reamkit 1.25.0 → 1.26.0

package/README.md

@@ -191,10 +191,12 @@ preview, flowed HTML export, and **docx + xlsx output** (write WordprocessingML
 tagged PDF from its structure tree (headings, tables, lists, reading order), an
 untagged one heuristically from glyph positions (lines, paragraphs, headings,
 and a clean two-column split). It lifts back the text (via each font's
-`/ToUnicode`, or the embedded program's own `cmap` where there is none), the
+`/ToUnicode`, or the embedded program's own `cmap` where there is none, or the
+glyph names its `/Encoding` states — which is all a PDF from TeX gives), the
 font programs themselves, raster images (JPEG verbatim; PNG/Flate/LZW/CCITT-fax
-decoded and re-encoded), `/Link` hyperlinks, form-XObject content, annotation
-appearances, and the page's artwork: filled / stroked / gradient shapes,
+and **JBIG2** decoded and re-encoded), `/Link` hyperlinks, form-XObject content,
+annotation appearances, colour set through a named space, and the page's
+artwork: filled / stroked / gradient shapes,
 clipping paths, tiling patterns, constant alpha, and the Type 3 glyphs that are
 drawings rather than letters. It reads modern compressed files (cross-reference
 + object streams) and encrypted ones (RC4 / AES — the user password is passed to

package/dist/esm/pdf-reader/content.d.ts

@@ -1,3 +1,4 @@
+import { ColorSpaceInfo } from './shading.js';
 import { ShapeGradient } from '../core/vector.js';
 import { PdfDict, PdfStream } from '../pdf/objects.js';
 /**
@@ -221,22 +222,7 @@ export type Matrix = readonly [number, number, number, number, number, number];
 export declare const IDENTITY: Matrix;
 /** Compose two {@link Matrix matrices}: `a` applied first, then `b`. */
 export declare function multiply(a: Matrix, b: Matrix): Matrix;
-/**
- * Walk a page's decoded content stream (§9.4) and extract its positioned text,
- * painted XObjects and painted paths. Tracks the graphics state (CTM via
- * `q`/`Q`/`cm`) and text state (text/line matrices, font, size, spacing), and
- * emits one {@link TextRun} per show operator (`Tj` / `TJ` / `'` / `"`) in page
- * space (points).
- *
- * @param bytes      The page's decoded content-stream bytes.
- * @param fonts      Page fonts by resource key (`Tf` name), for decoding + widths;
- *                   an unmapped key falls back to Latin-1 with a half-em advance.
- * @param initialCtm The starting CTM mapping user space to page space.
- * @param shadings   Shading patterns by name, selected by `scn`/`sc` (EP16c).
- * @param alphas     Constant fill alphas by `/ExtGState` name, selected by `gs`.
- * @returns The extracted text runs, image placements and vector paths.
- */
-export declare function interpretContent(bytes: Uint8Array, fonts: ReadonlyMap<string, ContentFont>, initialCtm?: Matrix, shadings?: ReadonlyMap<string, ShapeGradient>, alphas?: ReadonlyMap<string, number>): InterpretResult;
+export declare function interpretContent(bytes: Uint8Array, fonts: ReadonlyMap<string, ContentFont>, initialCtm?: Matrix, shadings?: ReadonlyMap<string, ShapeGradient>, alphas?: ReadonlyMap<string, number>, spaces?: ReadonlyMap<string, ColorSpaceInfo>): InterpretResult;
 /**
  * Whether a string is wholly right-to-left: at least one letter of an RTL
  * script and nothing of any other, spaces and joiners aside.

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

@@ -73,6 +73,8 @@ var FALLBACK_FONT = {
 var UPRIGHT_TOLERANCE_DEG = .5;
 function initialState() {
 	return {
+		fillSpace: void 0,
+		strokeSpace: void 0,
 		ctm: IDENTITY,
 		fontKey: "",
 		font: FALLBACK_FONT,
@@ -107,7 +109,49 @@ function initialState() {
 * @param alphas     Constant fill alphas by `/ExtGState` name, selected by `gs`.
 * @returns The extracted text runs, image placements and vector paths.
 */
-function interpretContent(bytes, fonts, initialCtm = IDENTITY, shadings = /* @__PURE__ */ new Map(), alphas = /* @__PURE__ */ new Map()) {
+/** §8.6.8 — the space a `cs` / `CS` operand names, if it is one we read. */
+function spaceOf(operands, spaces) {
+	const name = operands[operands.length - 1];
+	if (!(name instanceof PdfName)) return void 0;
+	const direct = spaces.get(name.value);
+	if (direct) return direct;
+	switch (name.value) {
+		case "DeviceGray": return {
+			kind: "gray",
+			components: 1
+		};
+		case "DeviceRGB": return {
+			kind: "rgb",
+			components: 3
+		};
+		case "DeviceCMYK": return {
+			kind: "cmyk",
+			components: 4
+		};
+		default: return;
+	}
+}
+/**
+* §8.6.8 — the colour a run of `sc` / `scn` components comes to.
+*
+* The space in force decides, and where it was not read the COUNT is the next
+* best witness: three numbers are RGB and four are CMYK on every device space
+* there is. One number is the ambiguous case — grey in a device space, but the
+* strength of a colorant in a Separation, where 1 is the ink at full and reads
+* dark — so a lone component is only taken where the space said what it means.
+*/
+function componentColor(operands, space) {
+	const nums = operands.filter((o) => typeof o === "number");
+	if (nums.length === 0) return void 0;
+	switch (space?.kind ?? (nums.length === 3 ? "rgb" : nums.length === 4 ? "cmyk" : void 0)) {
+		case "rgb": return nums.length >= 3 ? rgbHex(nums[0], nums[1], nums[2]) : void 0;
+		case "cmyk": return nums.length >= 4 ? cmykHex(nums[0], nums[1], nums[2], nums[3]) : void 0;
+		case "gray": return grayHex(nums[0]);
+		case "tint": return grayHex(1 - Math.max(...nums));
+		default: return;
+	}
+}
+function interpretContent(bytes, fonts, initialCtm = IDENTITY, shadings = /* @__PURE__ */ new Map(), alphas = /* @__PURE__ */ new Map(), spaces = /* @__PURE__ */ new Map()) {
 	const runs = [];
 	const images = [];
 	const vectors = [];
@@ -384,12 +428,34 @@ function interpretContent(bytes, fonts, initialCtm = IDENTITY, shadings = /* @__
 				state.fillGradient = void 0;
 				state.fillPattern = void 0;
 				break;
+			case "cs":
+				state.fillSpace = spaceOf(operands, spaces);
+				break;
+			case "CS":
+				state.strokeSpace = spaceOf(operands, spaces);
+				break;
 			case "scn":
 			case "sc": {
 				const last = operands[operands.length - 1];
 				const named = last instanceof PdfName ? last.value : void 0;
-				state.fillGradient = named !== void 0 ? shadings.get(named) : void 0;
-				state.fillPattern = named !== void 0 && !state.fillGradient ? named : void 0;
+				if (named !== void 0) {
+					state.fillGradient = shadings.get(named);
+					state.fillPattern = state.fillGradient ? void 0 : named;
+					break;
+				}
+				const hex = componentColor(operands, state.fillSpace);
+				if (hex !== void 0) {
+					state.fillColor = hex;
+					state.fillGradient = void 0;
+					state.fillPattern = void 0;
+				}
+				break;
+			}
+			case "SCN":
+			case "SC": {
+				if (operands[operands.length - 1] instanceof PdfName) break;
+				const hex = componentColor(operands, state.strokeSpace);
+				if (hex !== void 0) state.strokeColor = hex;
 				break;
 			}
 			case "RG":

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

@@ -2,6 +2,7 @@ import { parseTtf } from "../core/font/ttf-parser.js";
 import { PDF_NULL, PdfName, PdfStream } from "../pdf/objects.js";
 import { embeddedFontName } from "./embedded-fonts.js";
 import { parseToUnicodeCMap } from "./cmap.js";
+import { textForGlyphName } from "./glyph-names.js";
 //#region src/pdf-reader/font.ts
 /**
 * Build a {@link ContentFont} (the interpreter's decode + advance hooks) from a
@@ -25,7 +26,9 @@ function buildContentFont(file, fontDict) {
 		toUnicode = parsed.map;
 		if (isType0) codeBytes = parsed.codeBytes;
 	}
-	const unicode = (isType0 && toUnicode.size === 0 ? embeddedCmap(file, fontDict) : void 0) ?? toUnicode;
+	const fromProgram = isType0 && toUnicode.size === 0 ? embeddedCmap(file, fontDict) : void 0;
+	const fromNames = !isType0 ? namedGlyphs(file, fontDict) : void 0;
+	const unicode = fromProgram ?? (toUnicode.size > 0 ? toUnicode : fromNames ?? toUnicode);
 	const bytesPerCode = codeBytes;
 	const style = faceStyle(file, fontDict, isType0);
 	const name = embeddedFontName(file, fontDict);
@@ -148,6 +151,23 @@ function fontMatrix(file, fontDict) {
 		n[5]
 	];
 }
+/**
+* §9.6.6.1 — code → text for a simple font, out of the glyph names its
+* `/Encoding /Differences` gives.
+*
+* Returns `undefined` where the font names nothing, so the caller keeps its
+* Latin-1 reading rather than replacing it with an empty map.
+*/
+function namedGlyphs(file, fontDict) {
+	const names = differences(file, fontDict);
+	if (names.size === 0) return void 0;
+	const out = /* @__PURE__ */ new Map();
+	for (const [code, name] of names) {
+		const text = textForGlyphName(name);
+		if (text !== void 0) out.set(code, text);
+	}
+	return out.size > 0 ? out : void 0;
+}
 /** §9.6.6.1 `/Encoding` `/Differences` — code → glyph name, as the array runs. */
 function differences(file, fontDict) {
 	const out = /* @__PURE__ */ new Map();

package/dist/esm/pdf-reader/glyph-names.d.ts

@@ -0,0 +1,6 @@
+/**
+ * The text a glyph name stands for, or `undefined` when the name says nothing.
+ *
+ * @param name The name as `/Differences` (or a font program's charset) gives it.
+ */
+export declare function textForGlyphName(name: string): string | undefined;

package/dist/esm/pdf-reader/glyph-names.js

@@ -0,0 +1,408 @@
+//#region src/pdf-reader/glyph-names.ts
+/** Glyph name → the text it stands for. */
+var GLYPHS = {
+	AE: "Æ",
+	Aacute: "Á",
+	Abreve: "Ă",
+	Acaron: "Ǎ",
+	Acircumflex: "Â",
+	Adieresis: "Ä",
+	Adotaccent: "Ȧ",
+	Agrave: "À",
+	Amacron: "Ā",
+	Aogonek: "Ą",
+	Aring: "Å",
+	Atilde: "Ã",
+	Bdotaccent: "Ḃ",
+	Cacute: "Ć",
+	Ccaron: "Č",
+	Ccedilla: "Ç",
+	Ccircumflex: "Ĉ",
+	Cdotaccent: "Ċ",
+	Dcaron: "Ď",
+	Dcedilla: "Ḑ",
+	Ddotaccent: "Ḋ",
+	Eacute: "É",
+	Ebreve: "Ĕ",
+	Ecaron: "Ě",
+	Ecedilla: "Ȩ",
+	Ecircumflex: "Ê",
+	Edieresis: "Ë",
+	Edotaccent: "Ė",
+	Egrave: "È",
+	Emacron: "Ē",
+	Eogonek: "Ę",
+	Eth: "Ð",
+	Etilde: "Ẽ",
+	Euro: "€",
+	Fdotaccent: "Ḟ",
+	Gacute: "Ǵ",
+	Gbreve: "Ğ",
+	Gcaron: "Ǧ",
+	Gcedilla: "Ģ",
+	Gcircumflex: "Ĝ",
+	Gdotaccent: "Ġ",
+	Gmacron: "Ḡ",
+	Hcaron: "Ȟ",
+	Hcedilla: "Ḩ",
+	Hcircumflex: "Ĥ",
+	Hdieresis: "Ḧ",
+	Hdotaccent: "Ḣ",
+	Iacute: "Í",
+	Ibreve: "Ĭ",
+	Icaron: "Ǐ",
+	Icircumflex: "Î",
+	Idieresis: "Ï",
+	Idotaccent: "İ",
+	Igrave: "Ì",
+	Imacron: "Ī",
+	Iogonek: "Į",
+	Itilde: "Ĩ",
+	Jcircumflex: "Ĵ",
+	Kacute: "Ḱ",
+	Kcaron: "Ǩ",
+	Kcedilla: "Ķ",
+	Lacute: "Ĺ",
+	Lcaron: "Ľ",
+	Lcedilla: "Ļ",
+	Macute: "Ḿ",
+	Mdotaccent: "Ṁ",
+	Nacute: "Ń",
+	Ncaron: "Ň",
+	Ncedilla: "Ņ",
+	Ndotaccent: "Ṅ",
+	Ngrave: "Ǹ",
+	Ntilde: "Ñ",
+	OE: "Œ",
+	Oacute: "Ó",
+	Obreve: "Ŏ",
+	Ocaron: "Ǒ",
+	Ocircumflex: "Ô",
+	Odieresis: "Ö",
+	Odotaccent: "Ȯ",
+	Ograve: "Ò",
+	Ohungarumlaut: "Ő",
+	Omacron: "Ō",
+	Oogonek: "Ǫ",
+	Oslash: "Ø",
+	Otilde: "Õ",
+	Pacute: "Ṕ",
+	Pdotaccent: "Ṗ",
+	Racute: "Ŕ",
+	Rcaron: "Ř",
+	Rcedilla: "Ŗ",
+	Rdotaccent: "Ṙ",
+	Sacute: "Ś",
+	Scaron: "Š",
+	Scedilla: "Ş",
+	Scircumflex: "Ŝ",
+	Sdotaccent: "Ṡ",
+	Tcaron: "Ť",
+	Tcedilla: "Ţ",
+	Tdotaccent: "Ṫ",
+	Thorn: "Þ",
+	Uacute: "Ú",
+	Ubreve: "Ŭ",
+	Ucaron: "Ǔ",
+	Ucircumflex: "Û",
+	Udieresis: "Ü",
+	Ugrave: "Ù",
+	Uhungarumlaut: "Ű",
+	Umacron: "Ū",
+	Uogonek: "Ų",
+	Uring: "Ů",
+	Utilde: "Ũ",
+	Vtilde: "Ṽ",
+	Wacute: "Ẃ",
+	Wcircumflex: "Ŵ",
+	Wdieresis: "Ẅ",
+	Wdotaccent: "Ẇ",
+	Wgrave: "Ẁ",
+	Xdieresis: "Ẍ",
+	Xdotaccent: "Ẋ",
+	Yacute: "Ý",
+	Ycircumflex: "Ŷ",
+	Ydieresis: "Ÿ",
+	Ydotaccent: "Ẏ",
+	Ygrave: "Ỳ",
+	Ymacron: "Ȳ",
+	Ytilde: "Ỹ",
+	Zacute: "Ź",
+	Zcaron: "Ž",
+	Zcircumflex: "Ẑ",
+	Zdotaccent: "Ż",
+	aacute: "á",
+	abreve: "ă",
+	acaron: "ǎ",
+	acircumflex: "â",
+	acute: "´",
+	adieresis: "ä",
+	adotaccent: "ȧ",
+	ae: "æ",
+	agrave: "à",
+	amacron: "ā",
+	ampersand: "&",
+	aogonek: "ą",
+	aring: "å",
+	asciicircum: "^",
+	asciitilde: "~",
+	asterisk: "*",
+	at: "@",
+	atilde: "ã",
+	backslash: "\\",
+	bar: "|",
+	bdotaccent: "ḃ",
+	braceleft: "{",
+	braceright: "}",
+	bracketleft: "[",
+	bracketright: "]",
+	breve: "˘",
+	brokenbar: "¦",
+	bullet: "•",
+	cacute: "ć",
+	caron: "ˇ",
+	ccaron: "č",
+	ccedilla: "ç",
+	ccircumflex: "ĉ",
+	cdotaccent: "ċ",
+	cedilla: "¸",
+	cent: "¢",
+	circumflex: "ˆ",
+	colon: ":",
+	comma: ",",
+	copyright: "©",
+	currency: "¤",
+	dagger: "†",
+	daggerdbl: "‡",
+	dcaron: "ď",
+	dcedilla: "ḑ",
+	ddotaccent: "ḋ",
+	degree: "°",
+	dieresis: "¨",
+	divide: "÷",
+	dollar: "$",
+	dotaccent: "˙",
+	dotlessi: "ı",
+	eacute: "é",
+	ebreve: "ĕ",
+	ecaron: "ě",
+	ecedilla: "ȩ",
+	ecircumflex: "ê",
+	edieresis: "ë",
+	edotaccent: "ė",
+	egrave: "è",
+	eight: "8",
+	ellipsis: "…",
+	emacron: "ē",
+	emdash: "—",
+	endash: "–",
+	eogonek: "ę",
+	equal: "=",
+	eth: "ð",
+	etilde: "ẽ",
+	euro: "€",
+	exclam: "!",
+	exclamdown: "¡",
+	fdotaccent: "ḟ",
+	ff: "ff",
+	ffi: "ffi",
+	ffl: "ffl",
+	fi: "fi",
+	five: "5",
+	fl: "fl",
+	florin: "ƒ",
+	four: "4",
+	fraction: "⁄",
+	gacute: "ǵ",
+	gbreve: "ğ",
+	gcaron: "ǧ",
+	gcedilla: "ģ",
+	gcircumflex: "ĝ",
+	gdotaccent: "ġ",
+	germandbls: "ß",
+	gmacron: "ḡ",
+	grave: "`",
+	greater: ">",
+	guillemotleft: "«",
+	guillemotright: "»",
+	guilsinglleft: "‹",
+	guilsinglright: "›",
+	hcaron: "ȟ",
+	hcedilla: "ḩ",
+	hcircumflex: "ĥ",
+	hdieresis: "ḧ",
+	hdotaccent: "ḣ",
+	hungarumlaut: "˝",
+	hyphen: "-",
+	iacute: "í",
+	ibreve: "ĭ",
+	icaron: "ǐ",
+	icircumflex: "î",
+	idieresis: "ï",
+	igrave: "ì",
+	imacron: "ī",
+	iogonek: "į",
+	itilde: "ĩ",
+	jcaron: "ǰ",
+	jcircumflex: "ĵ",
+	kacute: "ḱ",
+	kcaron: "ǩ",
+	kcedilla: "ķ",
+	lacute: "ĺ",
+	lcaron: "ľ",
+	lcedilla: "ļ",
+	less: "<",
+	logicalnot: "¬",
+	macron: "¯",
+	macute: "ḿ",
+	mdotaccent: "ṁ",
+	minus: "−",
+	mu: "µ",
+	multiply: "×",
+	nacute: "ń",
+	nbspace: "\xA0",
+	ncaron: "ň",
+	ncedilla: "ņ",
+	ndotaccent: "ṅ",
+	ngrave: "ǹ",
+	nine: "9",
+	ntilde: "ñ",
+	numbersign: "#",
+	oacute: "ó",
+	obreve: "ŏ",
+	ocaron: "ǒ",
+	ocircumflex: "ô",
+	odieresis: "ö",
+	odotaccent: "ȯ",
+	oe: "œ",
+	ogonek: "˛",
+	ograve: "ò",
+	ohungarumlaut: "ő",
+	omacron: "ō",
+	one: "1",
+	onehalf: "½",
+	onequarter: "¼",
+	onesuperior: "¹",
+	oogonek: "ǫ",
+	ordfeminine: "ª",
+	ordmasculine: "º",
+	oslash: "ø",
+	otilde: "õ",
+	pacute: "ṕ",
+	paragraph: "¶",
+	parenleft: "(",
+	parenright: ")",
+	pdotaccent: "ṗ",
+	percent: "%",
+	period: ".",
+	perthousand: "‰",
+	plus: "+",
+	plusminus: "±",
+	question: "?",
+	questiondown: "¿",
+	quotedbl: "\"",
+	quotedblbase: "„",
+	quotedblleft: "“",
+	quotedblright: "”",
+	quoteleft: "‘",
+	quoteright: "’",
+	quotesinglbase: "‚",
+	quotesingle: "'",
+	racute: "ŕ",
+	rcaron: "ř",
+	rcedilla: "ŗ",
+	rdotaccent: "ṙ",
+	registered: "®",
+	ring: "˚",
+	sacute: "ś",
+	scaron: "š",
+	scedilla: "ş",
+	scircumflex: "ŝ",
+	sdotaccent: "ṡ",
+	section: "§",
+	semicolon: ";",
+	seven: "7",
+	sfthyphen: "­",
+	six: "6",
+	slash: "/",
+	softhyphen: "­",
+	space: " ",
+	sterling: "£",
+	tcaron: "ť",
+	tcedilla: "ţ",
+	tdieresis: "ẗ",
+	tdotaccent: "ṫ",
+	thorn: "þ",
+	three: "3",
+	threequarters: "¾",
+	threesuperior: "³",
+	tilde: "˜",
+	trademark: "™",
+	two: "2",
+	twosuperior: "²",
+	uacute: "ú",
+	ubreve: "ŭ",
+	ucaron: "ǔ",
+	ucircumflex: "û",
+	udieresis: "ü",
+	ugrave: "ù",
+	uhungarumlaut: "ű",
+	umacron: "ū",
+	underscore: "_",
+	uogonek: "ų",
+	uring: "ů",
+	utilde: "ũ",
+	vtilde: "ṽ",
+	wacute: "ẃ",
+	wcircumflex: "ŵ",
+	wdieresis: "ẅ",
+	wdotaccent: "ẇ",
+	wgrave: "ẁ",
+	wring: "ẘ",
+	xdieresis: "ẍ",
+	xdotaccent: "ẋ",
+	yacute: "ý",
+	ycircumflex: "ŷ",
+	ydieresis: "ÿ",
+	ydotaccent: "ẏ",
+	yen: "¥",
+	ygrave: "ỳ",
+	ymacron: "ȳ",
+	yring: "ẙ",
+	ytilde: "ỹ",
+	zacute: "ź",
+	zcaron: "ž",
+	zcircumflex: "ẑ",
+	zdotaccent: "ż",
+	zero: "0"
+};
+/**
+* The text a glyph name stands for, or `undefined` when the name says nothing.
+*
+* @param name The name as `/Differences` (or a font program's charset) gives it.
+*/
+function textForGlyphName(name) {
+	const base = name.includes(".") ? name.split(".")[0] ?? name : name;
+	if (base === "") return void 0;
+	if (base.includes("_")) {
+		const parts = base.split("_").map((p) => textForGlyphName(p));
+		return parts.every((p) => p !== void 0) ? parts.join("") : void 0;
+	}
+	const known = GLYPHS[base];
+	if (known !== void 0) return known;
+	if (base.length === 1) return base;
+	const uni = /^uni((?:[0-9A-Fa-f]{4})+)$/u.exec(base);
+	if (uni) {
+		const hex = uni[1];
+		let out = "";
+		for (let i = 0; i < hex.length; i += 4) out += String.fromCodePoint(parseInt(hex.slice(i, i + 4), 16));
+		return out;
+	}
+	const u = /^u([0-9A-Fa-f]{4,6})$/u.exec(base);
+	if (u) {
+		const cp = parseInt(u[1], 16);
+		return cp <= 1114111 ? String.fromCodePoint(cp) : void 0;
+	}
+}
+//#endregion
+export { textForGlyphName };

package/dist/esm/pdf-reader/image-decode.js

@@ -3,6 +3,7 @@ import { encodePng } from "../core/png-encode.js";
 import { lzwDecodeMsb } from "../core/lzw.js";
 import { reversePredictor } from "./predictor.js";
 import { decodeCcitt } from "./ccitt.js";
+import { decodeJbig2 } from "./jbig2.js";
 import { decodeJpeg } from "./jpeg.js";
 import { unzlibSync } from "fflate";
 //#region src/pdf-reader/image-decode.ts
@@ -52,7 +53,17 @@ function decodePdfImage(file, stream) {
 		heightPx: height,
 		degraded: "JPEG 2000 image — limited viewer support"
 	};
-	if (last === "JBIG2Decode") return fail("dropped", "JBIG2-encoded image not decoded");
+	if (last === "JBIG2Decode") {
+		const raster = decodeJbig2Image(file, stream, filters, width, height);
+		if (typeof raster === "string") return fail("dropped", raster);
+		return {
+			ok: true,
+			bytes: encodePng(width, height, raster.color, raster.samples),
+			format: "png",
+			widthPx: width,
+			heightPx: height
+		};
+	}
 	const decoded = last === "CCITTFaxDecode" || last === "CCF" ? decodeCcittImage(file, stream, filters, width, height) : decodeToSamples(file, stream, filters, width, height);
 	if (typeof decoded === "string") return fail("dropped", decoded);
 	const alpha = decodeSMask(file, d, width, height);
@@ -75,6 +86,26 @@ function decodeToSamples(file, stream, filters, width, height) {
 	const decodeArr = decodeArrayOf(file, d);
 	return toColor(cs, unpackSamples(raw, width, height, cs.components, bpc), width * height, bpc, decodeArr);
 }
+/**
+* §7.4.7 — a `/JBIG2Decode` image: the stream's own segments, plus the shared
+* ones a `/JBIG2Globals` holds (the symbol dictionary a scanned book stores
+* once for all its pages).
+*/
+function decodeJbig2Image(file, stream, filters, width, height) {
+	const parms = decodeParmsOf(file, stream.dict);
+	const globalsVal = parms ? file.get(parms, "JBIG2Globals") : void 0;
+	const globalsStream = globalsVal instanceof PdfStream ? globalsVal : void 0;
+	const globals = globalsStream ? file.streamData(globalsStream) : void 0;
+	const packed = decodeJbig2(applyChainExceptLast(filters, stream.data), globals, width, height);
+	if (!packed) return "JBIG2 image not decoded";
+	const rowBytes = width + 7 >> 3;
+	const samples = new Uint8Array(width * height);
+	for (let y = 0; y < height; y++) for (let x = 0; x < width; x++) samples[y * width + x] = packed[y * rowBytes + (x >> 3)] >> 7 - (x & 7) & 1 ? 0 : 255;
+	return {
+		color: "gray",
+		samples
+	};
+}
 function decodeCcittImage(file, stream, filters, width, height) {
 	const parms = decodeParmsOf(file, stream.dict);
 	const columns = (parms ? intOf(file.get(parms, "Columns")) : 0) || 1728;