pdf-codec 3.1.4 → 3.2.0

package/dist/read.js

@@ -1,17 +1,25 @@
+import "./notes-annotation-author.js";
+import { asArray, asName, asNumber, dictGet, pdfArray, pdfNull } from "./objects.js";
+import { decodePdfString, parsePdfDate } from "./pdf-text.js";
+import { concatBytes } from "./bytes/writer.js";
+import { serializeObjectToText } from "./serialize.js";
+import { applyMatrix, matrixRotationDegrees, matrixScaleX, matrixScaleY, multiplyMatrices, translationMatrix } from "./matrix.js";
+import { readPageAnnotations } from "./annotations.js";
+import { decodeStream } from "./filters.js";
 import { bytesToBase64 } from "./util/base64.js";
+import { readAttachments } from "./attachments.js";
 import { crc32 } from "./bytes/crc32.js";
-import { concatBytes } from "./bytes/writer.js";
+import { readAcroForm } from "./form.js";
+import { readXmpMetadata } from "./xmp.js";
+import { readOptionalContent } from "./optional-content.js";
 import "./layout.js";
 import { NOOP_DIAGNOSTIC_SINK, PdfParseError } from "./diagnostics.js";
-import { asArray, asName, asNumber, dictGet } from "./objects.js";
-import { decodeStream } from "./filters.js";
 import { openPdfDocument } from "./document.js";
 import { throwIfAborted } from "./util/abort.js";
 import { createFontResolver } from "./font-read.js";
 import { readImageXObject } from "./images-read.js";
-import { applyMatrix, matrixRotationDegrees, matrixScaleX, matrixScaleY, multiplyMatrices, translationMatrix } from "./matrix.js";
 import { interpretContentStream } from "./interpret.js";
-import "./notes-annotation-author.js";
+import { createDestinationRegistry, readOutline } from "./navigation.js";
 //#region src/read.ts
 const PDF_HEADER_BYTES = new TextEncoder().encode("%PDF-");
 const HEADER_SEARCH_WINDOW = 1024;
@@ -37,15 +45,28 @@ function readPdf(bytes, options) {
 	});
 	const images = {};
 	const imageIdCache = /* @__PURE__ */ new Map();
-	const pages = doc.pages().map((pageDict) => {
+	const pageDicts = doc.pages();
+	const destinationRegistry = createDestinationRegistry(doc.catalog, doc, (obj) => doc.pageIndex(obj), sink);
+	const outline = readOutline(doc.catalog, destinationRegistry, doc, sink);
+	const attachments = readAttachments(doc.catalog, pageDicts, doc, sink);
+	const optionalContent = readOptionalContent(doc.catalog, doc, sink);
+	const form = readAcroForm(doc.catalog, doc, (obj) => doc.pageIndex(obj), sink);
+	const source = readDocumentResidue(doc, sink);
+	const pages = pageDicts.map((pageDict) => {
 		throwIfAborted(signal);
-		return readPage(pageDict, doc, fontResolver, images, imageIdCache, sink);
+		return readPage(pageDict, doc, fontResolver, images, imageIdCache, destinationRegistry, optionalContent.layerNameOf, sink);
 	});
 	return {
 		formatVersion: 1,
-		metadata: readMetadata(doc.trailer, doc),
+		metadata: readMetadata(doc.trailer, doc, doc.catalog, sink),
 		pages,
-		images
+		images,
+		...destinationRegistry.entries.length > 0 ? { destinations: [...destinationRegistry.entries] } : {},
+		...outline.length > 0 ? { outline } : {},
+		...attachments.length > 0 ? { attachments } : {},
+		...optionalContent.layers.length > 0 ? { layers: [...optionalContent.layers] } : {},
+		...form.length > 0 ? { form } : {},
+		...Object.keys(source).length > 0 ? { source } : {}
 	};
 }
 function readMediaBox(page) {
@@ -135,7 +156,7 @@ function readPageContentBytes(page, resolver, sink) {
 	}
 	return /* @__PURE__ */ new Uint8Array(0);
 }
-function readPage(page, resolver, fontResolver, images, imageIdCache, sink) {
+function readPage(page, resolver, fontResolver, images, imageIdCache, destinationRegistry, layerNameOf, sink) {
 	const resources = resolver.resolveDict(dictGet(page, "Resources"));
 	const mediaBox = readMediaBox(page);
 	const rotationResult = pageRotationTransform(normalizeRotation(asNumber(dictGet(page, "Rotate"))), mediaBox.urx - mediaBox.llx, mediaBox.ury - mediaBox.lly);
@@ -146,7 +167,8 @@ function readPage(page, resolver, fontResolver, images, imageIdCache, sink) {
 		const extracted = interpretContentStream(contentBytes, resources, {
 			fontMetrics: fontResolver.metrics,
 			resolver,
-			sink
+			sink,
+			layerNameOf
 		});
 		for (const item of extracted) {
 			const converted = convertExtractedItem(item, pageMatrix, fontResolver, images, imageIdCache, resolver, sink);
@@ -157,13 +179,15 @@ function readPage(page, resolver, fontResolver, images, imageIdCache, sink) {
 		severity: "warning",
 		message: "page has no /Resources dict; its content stream cannot be interpreted"
 	});
-	items.push(...readLinkAnnotations(page, pageMatrix, resolver));
+	items.push(...readLinkAnnotations(page, pageMatrix, resolver, destinationRegistry));
 	const notes = readPageNotes(page, resolver);
+	const annotations = readPageAnnotations(page, pageMatrix, resolver, sink);
 	return {
 		widthPt: rotationResult.widthPt,
 		heightPt: rotationResult.heightPt,
 		items,
-		...notes !== void 0 ? { notes } : {}
+		...notes !== void 0 ? { notes } : {},
+		...annotations.length > 0 ? { annotations } : {}
 	};
 }
 function convertExtractedItem(item, pageMatrix, fontResolver, images, imageIdCache, resolver, sink) {
@@ -198,7 +222,10 @@ function convertText(item, pageMatrix, fontResolver) {
 		sizePt: sizePt > 0 ? sizePt : item.sizePt,
 		color: item.color,
 		widthPt,
-		rotationDeg: rotationDeg !== 0 ? rotationDeg : void 0
+		rotationDeg: rotationDeg !== 0 ? rotationDeg : void 0,
+		...item.layerName !== void 0 ? { layer: item.layerName } : {},
+		...item.actualText !== void 0 ? { actualText: item.actualText } : {},
+		...item.alt !== void 0 ? { alt: item.alt } : {}
 	};
 }
 function paintFields(paint) {
@@ -227,14 +254,16 @@ function convertRect(item, pageMatrix) {
 	return {
 		kind: "rect",
 		...transformBox(item, pageMatrix),
-		...paintFields(item)
+		...paintFields(item),
+		...item.layerName !== void 0 ? { layer: item.layerName } : {}
 	};
 }
 function convertEllipse(item, pageMatrix) {
 	return {
 		kind: "ellipse",
 		...transformBox(item, pageMatrix),
-		...paintFields(item)
+		...paintFields(item),
+		...item.layerName !== void 0 ? { layer: item.layerName } : {}
 	};
 }
 function convertLine(item, pageMatrix) {
@@ -253,7 +282,8 @@ function convertLine(item, pageMatrix) {
 		x2Pt: p2.x,
 		y2Pt: p2.y,
 		color: item.color,
-		widthPt: item.widthPt
+		widthPt: item.widthPt,
+		...item.layerName !== void 0 ? { layer: item.layerName } : {}
 	};
 }
 function transformSubpath(subpath, pageMatrix) {
@@ -308,7 +338,8 @@ function convertPath(item, pageMatrix) {
 		subpaths: item.subpaths.map((subpath) => transformSubpath(subpath, pageMatrix)),
 		...item.fill !== void 0 ? { fill: item.fill } : {},
 		...item.fill !== void 0 && item.fillRule === "evenodd" ? { fillRule: "evenodd" } : {},
-		...item.stroke !== void 0 ? { stroke: item.stroke } : {}
+		...item.stroke !== void 0 ? { stroke: item.stroke } : {},
+		...item.layerName !== void 0 ? { layer: item.layerName } : {}
 	};
 }
 function imagePlacementFrom(matrix) {
@@ -351,7 +382,8 @@ function convertImage(item, pageMatrix, images, cache, resolver, sink) {
 	return {
 		kind: "image",
 		imageId,
-		...imagePlacementFrom(multiplyMatrices(item.matrix, pageMatrix))
+		...imagePlacementFrom(multiplyMatrices(item.matrix, pageMatrix)),
+		...item.layerName !== void 0 ? { layer: item.layerName } : {}
 	};
 }
 function convertInlineImage(item, pageMatrix, images, resolver, sink) {
@@ -360,19 +392,22 @@ function convertInlineImage(item, pageMatrix, images, resolver, sink) {
 	return {
 		kind: "image",
 		imageId: registerExtractedImage(decoded.format, decoded.bytes, decoded.widthPx, decoded.heightPx, images),
-		...imagePlacementFrom(multiplyMatrices(item.matrix, pageMatrix))
+		...imagePlacementFrom(multiplyMatrices(item.matrix, pageMatrix)),
+		...item.layerName !== void 0 ? { layer: item.layerName } : {}
 	};
 }
-function readLinkAnnotations(page, pageMatrix, resolver) {
+function readLinkAnnotations(page, pageMatrix, resolver, destinationRegistry) {
 	const annotsArr = asArray(dictGet(page, "Annots"));
 	if (annotsArr === void 0) return [];
 	const links = [];
 	for (const annotRef of annotsArr) {
 		const annot = resolver.resolveDict(annotRef);
 		if (annot === void 0 || asName(dictGet(annot, "Subtype")) !== "Link") continue;
-		const uri = readLinkUri(annot, resolver);
 		const rectArr = asArray(dictGet(annot, "Rect"));
-		if (uri === void 0 || rectArr === void 0) continue;
+		if (rectArr === void 0) continue;
+		const uri = readLinkUri(annot, resolver);
+		const destination = uri === void 0 ? readInternalDestination(annot, resolver, destinationRegistry) : void 0;
+		if (uri === void 0 && destination === void 0) continue;
 		const x1 = asNumber(rectArr[0]) ?? 0;
 		const y1 = asNumber(rectArr[1]) ?? 0;
 		const x2 = asNumber(rectArr[2]) ?? 0;
@@ -385,13 +420,24 @@ function readLinkAnnotations(page, pageMatrix, resolver) {
 			x: Math.max(x1, x2),
 			y: Math.max(y1, y2)
 		});
-		links.push({
-			kind: "link",
-			uri,
+		const contentsObj = dictGet(annot, "Contents");
+		const title = contentsObj?.kind === "string" ? decodePdfString(contentsObj.bytes) : void 0;
+		const box = {
 			xPt: Math.min(p1.x, p2.x),
 			yPt: Math.min(p1.y, p2.y),
 			widthPt: Math.abs(p2.x - p1.x),
-			heightPt: Math.abs(p2.y - p1.y)
+			heightPt: Math.abs(p2.y - p1.y),
+			...title !== void 0 ? { title } : {}
+		};
+		if (uri !== void 0) links.push({
+			kind: "link",
+			uri,
+			...box
+		});
+		else if (destination !== void 0) links.push({
+			kind: "internalLink",
+			destination,
+			...box
 		});
 	}
 	return links;
@@ -402,6 +448,12 @@ function readLinkUri(annot, resolver) {
 	const uriObj = dictGet(action, "URI");
 	return uriObj?.kind === "string" ? decodePdfString(uriObj.bytes) : void 0;
 }
+function readInternalDestination(annot, resolver, destinationRegistry) {
+	const dest = dictGet(annot, "Dest");
+	if (dest !== void 0) return destinationRegistry.intern(dest);
+	const action = resolver.resolveDict(dictGet(annot, "A"));
+	if (action !== void 0 && asName(dictGet(action, "S")) === "GoTo") return destinationRegistry.intern(dictGet(action, "D"));
+}
 function readPageNotes(page, resolver) {
 	const annotsArr = asArray(dictGet(page, "Annots"));
 	if (annotsArr === void 0) return;
@@ -414,23 +466,7 @@ function readPageNotes(page, resolver) {
 		if (contentsObj?.kind === "string") return decodePdfString(contentsObj.bytes);
 	}
 }
-function decodePdfString(bytes) {
-	if (bytes.length >= 2 && bytes[0] === 254 && bytes[1] === 255) {
-		let out = "";
-		for (let i = 2; i + 1 < bytes.length; i += 2) out += String.fromCharCode((bytes[i] ?? 0) << 8 | (bytes[i + 1] ?? 0));
-		return out;
-	}
-	return Array.from(bytes, (b) => String.fromCharCode(b)).join("");
-}
-const PDF_DATE_PATTERN = /^D:(\d{4})(\d{2})?(\d{2})?(\d{2})?(\d{2})?(\d{2})?([+\-Z])?(\d{2})?'?(\d{2})?'?$/;
-function parsePdfDate(raw) {
-	if (raw === void 0) return;
-	const match = PDF_DATE_PATTERN.exec(raw);
-	if (match === null) return;
-	const [, year, month = "01", day = "01", hour = "00", minute = "00", second = "00", tzSign, tzHour = "00", tzMinute = "00"] = match;
-	return `${year}-${month}-${day}T${hour}:${minute}:${second}${tzSign === void 0 || tzSign === "Z" ? "Z" : `${tzSign}${tzHour}:${tzMinute}`}`;
-}
-function readMetadata(trailer, resolver) {
+function readMetadata(trailer, resolver, catalog, sink) {
 	const info = resolver.resolveDict(dictGet(trailer, "Info"));
 	if (info === void 0) return {};
 	const stringField = (key) => {
@@ -438,16 +474,52 @@ function readMetadata(trailer, resolver) {
 		return obj?.kind === "string" ? decodePdfString(obj.bytes) : void 0;
 	};
 	const keywords = stringField("Keywords")?.split(",").map((k) => k.trim()).filter((k) => k.length > 0);
+	const langObj = dictGet(catalog, "Lang");
+	const xmp = xmpPacket(catalog, resolver, sink);
+	const mirrored = readXmpMetadata(xmp ?? "");
 	return {
-		title: stringField("Title"),
-		author: stringField("Author"),
-		subject: stringField("Subject"),
-		keywords: keywords !== void 0 && keywords.length > 0 ? keywords : void 0,
-		creator: stringField("Creator"),
-		producer: stringField("Producer"),
-		createdIso: parsePdfDate(stringField("CreationDate")),
-		modifiedIso: parsePdfDate(stringField("ModDate"))
+		title: stringField("Title") ?? mirrored.title,
+		author: stringField("Author") ?? mirrored.author,
+		subject: stringField("Subject") ?? mirrored.subject,
+		keywords: (keywords !== void 0 && keywords.length > 0 ? keywords : void 0) ?? mirrored.keywords,
+		creator: stringField("Creator") ?? mirrored.creator,
+		producer: stringField("Producer") ?? mirrored.producer,
+		createdIso: parsePdfDate(stringField("CreationDate")) ?? mirrored.createdIso,
+		modifiedIso: parsePdfDate(stringField("ModDate")) ?? mirrored.modifiedIso,
+		...langObj?.kind === "string" ? { language: decodePdfString(langObj.bytes) } : {}
+	};
+}
+function xmpPacket(catalog, resolver, sink) {
+	const metadataObj = resolver.resolve(dictGet(catalog, "Metadata"));
+	if (metadataObj?.kind !== "stream") return;
+	const decoded = decodeStream(metadataObj.raw, metadataObj.dict, sink);
+	return new TextDecoder().decode(decoded.bytes);
+}
+function readDocumentResidue(doc, sink) {
+	const source = {};
+	const residue = (key, obj) => {
+		if (obj !== void 0) source[key] = {
+			format: "pdf",
+			xml: serializeObjectToText(obj)
+		};
+	};
+	const catalog = doc.catalog;
+	const packet = xmpPacket(catalog, doc, sink);
+	if (packet !== void 0) source.xmp = {
+		format: "pdf",
+		xml: packet
 	};
+	residue("viewer-preferences", dictGet(catalog, "ViewerPreferences"));
+	residue("page-mode", dictGet(catalog, "PageMode"));
+	residue("page-layout", dictGet(catalog, "PageLayout"));
+	residue("open-action", dictGet(catalog, "OpenAction"));
+	const outputIntents = asArray(dictGet(catalog, "OutputIntents"));
+	if (outputIntents !== void 0) residue("output-intents", pdfArray(outputIntents.map((intent) => doc.resolve(intent) ?? pdfNull())));
+	residue("piece-info", dictGet(catalog, "PieceInfo"));
+	residue("legal", dictGet(catalog, "Legal"));
+	residue("collection", dictGet(catalog, "Collection"));
+	residue("trailer-id", dictGet(doc.trailer, "ID"));
+	return source;
 }
 //#endregion
-export { decodePdfString, normalizeRotation, pageRotationTransform, parsePdfDate, readPdf };
+export { normalizeRotation, pageRotationTransform, readPdf };

package/dist/serialize.cjs

@@ -1,6 +1,6 @@
 Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
-const require_bytes_writer = require("./bytes/writer.cjs");
 const require_objects = require("./objects.cjs");
+const require_bytes_writer = require("./bytes/writer.cjs");
 //#region src/serialize.ts
 const NUMBER_DECIMAL_PLACES = 4;
 const NUMBER_EPSILON = 10 ** -4;
@@ -63,7 +63,11 @@ function serializeObject(obj) {
 	writeObject(writer, obj);
 	return writer.toBytes();
 }
+function serializeObjectToText(obj) {
+	return new TextDecoder().decode(serializeObject(obj));
+}
 //#endregion
 exports.formatNumber = formatNumber;
 exports.serializeObject = serializeObject;
+exports.serializeObjectToText = serializeObjectToText;
 exports.writeObject = writeObject;

package/dist/serialize.d.cts

@@ -1,8 +1,9 @@
-import { ByteWriter } from "./bytes/writer.cjs";
 import { PdfObject } from "./objects.cjs";
+import { ByteWriter } from "./bytes/writer.cjs";
 //#region src/serialize.d.ts
 declare function formatNumber(n: number): string;
 declare function writeObject(writer: ByteWriter, obj: PdfObject): void;
 declare function serializeObject(obj: PdfObject): Uint8Array<ArrayBuffer>;
+declare function serializeObjectToText(obj: PdfObject): string;
 //#endregion
-export { formatNumber, serializeObject, writeObject };
+export { formatNumber, serializeObject, serializeObjectToText, writeObject };

package/dist/serialize.d.ts

@@ -1,8 +1,9 @@
-import { ByteWriter } from "./bytes/writer.js";
 import { PdfObject } from "./objects.js";
+import { ByteWriter } from "./bytes/writer.js";
 //#region src/serialize.d.ts
 declare function formatNumber(n: number): string;
 declare function writeObject(writer: ByteWriter, obj: PdfObject): void;
 declare function serializeObject(obj: PdfObject): Uint8Array<ArrayBuffer>;
+declare function serializeObjectToText(obj: PdfObject): string;
 //#endregion
-export { formatNumber, serializeObject, writeObject };
+export { formatNumber, serializeObject, serializeObjectToText, writeObject };

package/dist/serialize.js

@@ -1,5 +1,5 @@
-import { ByteWriter } from "./bytes/writer.js";
 import { pdfNum } from "./objects.js";
+import { ByteWriter } from "./bytes/writer.js";
 //#region src/serialize.ts
 const NUMBER_DECIMAL_PLACES = 4;
 const NUMBER_EPSILON = 10 ** -4;
@@ -62,5 +62,8 @@ function serializeObject(obj) {
 	writeObject(writer, obj);
 	return writer.toBytes();
 }
+function serializeObjectToText(obj) {
+	return new TextDecoder().decode(serializeObject(obj));
+}
 //#endregion
-export { formatNumber, serializeObject, writeObject };
+export { formatNumber, serializeObject, serializeObjectToText, writeObject };

package/dist/write.cjs

@@ -1,15 +1,15 @@
 Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
 const require_afm_widths = require("./afm-widths-BoeTOK2r.cjs");
-const require_util_base64 = require("./util/base64.cjs");
-const require_bytes_writer = require("./bytes/writer.cjs");
+const require_notes_annotation_author = require("./notes-annotation-author.cjs");
 const require_objects = require("./objects.cjs");
+const require_bytes_writer = require("./bytes/writer.cjs");
+const require_serialize = require("./serialize.cjs");
 const require_bytes_flate = require("./bytes/flate.cjs");
+const require_util_base64 = require("./util/base64.cjs");
 const require_util_abort = require("./util/abort.cjs");
 const require_image_jpeg_info = require("./image/jpeg-info.cjs");
-const require_notes_annotation_author = require("./notes-annotation-author.cjs");
 const require_image_png_decode = require("./image/png-decode.cjs");
 const require_embedded_font = require("./embedded-font.cjs");
-const require_serialize = require("./serialize.cjs");
 const require_content_write = require("./content-write.cjs");
 const require_embedded_font_write = require("./embedded-font-write.cjs");
 const require_font_registry = require("./font-registry.cjs");
@@ -178,9 +178,55 @@ function buildLinkAnnotDict(link) {
 		})
 	});
 }
+function destinationViewArray(target) {
+	const n = (value) => value === void 0 ? require_objects.pdfNull() : require_objects.pdfNum(value);
+	if (target.kind === "xyz") return [
+		require_objects.pdfName("XYZ"),
+		n(target.leftPt),
+		n(target.topPt),
+		n(target.zoom)
+	];
+	if (target.kind === "fitH") return [require_objects.pdfName("FitH"), n(target.topPt)];
+	if (target.kind === "fitV") return [require_objects.pdfName("FitV"), n(target.leftPt)];
+	if (target.kind === "fitR") return [
+		require_objects.pdfName("FitR"),
+		n(target.leftPt),
+		n(target.bottomPt),
+		n(target.rightPt),
+		n(target.topPt)
+	];
+	if (target.kind === "fitBH") return [require_objects.pdfName("FitBH"), n(target.topPt)];
+	if (target.kind === "fitBV") return [require_objects.pdfName("FitBV"), n(target.leftPt)];
+	return [require_objects.pdfName(target.kind === "fitB" ? "FitB" : "Fit")];
+}
+function buildInternalLinkAnnotDict(link, doc, pageAllocs) {
+	const destination = doc.destinations?.find((d) => d.name === link.destination);
+	if (destination === void 0) throw new Error(`internal link names destination "${link.destination}", which the document's destinations table does not carry -- this is a caller-invariant violation`);
+	const targetPage = pageAllocs[destination.pageIndex];
+	if (targetPage === void 0) throw new Error(`destination "${link.destination}" names page index ${destination.pageIndex}, which is beyond the document's own pages -- this is a caller-invariant violation`);
+	return require_objects.pdfDict({
+		Type: require_objects.pdfName("Annot"),
+		Subtype: require_objects.pdfName("Link"),
+		Rect: require_objects.pdfArray([
+			link.xPt,
+			link.yPt,
+			link.xPt + link.widthPt,
+			link.yPt + link.heightPt
+		].map((v) => require_objects.pdfNum(v))),
+		Border: require_objects.pdfArray([
+			0,
+			0,
+			0
+		].map((v) => require_objects.pdfNum(v))),
+		Dest: require_objects.pdfArray([require_objects.pdfRef(targetPage.pageNum, 0), ...destinationViewArray(destination.target)])
+	});
+}
 function isLinkItem(item) {
 	return item.kind === "link";
 }
+function isInternalLinkItem(item) {
+	return item.kind === "internalLink";
+}
 const NOTES_ANNOTATION_HIDDEN_FLAG = 2;
 function buildNotesAnnotDict(notes) {
 	return require_objects.pdfDict({
@@ -436,7 +482,7 @@ function writePdf(doc, options = {}) {
 			num: contentsNum,
 			value: require_objects.pdfStream(contentsDict, finalContentBytes)
 		});
-		const annots = page.items.filter(isLinkItem).map((link) => buildLinkAnnotDict(link));
+		const annots = [...page.items.filter(isLinkItem).map((link) => buildLinkAnnotDict(link)), ...page.items.filter(isInternalLinkItem).map((link) => buildInternalLinkAnnotDict(link, doc, pageAllocs))];
 		if (page.notes !== void 0 && page.notes.length > 0) annots.push(buildNotesAnnotDict(page.notes));
 		const pageEntries = /* @__PURE__ */ new Map([
 			["Type", require_objects.pdfName("Page")],

package/dist/write.js

@@ -1,14 +1,14 @@
 import { a as winAnsiGlyphName, n as widthOfCode, t as STANDARD_METRICS } from "./afm-widths-Dxucrg7D.js";
-import { base64ToBytes } from "./util/base64.js";
+import { NOTES_ANNOTATION_AUTHOR } from "./notes-annotation-author.js";
+import { pdfArray, pdfDict, pdfHexString, pdfName, pdfNull, pdfNum, pdfRef, pdfStream } from "./objects.js";
 import { ByteWriter, concatBytes } from "./bytes/writer.js";
-import { pdfArray, pdfDict, pdfHexString, pdfName, pdfNum, pdfRef, pdfStream } from "./objects.js";
+import { writeObject } from "./serialize.js";
 import { deflate } from "./bytes/flate.js";
+import { base64ToBytes } from "./util/base64.js";
 import { throwIfAborted } from "./util/abort.js";
 import { readJpegInfo } from "./image/jpeg-info.js";
-import { NOTES_ANNOTATION_AUTHOR } from "./notes-annotation-author.js";
 import { decodePng } from "./image/png-decode.js";
 import { collectEmbeddedGlyphs } from "./embedded-font.js";
-import { writeObject } from "./serialize.js";
 import { writeContentStream } from "./content-write.js";
 import { buildEmbeddedFontObjects } from "./embedded-font-write.js";
 import { resolveFaceWithRegistry } from "./font-registry.js";
@@ -177,9 +177,55 @@ function buildLinkAnnotDict(link) {
 		})
 	});
 }
+function destinationViewArray(target) {
+	const n = (value) => value === void 0 ? pdfNull() : pdfNum(value);
+	if (target.kind === "xyz") return [
+		pdfName("XYZ"),
+		n(target.leftPt),
+		n(target.topPt),
+		n(target.zoom)
+	];
+	if (target.kind === "fitH") return [pdfName("FitH"), n(target.topPt)];
+	if (target.kind === "fitV") return [pdfName("FitV"), n(target.leftPt)];
+	if (target.kind === "fitR") return [
+		pdfName("FitR"),
+		n(target.leftPt),
+		n(target.bottomPt),
+		n(target.rightPt),
+		n(target.topPt)
+	];
+	if (target.kind === "fitBH") return [pdfName("FitBH"), n(target.topPt)];
+	if (target.kind === "fitBV") return [pdfName("FitBV"), n(target.leftPt)];
+	return [pdfName(target.kind === "fitB" ? "FitB" : "Fit")];
+}
+function buildInternalLinkAnnotDict(link, doc, pageAllocs) {
+	const destination = doc.destinations?.find((d) => d.name === link.destination);
+	if (destination === void 0) throw new Error(`internal link names destination "${link.destination}", which the document's destinations table does not carry -- this is a caller-invariant violation`);
+	const targetPage = pageAllocs[destination.pageIndex];
+	if (targetPage === void 0) throw new Error(`destination "${link.destination}" names page index ${destination.pageIndex}, which is beyond the document's own pages -- this is a caller-invariant violation`);
+	return pdfDict({
+		Type: pdfName("Annot"),
+		Subtype: pdfName("Link"),
+		Rect: pdfArray([
+			link.xPt,
+			link.yPt,
+			link.xPt + link.widthPt,
+			link.yPt + link.heightPt
+		].map((v) => pdfNum(v))),
+		Border: pdfArray([
+			0,
+			0,
+			0
+		].map((v) => pdfNum(v))),
+		Dest: pdfArray([pdfRef(targetPage.pageNum, 0), ...destinationViewArray(destination.target)])
+	});
+}
 function isLinkItem(item) {
 	return item.kind === "link";
 }
+function isInternalLinkItem(item) {
+	return item.kind === "internalLink";
+}
 const NOTES_ANNOTATION_HIDDEN_FLAG = 2;
 function buildNotesAnnotDict(notes) {
 	return pdfDict({
@@ -435,7 +481,7 @@ function writePdf(doc, options = {}) {
 			num: contentsNum,
 			value: pdfStream(contentsDict, finalContentBytes)
 		});
-		const annots = page.items.filter(isLinkItem).map((link) => buildLinkAnnotDict(link));
+		const annots = [...page.items.filter(isLinkItem).map((link) => buildLinkAnnotDict(link)), ...page.items.filter(isInternalLinkItem).map((link) => buildInternalLinkAnnotDict(link, doc, pageAllocs))];
 		if (page.notes !== void 0 && page.notes.length > 0) annots.push(buildNotesAnnotDict(page.notes));
 		const pageEntries = /* @__PURE__ */ new Map([
 			["Type", pdfName("Page")],

package/dist/xmp.cjs

@@ -0,0 +1,46 @@
+Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
+//#region src/xmp.ts
+function readXmpMetadata(packet) {
+	return {
+		...field(xmpValue(packet, "dc:title"), "title"),
+		...field(xmpValue(packet, "dc:creator"), "author"),
+		...field(xmpValue(packet, "dc:description"), "subject"),
+		...field(xmpValue(packet, "xmp:CreatorTool"), "creator"),
+		...field(xmpValue(packet, "pdf:Producer"), "producer"),
+		...field(xmpValue(packet, "xmp:CreateDate"), "createdIso"),
+		...field(xmpValue(packet, "xmp:ModifyDate"), "modifiedIso"),
+		...keywordsOf(packet)
+	};
+}
+function xmpValue(packet, element) {
+	const match = new RegExp(`<${element}(?:\\s[^>]*)?>([\\s\\S]*?)</${element}>`).exec(packet);
+	if (match === null) return;
+	const inner = match[1] ?? "";
+	const items = listItems(inner);
+	if (items.length > 0) return items;
+	const text = inner.trim();
+	return text.length > 0 ? text : void 0;
+}
+function listItems(inner) {
+	const items = [];
+	const pattern = /<rdf:li(?:\s[^>]*)?>([\s\S]*?)<\/rdf:li>/g;
+	let match;
+	while ((match = pattern.exec(inner)) !== null) {
+		const text = (match[1] ?? "").trim();
+		if (text.length > 0) items.push(text);
+	}
+	return items;
+}
+function field(value, key) {
+	if (typeof value === "string") return { [key]: value };
+	if (Array.isArray(value) && value.length > 0) return { [key]: value.join(", ") };
+	return {};
+}
+function keywordsOf(packet) {
+	const match = (/* @__PURE__ */ new RegExp("<dc:subject(?:\\s[^>]*)?>([\\s\\S]*?)</dc:subject>")).exec(packet);
+	if (match === null) return {};
+	const items = listItems(match[1] ?? "");
+	return items.length > 0 ? { keywords: items } : {};
+}
+//#endregion
+exports.readXmpMetadata = readXmpMetadata;

package/dist/xmp.d.cts

@@ -0,0 +1,14 @@
+//#region src/xmp.d.ts
+interface XmpMetadata {
+  readonly title?: string;
+  readonly author?: string;
+  readonly subject?: string;
+  readonly keywords?: string[];
+  readonly creator?: string;
+  readonly producer?: string;
+  readonly createdIso?: string;
+  readonly modifiedIso?: string;
+}
+declare function readXmpMetadata(packet: string): XmpMetadata;
+//#endregion
+export { XmpMetadata, readXmpMetadata };

package/dist/xmp.d.ts

@@ -0,0 +1,14 @@
+//#region src/xmp.d.ts
+interface XmpMetadata {
+  readonly title?: string;
+  readonly author?: string;
+  readonly subject?: string;
+  readonly keywords?: string[];
+  readonly creator?: string;
+  readonly producer?: string;
+  readonly createdIso?: string;
+  readonly modifiedIso?: string;
+}
+declare function readXmpMetadata(packet: string): XmpMetadata;
+//#endregion
+export { XmpMetadata, readXmpMetadata };

package/dist/xmp.js

@@ -0,0 +1,45 @@
+//#region src/xmp.ts
+function readXmpMetadata(packet) {
+	return {
+		...field(xmpValue(packet, "dc:title"), "title"),
+		...field(xmpValue(packet, "dc:creator"), "author"),
+		...field(xmpValue(packet, "dc:description"), "subject"),
+		...field(xmpValue(packet, "xmp:CreatorTool"), "creator"),
+		...field(xmpValue(packet, "pdf:Producer"), "producer"),
+		...field(xmpValue(packet, "xmp:CreateDate"), "createdIso"),
+		...field(xmpValue(packet, "xmp:ModifyDate"), "modifiedIso"),
+		...keywordsOf(packet)
+	};
+}
+function xmpValue(packet, element) {
+	const match = new RegExp(`<${element}(?:\\s[^>]*)?>([\\s\\S]*?)</${element}>`).exec(packet);
+	if (match === null) return;
+	const inner = match[1] ?? "";
+	const items = listItems(inner);
+	if (items.length > 0) return items;
+	const text = inner.trim();
+	return text.length > 0 ? text : void 0;
+}
+function listItems(inner) {
+	const items = [];
+	const pattern = /<rdf:li(?:\s[^>]*)?>([\s\S]*?)<\/rdf:li>/g;
+	let match;
+	while ((match = pattern.exec(inner)) !== null) {
+		const text = (match[1] ?? "").trim();
+		if (text.length > 0) items.push(text);
+	}
+	return items;
+}
+function field(value, key) {
+	if (typeof value === "string") return { [key]: value };
+	if (Array.isArray(value) && value.length > 0) return { [key]: value.join(", ") };
+	return {};
+}
+function keywordsOf(packet) {
+	const match = (/* @__PURE__ */ new RegExp("<dc:subject(?:\\s[^>]*)?>([\\s\\S]*?)</dc:subject>")).exec(packet);
+	if (match === null) return {};
+	const items = listItems(match[1] ?? "");
+	return items.length > 0 ? { keywords: items } : {};
+}
+//#endregion
+export { readXmpMetadata };