };
+export { buildFlowDoc, dedupeLosses, imageBlock, paragraphBlock, paragraphFromRuns, sectionFromPdfPages, shapeBlock };

package/dist/esm/pdf-reader/layout.js

@@ -1,5 +1,5 @@
 import { ResourceStore } from "../core/ir/resources.js";
-import { buildFlowDoc, dedupeLosses, imageBlock, paragraphFromRuns, shapeBlock } from "./flow-build.js";
+import { buildFlowDoc, dedupeLosses, imageBlock, paragraphFromRuns, sectionFromPdfPages, shapeBlock } from "./flow-build.js";
 import { collectPageImages } from "./images.js";
 import { extractPageText } from "./text.js";
 import { collectPageVectors } from "./vector.js";
@@ -45,7 +45,7 @@ function reconstructByLayout(file) {
 		for (const block of blocks) body.push(block.el);
 	});
 	return {
-		doc: buildFlowDoc(body, resources),
+		doc: buildFlowDoc(body, resources, sectionFromPdfPages(pages)),
 		losses: dedupeLosses(losses)
 	};
 }

package/dist/esm/pdf-reader/tagged.js

@@ -1,6 +1,6 @@
 import { pt } from "../core/ir/units.js";
 import { ResourceStore } from "../core/ir/resources.js";
-import { buildFlowDoc, dedupeLosses, imageBlock, paragraphBlock, paragraphFromRuns } from "./flow-build.js";
+import { buildFlowDoc, dedupeLosses, imageBlock, paragraphBlock, paragraphFromRuns, sectionFromPdfPages } from "./flow-build.js";
 import { collectPageImages } from "./images.js";
 import { extractPageText } from "./text.js";
 import { readStructTree } from "./struct-tree.js";
@@ -126,7 +126,7 @@ function reconstructTaggedPdf(file) {
 	for (const { img } of orphans) body.push(imageBlock(img, resources));
 	if (body.length === 0) return void 0;
 	return {
-		doc: buildFlowDoc(body, resources),
+		doc: buildFlowDoc(body, resources, sectionFromPdfPages(pages)),
 		losses: imageLosses
 	};
 }

package/dist/esm/pptx/pptx-reader.js

@@ -26,6 +26,10 @@ var parser = new XMLParser({
 	parseTagValue: false,
 	trimValues: false
 });
+function isSlideHidden(data) {
+	const root = decoder.decode(data.subarray(0, 4096)).match(/<p:sld\b[^>]*>/);
+	return root ? /\bshow\s*=\s*["']0["']/.test(root[0]) : false;
+}
 function readPptx(bytes) {
 	const losses = [];
 	const pkg = OpcPackage.open(bytes);
@@ -45,12 +49,23 @@ function readPptx(bytes) {
 		const slideRelById = new Map(pkg.getPartRelationships(presPath).filter((r) => r.type.endsWith("/slide")).map((r) => [r.id, r]));
 		const lst = kids.find((c) => poIs(c, "p:sldIdLst"));
 		const ids = lst ? poChildren(lst).filter((c) => poIs(c, "p:sldId")) : [];
+		let hidden = 0;
 		for (const sldId of ids) {
 			const rid = poAttr(sldId, "r:id");
 			const rel = rid !== void 0 ? slideRelById.get(rid) : void 0;
 			const part = rel ? pkg.resolveRelatedPart(presPath, rel) : void 0;
-			if (part) slideParts.push(part);
+			if (!part) continue;
+			if (isSlideHidden(part.data)) {
+				hidden++;
+				continue;
+			}
+			slideParts.push(part);
 		}
+		if (hidden > 0) losses.push({
+			severity: "dropped",
+			feature: FEATURES.text,
+			detail: `${hidden} hidden slide(s) omitted (p:sld@show="0")`
+		});
 	}
 	if (slideParts.length === 0) losses.push({
 		severity: "dropped",

package/dist/esm/word/doc/doc-reader.js

@@ -140,16 +140,22 @@ function readDoc(bytes) {
 		addStory(hf.firstFooter, "footer", "first");
 		addStory(hf.evenFooter, "footer", "even");
 	}
+	const ps = content.pageSize;
+	const pageSize = ps ? {
+		width: pt(ps.widthPt),
+		height: pt(ps.heightPt),
+		orientation: ps.widthPt > ps.heightPt ? "landscape" : "portrait"
+	} : {
+		width: pt(612),
+		height: pt(792)
+	};
 	return {
 		doc: {
 			kind: "flow",
 			body: resolveBodyStyles(body, EMPTY_STYLE_SHEET),
 			sections: [],
 			section: {
-				pageSize: {
-					width: pt(612),
-					height: pt(792)
-				},
+				pageSize,
 				margins: {
 					top: pt(72),
 					right: pt(72),

package/dist/esm/word/doc/doc-text.d.ts

@@ -74,6 +74,10 @@ export interface DocContent {
     readonly paragraphs: ReadonlyArray<DocParagraph>;
     readonly headerFooters?: DocHeaderFooters;
     readonly listTables?: DocListTables;
+    readonly pageSize?: {
+        readonly widthPt: number;
+        readonly heightPt: number;
+    };
     readonly encrypted: boolean;
 }
 export declare function extractDocContent(bytes: Uint8Array): DocContent;

package/dist/esm/word/doc/doc-text.js

@@ -13,6 +13,8 @@ var OFF_FCPLCFBTEPAPX = 258;
 var OFF_LCBPLCFBTEPAPX = 262;
 var OFF_FCCLX = 418;
 var OFF_LCBCLX = 422;
+var OFF_FCPLCFSED = 202;
+var OFF_LCBPLCFSED = 206;
 var OFF_FCPLFLST = 738;
 var OFF_LCBPLFLST = 742;
 var OFF_FCPLFLFO = 746;
@@ -43,6 +45,9 @@ var SPRM_T_DEF_TABLE_SHD = 54802;
 var MAX_ROW_CELLS = 256;
 var SPRM_P_ILVL = 9738;
 var SPRM_P_ILFO = 17931;
+var SPRM_S_XA_PAGE = 45087;
+var SPRM_S_YA_PAGE = 45088;
+var MAX_PAGE_TWIPS = 31680;
 var SPRA_LEN = [
 	1,
 	1,
@@ -63,6 +68,24 @@ var EMPTY = {
 	paragraphs: [],
 	encrypted: false
 };
+function readSectionPageSize(wd, table) {
+	const fc = u32(wd, OFF_FCPLCFSED);
+	const lcb = u32(wd, OFF_LCBPLCFSED);
+	if (lcb < 20 || fc + lcb > table.length) return void 0;
+	const fcSepx = u32(table.subarray(fc, fc + lcb), (Math.floor((lcb - 4) / 16) + 1) * 4 + 2);
+	if (fcSepx === 4294967295 || fcSepx + 2 > wd.length) return void 0;
+	const cb = u16(wd, fcSepx);
+	const grpprl = wd.subarray(fcSepx + 2, Math.min(wd.length, fcSepx + 2 + cb));
+	let xa = 0;
+	let ya = 0;
+	for (const { sprm, op } of sprms(grpprl)) if (sprm === SPRM_S_XA_PAGE) xa = u16(grpprl, op);
+	else if (sprm === SPRM_S_YA_PAGE) ya = u16(grpprl, op);
+	if (xa < 1 || ya < 1 || xa > MAX_PAGE_TWIPS || ya > MAX_PAGE_TWIPS) return void 0;
+	return {
+		widthPt: xa / 20,
+		heightPt: ya / 20
+	};
+}
 function extractDocContent(bytes) {
 	const cfb = openCfb(bytes);
 	const wd = cfb.readStream("WordDocument");
@@ -89,10 +112,12 @@ function extractDocContent(bytes) {
 	const main = buildParagraphs(wd, pieces, chpx, papx, 0, ccpText, data);
 	const headerFooters = extractHeaderFooters(wd, table, pieces, chpx, papx, data, ccpText);
 	const listTables = parseListTables(wd, table);
+	const pageSize = readSectionPageSize(wd, table);
 	return {
 		paragraphs: main,
 		...headerFooters ? { headerFooters } : {},
 		...listTables ? { listTables } : {},
+		...pageSize ? { pageSize } : {},
 		encrypted: false
 	};
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "reamkit",
-  "version": "1.15.0",
+  "version": "1.15.1",
   "description": "Ream — convert DOCX, XLSX, PPTX and PDF to PDF, SVG, HTML, DOCX and XLSX, built from scratch on the ECMA-376 and ISO 32000 specifications. Parse once, convert anywhere.",
   "license": "MIT",
   "author": "Alex Krassavin <info@reamkit.dev>",