From 796e0c20c767e214b104fa3d17e754ce58da6e73 Mon Sep 17 00:00:00 2001 From: Tan Kian-ting Date: Fri, 27 Oct 2023 23:51:43 +0800 Subject: [PATCH] add cjk splitter --- README.md | 2 + src/libclo/index.js | 96 +++++++++++++++++++++++++++--- src/libclo/index.ts | 138 +++++++++++++++++++++++++++++++++++++++++--- 3 files changed, 220 insertions(+), 16 deletions(-) diff --git a/README.md b/README.md index 4ae71cf..46dfc11 100644 --- a/README.md +++ b/README.md @@ -29,6 +29,8 @@ License: MIT - 20231012: clo->js converter successfully (maybe.) - 20231016：basic font guessing and `putText` function - 20231023-24:fix .ttc bug. + - 20231026-27 : clo basic interface, preprocessor of stream of text, + add cjk-english splitter, etc. ## 之後的做法 - 先做一個前處理註冊器，註冊下列的前處理 diff --git a/src/libclo/index.js b/src/libclo/index.js index 5bcc3f1..e5ae1bd 100644 --- a/src/libclo/index.js +++ b/src/libclo/index.js @@ -1,20 +1,102 @@ "use strict"; Object.defineProperty(exports, "__esModule", { value: true }); exports.a = exports.Clo = void 0; -function foo(arr) { +const canva_1 = require("../canva"); +/** + * TYPES + */ +/** + * text direction + * LTR - left to right + * TTB - top to bottom + * etc. + */ +var Direction; +(function (Direction) { + Direction[Direction["LTR"] = 0] = "LTR"; + Direction[Direction["RTL"] = 1] = "RTL"; + Direction[Direction["TTB"] = 2] = "TTB"; + Direction[Direction["BTT"] = 3] = "BTT"; +})(Direction || (Direction = {})); +/** + * DEFAULT CONST PART + */ +const A4_IN_PX = { "width": 793.7, + "height": 1122.5 }; +const defaultTextStyle = { + family: "FreeSans", + size: 12, + textWeight: canva_1.TextWeight.REGULAR, + textStyle: canva_1.TextStyle.ITALIC, +}; +const defaultFrameStyle = { + directionInsideLine: Direction.LTR, + direction: Direction.TTB, + baseLineskip: ptToPx(15), + fontStyle: defaultTextStyle, + x: A4_IN_PX.width * 0.10, + y: A4_IN_PX.height * 0.10, + width: A4_IN_PX.width * 0.80, + height: A4_IN_PX.height * 0.80, + content: null, +}; +const cjkvBlocksInRegex = ["Hani"]; +const cjkvRegexPattern = new RegExp("((?:" + + cjkvBlocksInRegex.map((x) => "\\p{Script_Extensions=" + x + "}").join("|") + ")+)", "gu"); +/** + * FUNCTION PART + */ +/** + * convert from ptToPx + * @param pt pt size value + * @returns the corresponding px value + */ +function ptToPx(pt) { + return pt * 4 / 3.0; +} +/** + * REGISTER PART + */ +/** + * split CJKV and non-CJKV + * + * @param arr : input tkTree + * @returns + */ +function splitCJKV(arr) { + console.log(arr); + var result = []; for (let i = 0; i < arr.length; i++) { + var item = arr[i]; + if (!Array.isArray(item)) { + console.log(item.split(cjkvRegexPattern)); + result = result.concat(item.split(cjkvRegexPattern)); + } + else { + result.push(item); + } } - if (Array.isArray(arr)) { - arr.push("balabala"); - } - return arr; + console.log(result); + return result; } class Clo { constructor() { this.preprocessors = []; this.mainStream = []; - this.attributes = { "page": [793.7, 1122.5] }; - this.preprocessorRegister(foo); + this.attributes = { "page": A4_IN_PX }; + // register the precessor functions + this.preprocessorRegister(splitCJKV); + } + setAttr(attr, val) { + Object.assign(this.attributes, attr, val); + } + getAttr(attr) { + if (Object.keys(this.attributes).length === 0) { + return this.attributes[attr]; + } + else { + return undefined; + } } /** * register a function of preprocessor diff --git a/src/libclo/index.ts b/src/libclo/index.ts index b83b050..7fd58c6 100644 --- a/src/libclo/index.ts +++ b/src/libclo/index.ts @@ -1,27 +1,147 @@ +import { isKeyObject, isStringObject } from "util/types"; import {tkTree} from "../parser"; +import {TextStyle, FontStyle, TextWeight} from "../canva"; +import { isString } from "util"; +/** + * TYPES + */ -function foo(arr : tkTree): tkTree{ +/** + * text direction + * LTR - left to right + * TTB - top to bottom + * etc. + */ +enum Direction{ + LTR, + RTL, + TTB, + BTT, +} + +/** + * frame box is a subclass of box + * - directionInsideLine : text direction inside a line + * - baselineskip : the distance between baselines in px + */ +interface FrameBox extends Box{ + directionInsideLine : Direction, + baseLineskip : number | null, +} + +/** + * a basic Box + */ +interface Box{ + x : number | null, + y : number | null, + fontStyle : FontStyle | null, + direction : Direction, + width : number, + height : number, + content : string | Box[] | null, +} + + +/** + * DEFAULT CONST PART + */ +const A4_IN_PX = {"width" : 793.7, + "height" : 1122.5}; + +const defaultTextStyle : FontStyle = { + family : "FreeSans", + size : 12, + textWeight : TextWeight.REGULAR, + textStyle : TextStyle.ITALIC, +} + +const defaultFrameStyle : FrameBox = { + directionInsideLine : Direction.LTR, + direction : Direction.TTB, + baseLineskip : ptToPx(15), + fontStyle : defaultTextStyle, + x : A4_IN_PX.width * 0.10, + y : A4_IN_PX.height * 0.10, + width : A4_IN_PX.width * 0.80, + height : A4_IN_PX.height * 0.80, + content : null, +}; + +const cjkvBlocksInRegex = ["Hani"]; + +const cjkvRegexPattern = new RegExp("((?:" + + cjkvBlocksInRegex.map((x)=>"\\p{Script_Extensions="+x+"}").join("|") + ")+)", "gu"); +/** + * FUNCTION PART + */ +/** + * convert from ptToPx + * @param pt pt size value + * @returns the corresponding px value + */ +function ptToPx(pt : number) : number{ + return pt * 4 / 3.0; +} + + + +/** + * REGISTER PART + */ + + +/** + * split CJKV and non-CJKV + * + * @param arr : input tkTree + * @returns + */ +function splitCJKV(arr : tkTree): tkTree{ + var result : tkTree = []; for (let i = 0; i < arr.length; i++) { - - } - if (Array.isArray(arr)){ - arr.push("balabala"); + var item = arr[i]; + if (!Array.isArray(item)){ + console.log(item.split(cjkvRegexPattern)); + result = result.concat(item.split(cjkvRegexPattern)); + } + else{ + result.push(item); + } } - return arr; + + return result; } export class Clo{ mainStream : Array; preprocessors : Array; - attributes: object ; // a4 size(x,y) + attributes: {[index: string]:any} ; // a4 size(x,y) constructor(){ this.preprocessors = []; this.mainStream = []; - this.attributes = {"page" : [793.7, 1122.5]}; - this.preprocessorRegister(foo); + this.attributes = {"page" : A4_IN_PX}; + + + + // register the precessor functions + this.preprocessorRegister(splitCJKV); + } + + public setAttr(attr : string, val : any):void{ + Object.assign(this.attributes, attr, val); + } + + public getAttr(attr:string) : any{ + if (Object.keys(this.attributes).length === 0){ + return this.attributes[attr]; + }else{ + return undefined; + } + } /** -- 2.39.2