pdf-codec.js 5.2.26 → 5.2.28

package/dist/interpret.js

@@ -7,6 +7,8 @@ import { COLOR_BLACK } from "document-schema.js";
 //#region src/interpret.ts
 const MAX_FORM_XOBJECT_DEPTH = 12;
 const FALLBACK_GLYPH_WIDTH_PER_1000 = 500;
+const GLYPH_SPACE_PER_TEXT_SPACE_UNIT = 1e3;
+const ASCII_SPACE = 32;
 const DEFAULT_LINE_WIDTH_PT = 1;
 function initialTextParameters() {
 	return {
@@ -71,31 +73,37 @@ function rgbColor(operands) {
 		b: numAt(operands, 2)
 	};
 }
+const CMYK_K_INDEX = 3;
 function cmykColor(operands) {
 	const c = numAt(operands, 0);
 	const m = numAt(operands, 1);
 	const y = numAt(operands, 2);
-	const k = numAt(operands, 3);
+	const k = numAt(operands, CMYK_K_INDEX);
 	return {
 		r: (1 - c) * (1 - k),
 		g: (1 - m) * (1 - k),
 		b: (1 - y) * (1 - k)
 	};
 }
+const RGB_COMPONENT_COUNT = 3;
+const CMYK_COMPONENT_COUNT = 4;
 function genericColor(operands) {
 	const numericOperands = operands.filter((o) => o.kind === "number");
 	if (numericOperands.length === 1) return grayColor(numAt(numericOperands, 0));
-	if (numericOperands.length === 3) return rgbColor(numericOperands);
-	if (numericOperands.length === 4) return cmykColor(numericOperands);
+	if (numericOperands.length === RGB_COMPONENT_COUNT) return rgbColor(numericOperands);
+	if (numericOperands.length === CMYK_COMPONENT_COUNT) return cmykColor(numericOperands);
 }
+const MATRIX_D_INDEX = 3;
+const MATRIX_E_INDEX = 4;
+const MATRIX_F_INDEX = 5;
 function matrixFromOperands(operands) {
 	return [
 		numAt(operands, 0),
 		numAt(operands, 1),
 		numAt(operands, 2),
-		numAt(operands, 3),
-		numAt(operands, 4),
-		numAt(operands, 5)
+		numAt(operands, MATRIX_D_INDEX),
+		numAt(operands, MATRIX_E_INDEX),
+		numAt(operands, MATRIX_F_INDEX)
 	];
 }
 const SHAPE_ABS_TOLERANCE_PT = .001;
@@ -139,9 +147,10 @@ function closedPolygonCorners(subpath) {
 	if (nearlyEqualPoints(first, last, shapeTolerance(Math.max(bounds.maxX - bounds.minX, bounds.maxY - bounds.minY)))) corners.pop();
 	return corners;
 }
+const RECTANGLE_CORNER_COUNT = 4;
 function detectRect(subpath, paint) {
 	const corners = closedPolygonCorners(subpath);
-	if (corners?.length !== 4) return;
+	if (corners?.length !== RECTANGLE_CORNER_COUNT) return;
 	const { minX, minY, maxX, maxY } = boundsOf(corners);
 	const widthPt = maxX - minX;
 	const heightPt = maxY - minY;
@@ -189,9 +198,10 @@ function expectedEllipseControl(atExtreme, atExtremeKind, otherEnd, center, kx,
 		y: atExtreme.y
 	};
 }
+const ELLIPSE_SEGMENT_COUNT = 4;
 function detectEllipse(subpath, paint) {
 	const segments = subpath.segments;
-	if (!subpath.closed || segments.length !== 4 || segments.some((segment) => segment.kind !== "cubic")) return;
+	if (!subpath.closed || segments.length !== ELLIPSE_SEGMENT_COUNT || segments.some((segment) => segment.kind !== "cubic")) return;
 	const start = {
 		x: subpath.startXPt,
 		y: subpath.startYPt
@@ -217,16 +227,16 @@ function detectEllipse(subpath, paint) {
 		y: finalSegment.yPt
 	}, start, Math.max(tolX, tolY))) return;
 	const extremes = onCurve.map((point) => cardinalExtremeOf(point, center, rx, ry, tolX, tolY));
-	if (extremes.some((extreme) => extreme === void 0) || new Set(extremes).size !== 4) return;
+	if (extremes.some((extreme) => extreme === void 0) || new Set(extremes).size !== ELLIPSE_SEGMENT_COUNT) return;
 	const kx = rx * BEZIER_KAPPA;
 	const ky = ry * BEZIER_KAPPA;
 	const controlTolerance = Math.max(tolX, tolY);
-	for (let i = 0; i < 4; i += 1) {
+	for (let i = 0; i < ELLIPSE_SEGMENT_COUNT; i += 1) {
 		const segment = segments[i];
 		const from = onCurve[i];
 		const fromKind = extremes[i];
-		const to = onCurve[(i + 1) % 4];
-		const toKind = extremes[(i + 1) % 4];
+		const to = onCurve[(i + 1) % ELLIPSE_SEGMENT_COUNT];
+		const toKind = extremes[(i + 1) % ELLIPSE_SEGMENT_COUNT];
 		if (segment?.kind !== "cubic" || from === void 0 || to === void 0 || fromKind === void 0 || toKind === void 0) return;
 		const expectedC1 = expectedEllipseControl(from, fromKind, to, center, kx, ky);
 		const expectedC2 = expectedEllipseControl(to, toKind, from, center, kx, ky);
@@ -273,17 +283,38 @@ function classifyShape(subpaths, paint, style) {
 	if (subpaths.length !== 1 || subpath === void 0) return;
 	return detectRect(subpath, paint) ?? detectEllipse(subpath, paint) ?? detectLine(subpath, paint, style);
 }
+var MarkedContentStack = class {
+	#frames;
+	constructor(seed = []) {
+		this.#frames = [...seed];
+	}
+	open(frame) {
+		this.#frames.push(frame);
+	}
+	close() {
+		this.#frames.pop();
+	}
+	inScope(key) {
+		for (let i = this.#frames.length - 1; i >= 0; i--) {
+			const frame = this.#frames[i];
+			if (frame !== void 0 && key in frame) return frame[key];
+		}
+	}
+};
+const RE_HEIGHT_INDEX = 3;
+const OPERAND_INDEX_3 = 3;
+const OPERAND_INDEX_4 = 4;
+const OPERAND_INDEX_5 = 5;
+const TZ_PERCENT_DIVISOR = 100;
 function interpretContentStream(bytes, resources, context) {
-	const items = [];
-	runContentStream(bytes, resources, {
+	return runContentStream(bytes, resources, {
 		ctm: IDENTITY_MATRIX,
 		fillColor: COLOR_BLACK,
 		strokeColor: COLOR_BLACK,
 		lineWidth: DEFAULT_LINE_WIDTH_PT,
 		dashArray: [],
 		...initialTextParameters()
-	}, context, items, 0, []);
-	return items;
+	}, context, 0, new MarkedContentStack());
 }
 function lastPointOf(subpath) {
 	const last = subpath.segments[subpath.segments.length - 1];
@@ -295,19 +326,15 @@ function lastPointOf(subpath) {
 		y: last.yPt
 	};
 }
-function runContentStream(bytes, resources, initialState, context, items, depth, markedContent) {
+function runContentStream(bytes, resources, initialState, context, depth, markedContent) {
+	const items = [];
 	const operations = readContentStream(bytes, context.sink);
 	const gsStack = [];
 	let gs = initialState;
 	let text = defaultTextObjectState();
 	let pathSubpaths = [];
 	let currentSubpath;
-	const inScope = (key) => {
-		for (let i = markedContent.length - 1; i >= 0; i--) {
-			const frame = markedContent[i];
-			if (frame !== void 0 && key in frame) return frame[key];
-		}
-	};
+	const inScope = (key) => markedContent.inScope(key);
 	const pushItem = (item) => {
 		if (item.kind === "text") {
 			const layer = item.layerName ?? inScope("layerName");
@@ -349,7 +376,7 @@ function runContentStream(bytes, resources, initialState, context, items, depth,
 		const x = numAt(operands, 0);
 		const y = numAt(operands, 1);
 		const w = numAt(operands, 2);
-		const h = numAt(operands, 3);
+		const h = numAt(operands, RE_HEIGHT_INDEX);
 		const p1 = applyMatrix(ctm, {
 			x,
 			y
@@ -431,10 +458,10 @@ function runContentStream(bytes, resources, initialState, context, items, depth,
 			});
 			const widthPer1000 = glyph?.widthPer1000 ?? FALLBACK_GLYPH_WIDTH_PER_1000;
 			const byteLength = glyph?.byteLengthConsumed ?? 1;
-			const isSingleByteSpace = byteLength === 1 && codes[offset] === 32;
+			const isSingleByteSpace = byteLength === 1 && codes[offset] === ASCII_SPACE;
 			const spacing = gs.charSpace + (isSingleByteSpace ? gs.wordSpace : 0);
 			const vertical = glyph?.vertical;
-			text.tm = vertical === void 0 ? multiplyMatrices(translationMatrix((widthPer1000 / 1e3 * gs.fontSizePt + spacing) * gs.horizScale, 0), text.tm) : multiplyMatrices(translationMatrix(0, vertical.displacementPer1000 / 1e3 * gs.fontSizePt + spacing), text.tm);
+			text.tm = vertical === void 0 ? multiplyMatrices(translationMatrix((widthPer1000 / GLYPH_SPACE_PER_TEXT_SPACE_UNIT * gs.fontSizePt + spacing) * gs.horizScale, 0), text.tm) : multiplyMatrices(translationMatrix(0, vertical.displacementPer1000 / GLYPH_SPACE_PER_TEXT_SPACE_UNIT * gs.fontSizePt + spacing), text.tm);
 			offset += byteLength;
 		}
 	};
@@ -444,8 +471,8 @@ function runContentStream(bytes, resources, initialState, context, items, depth,
 		const vertical = context.fontMetrics.glyphAdvance(fontResourceName, resources, codes, 0)?.vertical;
 		if (vertical === void 0) return;
 		return {
-			x: vertical.positionXPer1000 / 1e3,
-			y: vertical.positionYPer1000 / 1e3
+			x: vertical.positionXPer1000 / GLYPH_SPACE_PER_TEXT_SPACE_UNIT,
+			y: vertical.positionYPer1000 / GLYPH_SPACE_PER_TEXT_SPACE_UNIT
 		};
 	};
 	const showTextArray = (elements) => {
@@ -462,7 +489,7 @@ function runContentStream(bytes, resources, initialState, context, items, depth,
 			totalLength += el.bytes.length;
 			advanceThroughString(el.bytes);
 		} else if (el.kind === "number") {
-			const adjustment = -(el.value / 1e3) * gs.fontSizePt;
+			const adjustment = -(el.value / GLYPH_SPACE_PER_TEXT_SPACE_UNIT) * gs.fontSizePt;
 			text.tm = multiplyMatrices(vertical ? translationMatrix(0, adjustment) : translationMatrix(adjustment * gs.horizScale, 0), text.tm);
 		}
 		if (totalLength === 0) return;
@@ -540,10 +567,10 @@ function runContentStream(bytes, resources, initialState, context, items, depth,
 			};
 			const formLayer = ownLayer ?? inScope("layerName");
 			const formMcid = dictGet(xobj.dict, "StructParents") !== void 0 ? void 0 : inScope("mcid");
-			runContentStream(decoded.bytes, formResources, formState, context, items, depth + 1, [{
+			for (const formItem of runContentStream(decoded.bytes, formResources, formState, context, depth + 1, new MarkedContentStack([{
 				...formLayer !== void 0 ? { layerName: formLayer } : {},
 				...formMcid !== void 0 ? { mcid: formMcid } : {}
-			}]);
+			}]))) items.push(formItem);
 		}
 	};
 	for (const token of operations) {
@@ -660,11 +687,11 @@ function runContentStream(bytes, resources, initialState, context, items, depth,
 				});
 				const c2 = applyMatrix(gs.ctm, {
 					x: numAt(operands, 2),
-					y: numAt(operands, 3)
+					y: numAt(operands, OPERAND_INDEX_3)
 				});
 				const p = applyMatrix(gs.ctm, {
-					x: numAt(operands, 4),
-					y: numAt(operands, 5)
+					x: numAt(operands, OPERAND_INDEX_4),
+					y: numAt(operands, OPERAND_INDEX_5)
 				});
 				currentSubpath?.segments.push({
 					kind: "cubic",
@@ -686,7 +713,7 @@ function runContentStream(bytes, resources, initialState, context, items, depth,
 					});
 					const p = applyMatrix(gs.ctm, {
 						x: numAt(operands, 2),
-						y: numAt(operands, 3)
+						y: numAt(operands, OPERAND_INDEX_3)
 					});
 					currentSubpath.segments.push({
 						kind: "cubic",
@@ -706,7 +733,7 @@ function runContentStream(bytes, resources, initialState, context, items, depth,
 				});
 				const p = applyMatrix(gs.ctm, {
 					x: numAt(operands, 2),
-					y: numAt(operands, 3)
+					y: numAt(operands, OPERAND_INDEX_3)
 				});
 				currentSubpath?.segments.push({
 					kind: "cubic",
@@ -771,7 +798,7 @@ function runContentStream(bytes, resources, initialState, context, items, depth,
 			case "Tz":
 				gs = {
 					...gs,
-					horizScale: numAt(operands, 0) / 100
+					horizScale: numAt(operands, 0) / TZ_PERCENT_DIVISOR
 				};
 				break;
 			case "TL":
@@ -842,18 +869,19 @@ function runContentStream(bytes, resources, initialState, context, items, depth,
 					...layerName !== void 0 ? { layerName } : {},
 					...markedContentScopeProps(props)
 				};
-				markedContent.push(depth === 0 ? scopeProps : {
+				markedContent.open(depth === 0 ? scopeProps : {
 					...scopeProps,
 					mcid: void 0
 				});
 				break;
 			}
 			case "BMC":
-				markedContent.push({});
+				markedContent.open({});
 				break;
-			case "EMC": markedContent.pop();
+			case "EMC": markedContent.close();
 		}
 	}
+	return items;
 }
 //#endregion
 export { interpretContentStream };

package/dist/lexer.cjs

@@ -1,44 +1,79 @@
 Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
 const require_bytes_reader = require("./bytes/reader.cjs");
 //#region src/lexer.ts
+const LEFT_PAREN = 40;
+const RIGHT_PAREN = 41;
+const LESS_THAN = 60;
+const GREATER_THAN = 62;
+const LEFT_BRACKET = 91;
+const RIGHT_BRACKET = 93;
+const LEFT_BRACE = 123;
+const RIGHT_BRACE = 125;
+const SOLIDUS = 47;
+const PERCENT_SIGN = 37;
+const REVERSE_SOLIDUS = 92;
+const NUMBER_SIGN = 35;
+const PLUS_SIGN = 43;
+const MINUS_SIGN = 45;
+const FULL_STOP = 46;
+const LINE_FEED = 10;
+const CARRIAGE_RETURN = 13;
+const HORIZONTAL_TAB = 9;
+const BACKSPACE = 8;
+const FORM_FEED = 12;
+const DIGIT_ZERO = 48;
+const DIGIT_SEVEN = 55;
+const DIGIT_NINE = 57;
+const UPPER_A = 65;
+const UPPER_F = 70;
+const LOWER_A = 97;
+const LOWER_B = 98;
+const LOWER_F = 102;
+const LOWER_N = 110;
+const LOWER_R = 114;
+const LOWER_T = 116;
+const HEX_DIGIT_ALPHA_OFFSET = 10;
+const HEX_RADIX = 16;
+const OCTAL_RADIX = 8;
+const BYTE_MASK = 255;
 const DELIMITER_BYTES = /* @__PURE__ */ new Set([
-	40,
-	41,
-	60,
-	62,
-	91,
-	93,
-	123,
-	125,
-	47,
-	37
+	LEFT_PAREN,
+	RIGHT_PAREN,
+	LESS_THAN,
+	GREATER_THAN,
+	LEFT_BRACKET,
+	RIGHT_BRACKET,
+	LEFT_BRACE,
+	RIGHT_BRACE,
+	SOLIDUS,
+	PERCENT_SIGN
 ]);
 function isRegularByte(byte) {
 	return !require_bytes_reader.isAsciiWhitespace(byte) && !DELIMITER_BYTES.has(byte);
 }
 function isDigit(byte) {
-	return byte !== void 0 && byte >= 48 && byte <= 57;
+	return byte !== void 0 && byte >= DIGIT_ZERO && byte <= DIGIT_NINE;
 }
 function isHexDigit(byte) {
-	return byte !== void 0 && (byte >= 48 && byte <= 57 || byte >= 65 && byte <= 70 || byte >= 97 && byte <= 102);
+	return byte !== void 0 && (byte >= DIGIT_ZERO && byte <= DIGIT_NINE || byte >= UPPER_A && byte <= UPPER_F || byte >= LOWER_A && byte <= LOWER_F);
 }
 function hexDigitValue(byte) {
-	if (byte >= 48 && byte <= 57) return byte - 48;
-	if (byte >= 65 && byte <= 70) return byte - 65 + 10;
-	return byte - 97 + 10;
+	if (byte >= DIGIT_ZERO && byte <= DIGIT_NINE) return byte - DIGIT_ZERO;
+	if (byte >= UPPER_A && byte <= UPPER_F) return byte - UPPER_A + HEX_DIGIT_ALPHA_OFFSET;
+	return byte - LOWER_A + HEX_DIGIT_ALPHA_OFFSET;
 }
 function skipWhitespaceAndComments(reader) {
 	for (;;) {
 		reader.skipWhitespace();
-		if (reader.peek() !== 37) return;
-		while (!reader.atEnd() && reader.peek() !== 10 && reader.peek() !== 13) reader.next();
+		if (reader.peek() !== PERCENT_SIGN) return;
+		while (!reader.atEnd() && reader.peek() !== LINE_FEED && reader.peek() !== CARRIAGE_RETURN) reader.next();
 	}
 }
 function readNumberToken(reader) {
 	const start = reader.offset;
-	if (reader.peek() === 43 || reader.peek() === 45) reader.next();
+	if (reader.peek() === PLUS_SIGN || reader.peek() === MINUS_SIGN) reader.next();
 	while (isDigit(reader.peek())) reader.next();
-	if (reader.peek() === 46) {
+	if (reader.peek() === FULL_STOP) {
 		reader.next();
 		while (isDigit(reader.peek())) reader.next();
 	}
@@ -54,11 +89,11 @@ function readNameToken(reader) {
 	for (;;) {
 		const byte = reader.peek();
 		if (byte === void 0 || !isRegularByte(byte)) break;
-		if (byte === 35 && isHexDigit(reader.peek(1)) && isHexDigit(reader.peek(2))) {
+		if (byte === NUMBER_SIGN && isHexDigit(reader.peek(1)) && isHexDigit(reader.peek(2))) {
 			reader.next();
 			const hi = hexDigitValue(reader.next());
 			const lo = hexDigitValue(reader.next());
-			bytes.push(hi * 16 + lo);
+			bytes.push(hi * HEX_RADIX + lo);
 		} else bytes.push(reader.next());
 	}
 	return {
@@ -73,38 +108,38 @@ function readLiteralStringToken(reader) {
 	for (;;) {
 		const byte = reader.next();
 		if (byte === void 0) break;
-		if (byte === 92) {
+		if (byte === REVERSE_SOLIDUS) {
 			const esc = reader.next();
 			if (esc === void 0) break;
-			if (esc === 110) bytes.push(10);
-			else if (esc === 114) bytes.push(13);
-			else if (esc === 116) bytes.push(9);
-			else if (esc === 98) bytes.push(8);
-			else if (esc === 102) bytes.push(12);
-			else if (esc === 40 || esc === 41 || esc === 92) bytes.push(esc);
-			else if (esc === 13) {
-				if (reader.peek() === 10) reader.next();
-			} else if (esc === 10) {} else if (esc >= 48 && esc <= 55) {
-				let value = esc - 48;
-				for (let i = 0; i < 2 && reader.peek() !== void 0 && reader.peek() >= 48 && reader.peek() <= 55; i++) value = value * 8 + (reader.next() - 48);
-				bytes.push(value & 255);
+			if (esc === LOWER_N) bytes.push(LINE_FEED);
+			else if (esc === LOWER_R) bytes.push(CARRIAGE_RETURN);
+			else if (esc === LOWER_T) bytes.push(HORIZONTAL_TAB);
+			else if (esc === LOWER_B) bytes.push(BACKSPACE);
+			else if (esc === LOWER_F) bytes.push(FORM_FEED);
+			else if (esc === LEFT_PAREN || esc === RIGHT_PAREN || esc === REVERSE_SOLIDUS) bytes.push(esc);
+			else if (esc === CARRIAGE_RETURN) {
+				if (reader.peek() === LINE_FEED) reader.next();
+			} else if (esc === LINE_FEED) {} else if (esc >= DIGIT_ZERO && esc <= DIGIT_SEVEN) {
+				let value = esc - DIGIT_ZERO;
+				for (let i = 0; i < 2 && reader.peek() !== void 0 && reader.peek() >= DIGIT_ZERO && reader.peek() <= DIGIT_SEVEN; i++) value = value * OCTAL_RADIX + (reader.next() - DIGIT_ZERO);
+				bytes.push(value & BYTE_MASK);
 			} else bytes.push(esc);
 			continue;
 		}
-		if (byte === 40) {
+		if (byte === LEFT_PAREN) {
 			depth++;
 			bytes.push(byte);
 			continue;
 		}
-		if (byte === 41) {
+		if (byte === RIGHT_PAREN) {
 			depth--;
 			if (depth === 0) break;
 			bytes.push(byte);
 			continue;
 		}
-		if (byte === 13) {
-			if (reader.peek() === 10) reader.next();
-			bytes.push(10);
+		if (byte === CARRIAGE_RETURN) {
+			if (reader.peek() === LINE_FEED) reader.next();
+			bytes.push(LINE_FEED);
 			continue;
 		}
 		bytes.push(byte);
@@ -119,12 +154,12 @@ function readHexStringToken(reader) {
 	const digits = [];
 	for (;;) {
 		const byte = reader.next();
-		if (byte === void 0 || byte === 62) break;
+		if (byte === void 0 || byte === GREATER_THAN) break;
 		if (isHexDigit(byte)) digits.push(hexDigitValue(byte));
 	}
 	if (digits.length % 2 === 1) digits.push(0);
 	const bytes = new Uint8Array(digits.length / 2);
-	for (let i = 0; i < bytes.length; i++) bytes[i] = digits[i * 2] * 16 + digits[i * 2 + 1];
+	for (let i = 0; i < bytes.length; i++) bytes[i] = digits[i * 2] * HEX_RADIX + digits[i * 2 + 1];
 	return {
 		kind: "hexString",
 		value: bytes
@@ -142,18 +177,18 @@ function nextToken(reader) {
 	skipWhitespaceAndComments(reader);
 	const byte = reader.peek();
 	if (byte === void 0) return;
-	if (byte === 47) return readNameToken(reader);
-	if (byte === 40) return readLiteralStringToken(reader);
-	if (byte === 60) {
-		if (reader.peek(1) === 60) {
+	if (byte === SOLIDUS) return readNameToken(reader);
+	if (byte === LEFT_PAREN) return readLiteralStringToken(reader);
+	if (byte === LESS_THAN) {
+		if (reader.peek(1) === LESS_THAN) {
 			reader.next();
 			reader.next();
 			return { kind: "dictStart" };
 		}
 		return readHexStringToken(reader);
 	}
-	if (byte === 62) {
-		if (reader.peek(1) === 62) {
+	if (byte === GREATER_THAN) {
+		if (reader.peek(1) === GREATER_THAN) {
 			reader.next();
 			reader.next();
 			return { kind: "dictEnd" };
@@ -161,19 +196,19 @@ function nextToken(reader) {
 		reader.next();
 		return nextToken(reader);
 	}
-	if (byte === 91) {
+	if (byte === LEFT_BRACKET) {
 		reader.next();
 		return { kind: "arrayStart" };
 	}
-	if (byte === 93) {
+	if (byte === RIGHT_BRACKET) {
 		reader.next();
 		return { kind: "arrayEnd" };
 	}
-	if (byte === 123 || byte === 125) {
+	if (byte === LEFT_BRACE || byte === RIGHT_BRACE) {
 		reader.next();
 		return nextToken(reader);
 	}
-	if (byte === 43 || byte === 45 || byte === 46 || isDigit(byte)) return readNumberToken(reader);
+	if (byte === PLUS_SIGN || byte === MINUS_SIGN || byte === FULL_STOP || isDigit(byte)) return readNumberToken(reader);
 	return readKeywordToken(reader);
 }
 //#endregion