text-shaper 0.0.1

package/dist/unicode/segmentation.d.ts

@@ -0,0 +1,104 @@
+/**
+ * Unicode Text Segmentation (UAX #29)
+ * Grapheme cluster and word boundaries
+ */
+import type { GlyphInfo } from "../types.ts";
+/**
+ * Grapheme cluster break property
+ */
+export declare enum GraphemeBreakProperty {
+    Other = 0,
+    CR = 1,
+    LF = 2,
+    Control = 3,
+    Extend = 4,
+    ZWJ = 5,
+    Regional_Indicator = 6,
+    Prepend = 7,
+    SpacingMark = 8,
+    L = 9,// Hangul L
+    V = 10,// Hangul V
+    T = 11,// Hangul T
+    LV = 12,// Hangul LV
+    LVT = 13,// Hangul LVT
+    Extended_Pictographic = 14
+}
+/**
+ * Word break property
+ */
+export declare enum WordBreakProperty {
+    Other = 0,
+    CR = 1,
+    LF = 2,
+    Newline = 3,
+    Extend = 4,
+    ZWJ = 5,
+    Regional_Indicator = 6,
+    Format = 7,
+    Katakana = 8,
+    Hebrew_Letter = 9,
+    ALetter = 10,
+    Single_Quote = 11,
+    Double_Quote = 12,
+    MidNumLet = 13,
+    MidLetter = 14,
+    MidNum = 15,
+    Numeric = 16,
+    ExtendNumLet = 17,
+    WSegSpace = 18,
+    Extended_Pictographic = 19
+}
+/**
+ * Get grapheme break property for codepoint
+ */
+export declare function getGraphemeBreakProperty(cp: number): GraphemeBreakProperty;
+/**
+ * Get word break property for codepoint
+ */
+export declare function getWordBreakProperty(cp: number): WordBreakProperty;
+/**
+ * Grapheme cluster boundary result
+ */
+export interface GraphemeBoundaries {
+    /** Boundary positions (indices where clusters end) */
+    boundaries: number[];
+    /** Grapheme break properties */
+    properties: GraphemeBreakProperty[];
+}
+/**
+ * Find grapheme cluster boundaries in codepoints
+ */
+export declare function findGraphemeBoundaries(codepoints: number[]): GraphemeBoundaries;
+/**
+ * Word boundary result
+ */
+export interface WordBoundaries {
+    /** Boundary positions */
+    boundaries: number[];
+    /** Word break properties */
+    properties: WordBreakProperty[];
+}
+/**
+ * Find word boundaries in codepoints
+ */
+export declare function findWordBoundaries(codepoints: number[]): WordBoundaries;
+/**
+ * Split text into grapheme clusters
+ */
+export declare function splitGraphemes(text: string): string[];
+/**
+ * Split text into words
+ */
+export declare function splitWords(text: string): string[];
+/**
+ * Count grapheme clusters in text
+ */
+export declare function countGraphemes(text: string): number;
+/**
+ * Analyze grapheme boundaries for glyph infos
+ */
+export declare function analyzeGraphemesForGlyphs(infos: GlyphInfo[]): GraphemeBoundaries;
+/**
+ * Analyze word boundaries for glyph infos
+ */
+export declare function analyzeWordsForGlyphs(infos: GlyphInfo[]): WordBoundaries;

package/package.json

@@ -0,0 +1,33 @@
+{
+  "name": "text-shaper",
+  "version": "0.0.1",
+  "description": "Pure TypeScript text shaping engine",
+  "type": "module",
+  "main": "./dist/index.js",
+  "module": "./dist/index.js",
+  "types": "./dist/index.d.ts",
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "import": "./dist/index.js"
+    }
+  },
+  "files": ["dist"],
+  "scripts": {
+    "build": "bun build ./src/index.ts --outdir ./dist --target browser --minify --sourcemap=linked",
+    "build:prod": "bun build ./src/index.ts --outdir ./dist --target browser --production --sourcemap=external",
+    "build:analyze": "bun build ./src/index.ts --outdir ./dist --target browser --minify --metafile=./dist/meta.json",
+    "build:dts": "tsc -p tsconfig.build.json",
+    "typecheck": "tsc --noEmit",
+    "lint": "biome check --write --unsafe --max-diagnostics 99999",
+    "check": "bun run typecheck && bun run lint",
+    "prepublishOnly": "bun run build:prod && bun run build:dts"
+  },
+  "devDependencies": {
+    "@biomejs/biome": "2.3.8",
+    "@types/bun": "latest"
+  },
+  "peerDependencies": {
+    "typescript": "^5"
+  }
+}