reamkit 0.1.0-alpha.0 → 1.0.0

package/package.json

@@ -1,14 +1,14 @@
 {
   "name": "reamkit",
-  "version": "0.1.0-alpha.0",
-  "description": "Ream — DOCX/XLSX to PDF, built from scratch on the ECMA-376 and ISO 32000 specifications (alpha; more document formats to come)",
+  "version": "1.0.0",
+  "description": "Ream — DOCX/XLSX to PDF (and SVG), built from scratch on the ECMA-376 and ISO 32000 specifications. Parse once, convert anywhere.",
   "license": "MIT",
   "author": "Alex Krassavin",
   "repository": {
     "type": "git",
     "url": "git+https://github.com/alex-krassavin/reamkit.git"
   },
-  "homepage": "https://github.com/alex-krassavin/reamkit#readme",
+  "homepage": "https://reamkit.dev",
   "bugs": {
     "url": "https://github.com/alex-krassavin/reamkit/issues"
   },
@@ -21,8 +21,8 @@
       "import": "./dist/esm/index.js"
     },
     "./document-model": {
-      "types": "./dist/esm/document-model/index.d.ts",
-      "import": "./dist/esm/document-model/index.js"
+      "types": "./dist/esm/core/document-model/index.d.ts",
+      "import": "./dist/esm/core/document-model/index.js"
     },
     "./package.json": "./package.json"
   },
@@ -63,7 +63,6 @@
     "lint:fix": "eslint . --fix",
     "format": "prettier --write \"src/**/*.ts\" \"tests/**/*.ts\" \"scripts/**/*.ts\"",
     "format:check": "prettier --check \"src/**/*.ts\" \"tests/**/*.ts\" \"scripts/**/*.ts\"",
-    "docs:api": "tsx scripts/docs/generate-api.ts",
     "corpus:build": "tsx scripts/corpus/build-corpus.ts",
     "corpus:fetch": "tsx scripts/corpus/fetch-corpus.ts",
     "corpus:sandbox:build": "docker build -t docgen-losandbox:latest scripts/corpus/sandbox/",

package/dist/esm/converter/docx-to-pdf.d.ts

@@ -1,15 +0,0 @@
-import { FontBytesByVariant } from '../font/index.js';
-import { FamilyKey, FetchLike } from '../fonts/index.js';
-import { SignatureOptions, StyledRenderOptions } from '../pdf/index.js';
-export interface ConvertDocxOptions extends Omit<StyledRenderOptions, 'registry' | 'styles'> {
-    readonly fontBytes?: Uint8Array;
-    readonly fonts?: FontBytesByVariant;
-    readonly fontFamily?: string;
-    readonly fontFetch?: FetchLike;
-    readonly embedSource?: boolean;
-    readonly signature?: SignatureOptions;
-}
-export declare function convertDocxToPdfSync(docx: Uint8Array, options: ConvertDocxOptions): Uint8Array;
-export declare function detectDocxFamilyKeys(docx: Uint8Array): Set<FamilyKey>;
-export declare function convertDocxToPdf(docx: Uint8Array, options?: ConvertDocxOptions): Promise<Uint8Array>;
-export declare function detectDocxFontFamily(docx: Uint8Array): string | undefined;

package/dist/esm/converter/docx-to-pdf.js

@@ -1,227 +0,0 @@
-import { FontRegistry } from "../font/font-registry.js";
-import { fetchFontSet, resolveFamilyKey } from "../fonts/remote-fonts.js";
-import { parseChart } from "../ooxml/drawingml/chart-parser.js";
-import { DEFAULT_THEME_PALETTE, makeColorResolver } from "../ooxml/drawingml/colors.js";
-import { parseTheme } from "../ooxml/drawingml/theme-parser.js";
-import { OpcPackage } from "../opc/package.js";
-import { parseCoreProperties } from "../opc/core-properties.js";
-import "../opc/index.js";
-import { EMPTY_SECTION, parseDocument, parseHeaderFooter, parseSections } from "../ooxml/wordproc/document-parser.js";
-import { EMPTY_STYLE_SHEET, parseStyles } from "../ooxml/wordproc/styles-parser.js";
-import { loadEmbeddedFonts } from "../ooxml/wordproc/font-table.js";
-import { EMPTY_NUMBERING, parseNumbering } from "../ooxml/wordproc/numbering-parser.js";
-import { EMPTY_SETTINGS, parseSettings } from "../ooxml/wordproc/settings-parser.js";
-import "../ooxml/wordproc/index.js";
-import { signPdf } from "../pdf/signature.js";
-import { renderStyledPdf } from "../pdf/styled-page-renderer.js";
-import "../pdf/index.js";
-//#region src/converter/docx-to-pdf.ts
-var STYLES_PART = "word/styles.xml";
-var NUMBERING_PART = "word/numbering.xml";
-var SETTINGS_PART = "word/settings.xml";
-var CORE_PROPS_PART = "docProps/core.xml";
-var MAIN_DOCUMENT_PART = "word/document.xml";
-var REL_HEADER = "http://schemas.openxmlformats.org/officeDocument/2006/relationships/header";
-var REL_FOOTER = "http://schemas.openxmlformats.org/officeDocument/2006/relationships/footer";
-var REL_IMAGE = "http://schemas.openxmlformats.org/officeDocument/2006/relationships/image";
-var REL_THEME = "http://schemas.openxmlformats.org/officeDocument/2006/relationships/theme";
-var THEME_PART = "word/theme/theme1.xml";
-var REL_CHART = "http://schemas.openxmlformats.org/officeDocument/2006/relationships/chart";
-function convertDocxToPdfSync(docx, options) {
-	const fonts = options.fonts ?? (options.fontBytes ? { regular: options.fontBytes } : void 0);
-	if (!fonts) throw new Error("convertDocxToPdfSync requires options.fonts or options.fontBytes");
-	const registry = FontRegistry.fromBytes(fonts);
-	const pkg = OpcPackage.open(docx);
-	const main = pkg.getMainDocument();
-	const resolveColor = buildColorResolver(pkg);
-	const body = parseDocument(main.data, resolveColor);
-	const rawSections = parseSections(main.data);
-	const stylesData = pkg.getPart(STYLES_PART);
-	const styles = stylesData ? parseStyles(stylesData) : EMPTY_STYLE_SHEET;
-	const numberingData = pkg.getPart(NUMBERING_PART);
-	const numbering = numberingData ? parseNumbering(numberingData) : EMPTY_NUMBERING;
-	const settingsData = pkg.getPart(SETTINGS_PART);
-	const settings = settingsData ? parseSettings(settingsData) : EMPTY_SETTINGS;
-	const sections = rawSections.length > 0 ? rawSections.map((s) => ({
-		...s,
-		properties: settings.evenAndOddHeaders ? {
-			...s.properties,
-			evenAndOddHeaders: true
-		} : s.properties
-	})) : [{
-		properties: settings.evenAndOddHeaders ? {
-			...EMPTY_SECTION,
-			evenAndOddHeaders: true
-		} : EMPTY_SECTION,
-		endIndex: body.length
-	}];
-	const headersFooters = loadHeadersFootersForSections(pkg, sections, resolveColor);
-	const images = loadImages(pkg);
-	const charts = loadCharts(pkg, resolveColor);
-	const embeddedFonts = loadEmbeddedFonts(pkg);
-	const coreData = pkg.getPart(CORE_PROPS_PART);
-	const coreProps = coreData ? parseCoreProperties(coreData) : void 0;
-	const { fontBytes: _ignoreA, fonts: _ignoreB, info: callerInfo, embedSource, attachments: callerAttachments, signature: _ignoreSig, ...renderOptions } = options;
-	const info = mergeInfo(coreProps, callerInfo);
-	const language = detectDocxLanguage(stylesData, main.data);
-	const attachments = [...callerAttachments ?? []];
-	if (embedSource && options.pdfA?.startsWith("PDF/A-3")) attachments.push({
-		name: "source.docx",
-		bytes: docx,
-		mimeType: "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
-		relationship: "Source",
-		description: "Source Word document"
-	});
-	return renderStyledPdf(body, {
-		registry,
-		styles,
-		numbering,
-		sections,
-		headersFooters,
-		images,
-		charts,
-		...embeddedFonts.size > 0 ? { embeddedFonts } : {},
-		...info ? { info } : {},
-		...language ? { language } : {},
-		...attachments.length > 0 ? { attachments } : {},
-		...renderOptions
-	});
-}
-function detectDocxLanguage(stylesData, documentData) {
-	const re = /<w:lang\b[^>]*\bw:val="([^"]+)"/;
-	const decoder = new TextDecoder();
-	for (const data of [stylesData, documentData]) {
-		if (!data) continue;
-		const m = re.exec(decoder.decode(data));
-		if (m?.[1]) return m[1];
-	}
-}
-function detectDocxFamilyKeys(docx) {
-	const keys = new Set(["roboto"]);
-	let pkg;
-	try {
-		pkg = OpcPackage.open(docx);
-	} catch {
-		return keys;
-	}
-	const decoder = new TextDecoder("utf-8");
-	const re = /<w:rFonts[^>]*\bw:ascii="([^"]+)"/g;
-	for (const part of [STYLES_PART, MAIN_DOCUMENT_PART]) {
-		const data = pkg.getPart(part);
-		if (!data) continue;
-		const xml = decoder.decode(data);
-		let m;
-		while ((m = re.exec(xml)) !== null) keys.add(resolveFamilyKey(m[1]));
-	}
-	return keys;
-}
-async function convertDocxToPdf(docx, options = {}) {
-	const { signature } = options;
-	const pdf = await buildUnsignedDocxPdf(docx, signature ? {
-		...options,
-		signaturePlaceholder: signature
-	} : options);
-	return signature ? signPdf(pdf, signature) : pdf;
-}
-async function buildUnsignedDocxPdf(docx, options) {
-	if (options.fonts ?? options.fontBytes) return convertDocxToPdfSync(docx, options);
-	const fetchOpt = options.fontFetch ? { fetch: options.fontFetch } : {};
-	const keys = options.fontFamily ? new Set([resolveFamilyKey(options.fontFamily)]) : detectDocxFamilyKeys(docx);
-	if (keys.size <= 1) {
-		const family = keys.values().next().value;
-		const fonts = await fetchFontSet({
-			...family ? { family } : {},
-			...fetchOpt
-		});
-		return convertDocxToPdfSync(docx, {
-			...options,
-			fonts
-		});
-	}
-	const registriesByFamily = /* @__PURE__ */ new Map();
-	let baseBytes;
-	for (const key of keys) {
-		const bytes = await fetchFontSet({
-			family: key,
-			...fetchOpt
-		});
-		registriesByFamily.set(key, FontRegistry.fromBytes(bytes));
-		if (key === "roboto" || !baseBytes) baseBytes = bytes;
-	}
-	return convertDocxToPdfSync(docx, {
-		...options,
-		fonts: baseBytes,
-		registriesByFamily
-	});
-}
-function mergeInfo(core, caller) {
-	if (!core && !caller) return void 0;
-	return {
-		...core?.title ? { title: core.title } : {},
-		...core?.creator ? { author: core.creator } : {},
-		...core?.subject ? { subject: core.subject } : {},
-		...core?.keywords ? { keywords: core.keywords } : {},
-		...core?.created ? { creationDate: core.created } : {},
-		...core?.modified ? { modificationDate: core.modified } : {},
-		...caller
-	};
-}
-function loadImages(pkg) {
-	const rels = pkg.getPartRelationships(MAIN_DOCUMENT_PART);
-	if (rels.length === 0) return /* @__PURE__ */ new Map();
-	const out = /* @__PURE__ */ new Map();
-	for (const rel of rels) {
-		if (rel.type !== REL_IMAGE) continue;
-		const resolved = pkg.resolveRelatedPart(MAIN_DOCUMENT_PART, rel);
-		if (!resolved) continue;
-		out.set(rel.id, resolved.data);
-	}
-	return out;
-}
-function loadCharts(pkg, resolveColor) {
-	const out = /* @__PURE__ */ new Map();
-	for (const rel of pkg.getPartRelationships(MAIN_DOCUMENT_PART)) {
-		if (rel.type !== REL_CHART) continue;
-		const resolved = pkg.resolveRelatedPart(MAIN_DOCUMENT_PART, rel);
-		if (!resolved) continue;
-		const chart = parseChart(resolved.data, resolveColor);
-		if (chart) out.set(rel.id, chart);
-	}
-	return out;
-}
-function buildColorResolver(pkg) {
-	const themeData = loadTheme(pkg);
-	if (!themeData) return makeColorResolver(DEFAULT_THEME_PALETTE);
-	const palette = new Map(DEFAULT_THEME_PALETTE);
-	for (const [slot, hex] of parseTheme(themeData)) palette.set(slot, hex);
-	return makeColorResolver(palette);
-}
-function loadTheme(pkg) {
-	for (const rel of pkg.getPartRelationships(MAIN_DOCUMENT_PART)) {
-		if (rel.type !== REL_THEME) continue;
-		const resolved = pkg.resolveRelatedPart(MAIN_DOCUMENT_PART, rel);
-		if (resolved) return resolved.data;
-	}
-	return pkg.getPart(THEME_PART);
-}
-function loadHeadersFootersForSections(pkg, sections, resolveColor) {
-	const wanted = /* @__PURE__ */ new Set();
-	for (const s of sections) {
-		for (const h of s.properties.headers) wanted.add(h.relationshipId);
-		for (const f of s.properties.footers) wanted.add(f.relationshipId);
-	}
-	if (wanted.size === 0) return /* @__PURE__ */ new Map();
-	const rels = pkg.getPartRelationships(MAIN_DOCUMENT_PART);
-	if (rels.length === 0) return /* @__PURE__ */ new Map();
-	const out = /* @__PURE__ */ new Map();
-	for (const rel of rels) {
-		if (!wanted.has(rel.id)) continue;
-		if (rel.type !== REL_HEADER && rel.type !== REL_FOOTER) continue;
-		const resolved = pkg.resolveRelatedPart(MAIN_DOCUMENT_PART, rel);
-		if (!resolved) continue;
-		out.set(rel.id, parseHeaderFooter(resolved.data, resolveColor));
-	}
-	return out;
-}
-//#endregion
-export { convertDocxToPdf, convertDocxToPdfSync };

package/dist/esm/converter/index.d.ts

@@ -1,4 +0,0 @@
-export { convertDocxToPdf, convertDocxToPdfSync, detectDocxFontFamily, } from './docx-to-pdf.js';
-export type { ConvertDocxOptions } from './docx-to-pdf.js';
-export { convertXlsxToPdf, convertXlsxToPdfSync } from './xlsx-to-pdf.js';
-export type { ConvertXlsxOptions } from './xlsx-to-pdf.js';

package/dist/esm/converter/index.js

@@ -1,2 +0,0 @@
-import "./docx-to-pdf.js";
-import "./xlsx-to-pdf.js";

package/dist/esm/document-model/index.d.ts

@@ -1 +0,0 @@
-export type { AbstractNumbering, Alignment, Border, BorderStyle, BodyElement, CellBorders, CellMargins, CellProperties, CellShading, Chart, ChartBlock, ChartDataPoint, ChartSeries, ChartType, CustomGeometry, CustomPathCmd, DocumentModel, FontFamilyMap, HeaderFooterReference, HeaderFooterType, ImageBlock, InlineImage, MathAccent, MathBar, MathDelimiter, MathEqArray, MathFraction, MathFunc, MathGroupChr, MathLimit, MathMatrix, MathNary, MathNode, MathRadical, MathRow, MathRun, MathScript, Numbering, NumberingFormat, NumberingInstance, NumberingLevel, NumberingReference, PageMargins, PageSize, Paragraph, ParagraphProperties, Run, RunProperties, RowProperties, Section, SectionProperties, ShapeBlock, ShapeDash, ShapeFill, ShapeFillKind, ShapeGeometry, ShapeLine, ShapeTextBody, ShapeTransform, Style, StyleSheet, StyleType, Table, TableCell, TableProperties, TableRow, UnderlineStyle, VerticalAlign, } from './types.js';

package/dist/esm/line-breaker/index.d.ts

@@ -1,2 +0,0 @@
-export { breakLines } from './knuth-plass.js';
-export type { Item, BoxItem, GlueItem, PenaltyItem, BreakResult } from './knuth-plass.js';

package/dist/esm/ooxml/index.d.ts

@@ -1 +0,0 @@
-export * as wordproc from './wordproc/index.js';

package/dist/esm/ooxml/wordproc/document-parser.d.ts

@@ -1,9 +0,0 @@
-import { BodyElement, Section, SectionProperties } from '../../document-model/index.js';
-import { ColorResolver } from '../drawingml/colors.js';
-import { PoNode } from './po-helpers.js';
-export declare function parseDocument(documentXml: Uint8Array, resolveColor?: ColorResolver): Array<BodyElement>;
-export declare const EMPTY_SECTION: SectionProperties;
-export declare function parseSection(documentXml: Uint8Array): SectionProperties;
-export declare function parseSections(documentXml: Uint8Array): Array<Section>;
-export declare function parseHeaderFooter(xml: Uint8Array, resolveColor?: ColorResolver): Array<BodyElement>;
-export declare function parseBodyElements(children: ReadonlyArray<PoNode>, resolveColor?: ColorResolver): Array<BodyElement>;

package/dist/esm/ooxml/wordproc/styles-parser.d.ts

@@ -1,3 +0,0 @@
-import { StyleSheet } from '../../document-model/index.js';
-export declare const EMPTY_STYLE_SHEET: StyleSheet;
-export declare function parseStyles(stylesXml: Uint8Array): StyleSheet;

package/dist/esm/ooxml/wordproc/table-parser.d.ts

@@ -1,3 +0,0 @@
-import { Table } from '../../document-model/index.js';
-import { PoNode } from './po-helpers.js';
-export declare function parseTable(tbl: PoNode): Table;