pdf-codec 5.7.3 → 5.8.0

package/dist/text-group.d.ts

@@ -37,6 +37,12 @@ interface PdfTextGroupingOptions {
 declare const DEFAULT_BASELINE_TOLERANCE_EM: number;
 declare const DEFAULT_WORD_GAP_EM: number;
 declare const DEFAULT_COLUMN_GAP_EM = 1;
+/**
+ * An anticlockwise rotation as one canonical number: normalised into [0, 360) and rounded to remove the last-bit noise atan2 leaves, so two runs set at one angle report the identical value and a comparison against a quarter turn is exact.
+ * @param rotationDeg - a rotation in degrees, anticlockwise from the page's x-axis; absent means zero
+ * @returns the same angle in [0, 360)
+ */
+declare function normalisedRotationDeg(rotationDeg: number | undefined): number;
 /**
  * Whether two positioned runs sit on one visual line.
  *
@@ -47,6 +53,12 @@ declare const DEFAULT_COLUMN_GAP_EM = 1;
  * @returns true when the two baselines are within tolerance of each other
  */
 declare function runsShareBaseline(a: PdfTextRunGeometry, b: PdfTextRunGeometry, options?: PdfTextGroupingOptions): boolean;
+/**
+ * Where a run starts along the direction it advances in: the coordinate that puts the runs of one line in reading order, at whatever angle the line is set. For an upright run it is the run's `xPt`; for a run turned a quarter turn anticlockwise, read upward, it is its `yPt`.
+ * @param run - a positioned run
+ * @returns the run's origin projected onto its own baseline, in points
+ */
+declare function runAlongBaselinePt(run: PdfTextRunGeometry): number;
 /**
  * The horizontal gap along the baseline between the end of `previous` and the start of `next`, or undefined when the geometry does not state one.
  *
@@ -95,4 +107,4 @@ declare function attachScriptRuns<TLine, TRun extends PdfTextRunGeometry>(lines:
  */
 declare function groupPdfTextRuns<TRun extends PdfTextRunGeometry>(runs: readonly TRun[], options?: PdfTextGroupingOptions): PdfTextLine<TRun>[];
 //#endregion
-export { DEFAULT_BASELINE_TOLERANCE_EM, DEFAULT_COLUMN_GAP_EM, DEFAULT_WORD_GAP_EM, PdfLineWithScripts, PdfScriptPosition, PdfTextBox, PdfTextGroupingOptions, PdfTextLine, PdfTextRunGeometry, PdfTextWord, PdfWordSeparator, attachScriptRuns, groupPdfTextRuns, runGapPt, runsShareBaseline, scriptPositionOf };
+export { DEFAULT_BASELINE_TOLERANCE_EM, DEFAULT_COLUMN_GAP_EM, DEFAULT_WORD_GAP_EM, PdfLineWithScripts, PdfScriptPosition, PdfTextBox, PdfTextGroupingOptions, PdfTextLine, PdfTextRunGeometry, PdfTextWord, PdfWordSeparator, attachScriptRuns, groupPdfTextRuns, normalisedRotationDeg, runAlongBaselinePt, runGapPt, runsShareBaseline, scriptPositionOf };

package/dist/text-group.js

@@ -4,9 +4,16 @@ const DEFAULT_COLUMN_GAP_EM = 1;
 const ROTATION_NOISE_DEG = 1e-6;
 const DEGREES_PER_TURN = 360;
 const VERTICAL_ADVANCE_OFFSET_DEG = -90;
+/**
+* An anticlockwise rotation as one canonical number: normalised into [0, 360) and rounded to remove the last-bit noise atan2 leaves, so two runs set at one angle report the identical value and a comparison against a quarter turn is exact.
+* @param rotationDeg - a rotation in degrees, anticlockwise from the page's x-axis; absent means zero
+* @returns the same angle in [0, 360)
+*/
+function normalisedRotationDeg(rotationDeg) {
+	return (Math.round((rotationDeg ?? 0) / ROTATION_NOISE_DEG) * ROTATION_NOISE_DEG % DEGREES_PER_TURN + DEGREES_PER_TURN) % DEGREES_PER_TURN;
+}
 function advanceAngleDeg(run) {
-	const raw = (run.rotationDeg ?? 0) + (run.writingMode === "vertical" ? VERTICAL_ADVANCE_OFFSET_DEG : 0);
-	return (Math.round(raw / ROTATION_NOISE_DEG) * ROTATION_NOISE_DEG % DEGREES_PER_TURN + DEGREES_PER_TURN) % DEGREES_PER_TURN;
+	return normalisedRotationDeg((run.rotationDeg ?? 0) + (run.writingMode === "vertical" ? VERTICAL_ADVANCE_OFFSET_DEG : 0));
 }
 const RADIANS_PER_TURN = 2 * Math.PI;
 function baselineAxis(rotationDeg) {
@@ -48,6 +55,14 @@ function runsShareBaseline(a, b, options = {}) {
 	return Math.abs(project(a, rotationDeg).acrossPt - project(b, rotationDeg).acrossPt) <= toleranceEm * Math.min(a.sizePt, b.sizePt);
 }
 /**
+* Where a run starts along the direction it advances in: the coordinate that puts the runs of one line in reading order, at whatever angle the line is set. For an upright run it is the run's `xPt`; for a run turned a quarter turn anticlockwise, read upward, it is its `yPt`.
+* @param run - a positioned run
+* @returns the run's origin projected onto its own baseline, in points
+*/
+function runAlongBaselinePt(run) {
+	return project(run, advanceAngleDeg(run)).alongPt;
+}
+/**
 * The horizontal gap along the baseline between the end of `previous` and the start of `next`, or undefined when the geometry does not state one.
 *
 * Undefined has exactly two causes, and a caller must treat both as "no evidence of a space" rather than substituting zero: `previous` stated no advance width, so where it ends is unknown; or the two runs are set at different angles, so they share no axis to measure along. A negative result is real and means the two runs overlap.
@@ -292,4 +307,4 @@ function groupPdfTextRuns(runs, options = {}) {
 	return result;
 }
 //#endregion
-export { DEFAULT_BASELINE_TOLERANCE_EM, DEFAULT_COLUMN_GAP_EM, DEFAULT_WORD_GAP_EM, attachScriptRuns, groupPdfTextRuns, runGapPt, runsShareBaseline, scriptPositionOf };
+export { DEFAULT_BASELINE_TOLERANCE_EM, DEFAULT_COLUMN_GAP_EM, DEFAULT_WORD_GAP_EM, attachScriptRuns, groupPdfTextRuns, normalisedRotationDeg, runAlongBaselinePt, runGapPt, runsShareBaseline, scriptPositionOf };

package/dist/winansi.cjs

@@ -1,5 +1,163 @@
 Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
-const require_afm_widths = require("./afm-widths-sjaRvmQX.cjs");
-exports.encodeForShow = require_afm_widths.encodeForShow;
-exports.sanitizeToWinAnsi = require_afm_widths.sanitizeToWinAnsi;
-exports.winAnsiCodeToUnicode = require_afm_widths.winAnsiCodeToUnicode;
+const require_afm_widths = require("./afm-widths.cjs");
+//#region src/winansi.ts
+const ASCII_PRINTABLE_START = 32;
+const ASCII_PRINTABLE_END = 127;
+const LATIN1_SUPPLEMENT_START = 160;
+const LATIN1_SUPPLEMENT_END = 255;
+function identityRangeEntries(start, end) {
+	const entries = [];
+	for (let code = start; code <= end; code++) entries.push([code, code]);
+	return entries;
+}
+const CP1252_EXTENDED_MAPPINGS = [
+	{
+		unicode: 338,
+		winAnsi: 140
+	},
+	{
+		unicode: 339,
+		winAnsi: 156
+	},
+	{
+		unicode: 352,
+		winAnsi: 138
+	},
+	{
+		unicode: 353,
+		winAnsi: 154
+	},
+	{
+		unicode: 376,
+		winAnsi: 159
+	},
+	{
+		unicode: 381,
+		winAnsi: 142
+	},
+	{
+		unicode: 382,
+		winAnsi: 158
+	},
+	{
+		unicode: 402,
+		winAnsi: 131
+	},
+	{
+		unicode: 710,
+		winAnsi: 136
+	},
+	{
+		unicode: 732,
+		winAnsi: 152
+	},
+	{
+		unicode: 8211,
+		winAnsi: 150
+	},
+	{
+		unicode: 8212,
+		winAnsi: 151
+	},
+	{
+		unicode: 8216,
+		winAnsi: 145
+	},
+	{
+		unicode: 8217,
+		winAnsi: 146
+	},
+	{
+		unicode: 8218,
+		winAnsi: 130
+	},
+	{
+		unicode: 8220,
+		winAnsi: 147
+	},
+	{
+		unicode: 8221,
+		winAnsi: 148
+	},
+	{
+		unicode: 8222,
+		winAnsi: 132
+	},
+	{
+		unicode: 8224,
+		winAnsi: 134
+	},
+	{
+		unicode: 8225,
+		winAnsi: 135
+	},
+	{
+		unicode: 8226,
+		winAnsi: 149
+	},
+	{
+		unicode: 8230,
+		winAnsi: 133
+	},
+	{
+		unicode: 8240,
+		winAnsi: 137
+	},
+	{
+		unicode: 8249,
+		winAnsi: 139
+	},
+	{
+		unicode: 8250,
+		winAnsi: 155
+	},
+	{
+		unicode: 8364,
+		winAnsi: 128
+	},
+	{
+		unicode: 8482,
+		winAnsi: 153
+	}
+];
+const UNICODE_TO_WINANSI = new Map([
+	...identityRangeEntries(ASCII_PRINTABLE_START, ASCII_PRINTABLE_END),
+	...identityRangeEntries(LATIN1_SUPPLEMENT_START, LATIN1_SUPPLEMENT_END),
+	...CP1252_EXTENDED_MAPPINGS.map((entry) => [entry.unicode, entry.winAnsi])
+]);
+const FALLBACK_BYTE = 63;
+const FALLBACK_CHAR = "?";
+function sanitizeToWinAnsi(text) {
+	const codes = [];
+	const substitutions = [];
+	for (const ch of text) {
+		const codePoint = ch.codePointAt(0);
+		const code = codePoint === void 0 ? void 0 : UNICODE_TO_WINANSI.get(codePoint);
+		if (code !== void 0) {
+			codes.push(code);
+			continue;
+		}
+		codes.push(FALLBACK_BYTE);
+		substitutions.push({
+			from: ch,
+			to: FALLBACK_CHAR
+		});
+	}
+	return {
+		codes: new Uint8Array(codes),
+		substitutions
+	};
+}
+function encodeForShow(text, font) {
+	const { codes, substitutions } = sanitizeToWinAnsi(text);
+	let width1000 = 0;
+	for (const code of codes) width1000 += require_afm_widths.widthOfCode(font, code);
+	return {
+		codes,
+		width1000,
+		substitutions
+	};
+}
+//#endregion
+exports.encodeForShow = encodeForShow;
+exports.sanitizeToWinAnsi = sanitizeToWinAnsi;

package/dist/winansi.d.cts

@@ -1,6 +1,5 @@
 import { r as StandardFontName } from "./afm-widths-D0Lu9kLv.cjs";
 //#region src/winansi.d.ts
-declare function winAnsiCodeToUnicode(code: number): number | undefined;
 interface WinAnsiSubstitution {
   readonly from: string;
   readonly to: string;
@@ -17,4 +16,4 @@ interface EncodedShow {
 }
 declare function encodeForShow(text: string, font: StandardFontName): EncodedShow;
 //#endregion
-export { EncodedShow, SanitizeResult, WinAnsiSubstitution, encodeForShow, sanitizeToWinAnsi, winAnsiCodeToUnicode };
+export { EncodedShow, SanitizeResult, WinAnsiSubstitution, encodeForShow, sanitizeToWinAnsi };

package/dist/winansi.d.ts

@@ -1,6 +1,5 @@
 import { r as StandardFontName } from "./afm-widths-D0Lu9kLv.js";
 //#region src/winansi.d.ts
-declare function winAnsiCodeToUnicode(code: number): number | undefined;
 interface WinAnsiSubstitution {
   readonly from: string;
   readonly to: string;
@@ -17,4 +16,4 @@ interface EncodedShow {
 }
 declare function encodeForShow(text: string, font: StandardFontName): EncodedShow;
 //#endregion
-export { EncodedShow, SanitizeResult, WinAnsiSubstitution, encodeForShow, sanitizeToWinAnsi, winAnsiCodeToUnicode };
+export { EncodedShow, SanitizeResult, WinAnsiSubstitution, encodeForShow, sanitizeToWinAnsi };

package/dist/winansi.js

@@ -1,2 +1,161 @@
-import { c as encodeForShow, l as sanitizeToWinAnsi, u as winAnsiCodeToUnicode } from "./afm-widths-BNI8qa2S.js";
-export { encodeForShow, sanitizeToWinAnsi, winAnsiCodeToUnicode };
+import { widthOfCode } from "./afm-widths.js";
+//#region src/winansi.ts
+const ASCII_PRINTABLE_START = 32;
+const ASCII_PRINTABLE_END = 127;
+const LATIN1_SUPPLEMENT_START = 160;
+const LATIN1_SUPPLEMENT_END = 255;
+function identityRangeEntries(start, end) {
+	const entries = [];
+	for (let code = start; code <= end; code++) entries.push([code, code]);
+	return entries;
+}
+const CP1252_EXTENDED_MAPPINGS = [
+	{
+		unicode: 338,
+		winAnsi: 140
+	},
+	{
+		unicode: 339,
+		winAnsi: 156
+	},
+	{
+		unicode: 352,
+		winAnsi: 138
+	},
+	{
+		unicode: 353,
+		winAnsi: 154
+	},
+	{
+		unicode: 376,
+		winAnsi: 159
+	},
+	{
+		unicode: 381,
+		winAnsi: 142
+	},
+	{
+		unicode: 382,
+		winAnsi: 158
+	},
+	{
+		unicode: 402,
+		winAnsi: 131
+	},
+	{
+		unicode: 710,
+		winAnsi: 136
+	},
+	{
+		unicode: 732,
+		winAnsi: 152
+	},
+	{
+		unicode: 8211,
+		winAnsi: 150
+	},
+	{
+		unicode: 8212,
+		winAnsi: 151
+	},
+	{
+		unicode: 8216,
+		winAnsi: 145
+	},
+	{
+		unicode: 8217,
+		winAnsi: 146
+	},
+	{
+		unicode: 8218,
+		winAnsi: 130
+	},
+	{
+		unicode: 8220,
+		winAnsi: 147
+	},
+	{
+		unicode: 8221,
+		winAnsi: 148
+	},
+	{
+		unicode: 8222,
+		winAnsi: 132
+	},
+	{
+		unicode: 8224,
+		winAnsi: 134
+	},
+	{
+		unicode: 8225,
+		winAnsi: 135
+	},
+	{
+		unicode: 8226,
+		winAnsi: 149
+	},
+	{
+		unicode: 8230,
+		winAnsi: 133
+	},
+	{
+		unicode: 8240,
+		winAnsi: 137
+	},
+	{
+		unicode: 8249,
+		winAnsi: 139
+	},
+	{
+		unicode: 8250,
+		winAnsi: 155
+	},
+	{
+		unicode: 8364,
+		winAnsi: 128
+	},
+	{
+		unicode: 8482,
+		winAnsi: 153
+	}
+];
+const UNICODE_TO_WINANSI = new Map([
+	...identityRangeEntries(ASCII_PRINTABLE_START, ASCII_PRINTABLE_END),
+	...identityRangeEntries(LATIN1_SUPPLEMENT_START, LATIN1_SUPPLEMENT_END),
+	...CP1252_EXTENDED_MAPPINGS.map((entry) => [entry.unicode, entry.winAnsi])
+]);
+const FALLBACK_BYTE = 63;
+const FALLBACK_CHAR = "?";
+function sanitizeToWinAnsi(text) {
+	const codes = [];
+	const substitutions = [];
+	for (const ch of text) {
+		const codePoint = ch.codePointAt(0);
+		const code = codePoint === void 0 ? void 0 : UNICODE_TO_WINANSI.get(codePoint);
+		if (code !== void 0) {
+			codes.push(code);
+			continue;
+		}
+		codes.push(FALLBACK_BYTE);
+		substitutions.push({
+			from: ch,
+			to: FALLBACK_CHAR
+		});
+	}
+	return {
+		codes: new Uint8Array(codes),
+		substitutions
+	};
+}
+function encodeForShow(text, font) {
+	const { codes, substitutions } = sanitizeToWinAnsi(text);
+	let width1000 = 0;
+	for (const code of codes) width1000 += widthOfCode(font, code);
+	return {
+		codes,
+		width1000,
+		substitutions
+	};
+}
+//#endregion
+export { encodeForShow, sanitizeToWinAnsi };

package/dist/write.cjs

@@ -1,5 +1,5 @@
 Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
-const require_afm_widths = require("./afm-widths-sjaRvmQX.cjs");
+const require_afm_widths = require("./afm-widths.cjs");
 const require_objects = require("./objects.cjs");
 const require_bytes_writer = require("./bytes/writer.cjs");
 const require_serialize = require("./serialize.cjs");

package/dist/write.js

@@ -1,4 +1,4 @@
-import { r as widthOfCode, t as STANDARD_METRICS } from "./afm-widths-BNI8qa2S.js";
+import { STANDARD_METRICS, widthOfCode } from "./afm-widths.js";
 import { pdfArray, pdfDict, pdfHexString, pdfLiteralString, pdfName, pdfNull, pdfNum, pdfRef, pdfStream } from "./objects.js";
 import { ByteWriter } from "./bytes/writer.js";
 import { writeObject } from "./serialize.js";

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pdf-codec",
-  "version": "5.7.3",
+  "version": "5.8.0",
   "description": "Hand-written, dependency-minimal PDF codec: parses arbitrary real-world PDFs and generates new ones, built on its own codec-owned LayoutDocument item model and Zod 4 codecs.",
   "type": "module",
   "repository": {