git.kianting.info Git - clo/blob - src/parser.ts

   1 /**
   2  * parser.ts - parser and js generator of clo.
   3  */
   4 import { text } from 'pdfkit';
   5 import * as p from 'typescript-parsec';
   6 import { Token } from 'typescript-parsec';
   7
   8 /**
   9  *
  10  * # REPRESENTATION
  11  */
  12
  13 /**
  14  * convert a `tkTree` AST to S-expr string
  15  * @param t the `tkTree`
  16  * @returns S-expr String
  17  *
  18 export function tkTreeToSExp(t: tkTree): string{
  19     var str = "";
  20
  21     if (Array.isArray(t)){
  22         let strArray = t.map((x)=>tkTreeToSExp(x));
  23         str = "(" + strArray.join("◎") + ")";
  24     }else{
  25         if (t=== undefined){
  26             str = "%undefined"
  27         }else{
  28             str = t;
  29         }
  30     }
  31
  32     return str;
  33 }*/
  34
  35 export type tkTree = string | tkTree[];
  36
  37 export enum TokenKind {
  38     Seperator, // ---
  39     Semicolon, // ;
  40     Number,
  41     Op,
  42     ExprMark, // @
  43     ExcapeAt, // \@
  44     Paren,
  45     SpaceNL, // \s\t\n\r
  46     Id,
  47     Str,
  48     Comment, // /* ooo */
  49 }
  50
  51 /**
  52  * Parsing
  53  */
  54 export const lexer = p.buildLexer([
  55     [true, /^\d+(\.\d+)?/g, TokenKind.Number],
  56     [true, /^[\\][\\]/g, TokenKind.Op],
  57     [true, /^\\\@/g, TokenKind.ExcapeAt],
  58     [true, /^\/\*([^/]|\/[^*])*\*\//g, TokenKind.Comment],
  59     [true, /^\;/g, TokenKind.Semicolon],
  60     [true, /^[-][-][-]/g, TokenKind.Seperator],
  61     [true, /^[\+\-\*\/\&\|\!\^\<\>\~\=\?]+/g, TokenKind.Op],
  62     [true, /^\@/g, TokenKind.ExprMark],
  63     [true, /^[()\[\]{}]/g, TokenKind.Paren],
  64     [true, /^[\`]([^\`]|[\\].)*[\`]/g, TokenKind.Str],
  65     [true, /^[\"]([^\"]|[\\].)*[\"]/g, TokenKind.Str],
  66     [true, /^[\']([^\']|[\\].)*[\']/g, TokenKind.Str],
  67     [true, /^[()\[\]{}]/g, TokenKind.Paren],
  68     [true, /^[^\/\\\@\s\n\t\r;]+/g, TokenKind.Id],
  69     [true, /^(\s|\n|\r|\t)+/g, TokenKind.SpaceNL],
  70
  71 ]);
  72
  73 /**
  74  *
  75  * # TEST
  76  */
  77
  78
  79
  80 export const PROG = p.rule<TokenKind, tkTree>();
  81 export const SEGMENT = p.rule<TokenKind, tkTree>();
  82 export const IMPORT = p.rule<TokenKind, tkTree>();
  83 export const IMPORTS = p.rule<TokenKind, tkTree>();
  84 export const SEMICOLON = p.rule<TokenKind, tkTree>();
  85 export const NOT_AT_TEXT = p.rule<TokenKind, tkTree>();
  86 export const CONTENT = p.rule<TokenKind, tkTree>();
  87
  88
  89 export function applySegment(input: [Token<TokenKind>, Token<TokenKind>[],
  90         Token<TokenKind>]): tkTree[]{
  91     let unpackedInnerExprs = input[1].map((x)=>{return x.text});
  92     return ["%exprs", unpackedInnerExprs];
  93 }
  94
  95 export function applySemiColon(value: Token<TokenKind.Semicolon>): tkTree{
  96     return value.text;
  97 }
  98
  99 export function applyParts(first: tkTree,
 100                     second: [Token<TokenKind>, Token<TokenKind>, tkTree]):tkTree {
 101     return ["%clo", first , second[2]];
 102 }
 103
 104 export function applyPartsWithoutImport(
 105         parsed: [Token<TokenKind>, Token<TokenKind>, tkTree]):tkTree {
 106 return ["%clo", "" , parsed[2]];
 107 }
 108
 109
 110 export function applyComment(value: Token<TokenKind.Comment>): tkTree[]{
 111     return [value.text];
 112 }
 113
 114
 115 export function applyImport(input: [Token<TokenKind>,Token<TokenKind>[], tkTree]) : tkTree{
 116     let importTail = input[1].map(x=>x.text);
 117     return ["import"].concat(importTail);
 118 };
 119
 120
 121 /*
 122 function applyImportComment(input: [Token<TokenKind>,Token<TokenKind>[],
 123     tkTree, Token<TokenKind.Comment>]) : tkTree{
 124     let importTail = input[1].map(x=>x.text);
 125     let comment = [input[3].text];
 126     return ["import"].concat(importTail).concat(comment);
 127 };*/
 128
 129 export function applyImports(input : [tkTree, tkTree[]]): tkTree{
 130     let resultBody = [input[0]].concat(input[1]);
 131     let resultWrapper = ["%import", resultBody];
 132     return resultWrapper;
 133 };
 134
 135
 136
 137
 138 export function applyNotAtText(value : Token<TokenKind>): tkTree{
 139     if (value.text == "\\\@"){
 140         return '@';
 141     }
 142     else{return value.text;}
 143 };
 144
 145 export function applyText (input : tkTree): tkTree[]{
 146
 147     return ["%text", input];
 148 };
 149
 150
 151
 152 export function applyContent(input : tkTree[]): tkTree[]{
 153     return ["%content", input];
 154 };
 155
 156 export function applySpaceNL(value : Token<TokenKind.SpaceNL>): tkTree{
 157     return value.text;
 158 }
 159
 160 /**
 161  * IMPORTEE:  Number, Op, Paren, Id, Str, Comment,
 162  */
 163 export let IMPORTEE = p.alt(p.tok(TokenKind.Number),
 164     p.tok(TokenKind.Op),
 165     p.tok(TokenKind.Paren),
 166     p.tok(TokenKind.Id),
 167     p.tok(TokenKind.Str),
 168     p.tok(TokenKind.SpaceNL),
 169     p.tok(TokenKind.Comment));
 170
 171 export let NOT_AT = p.alt(p.tok(TokenKind.Seperator),
 172     p.tok(TokenKind.Semicolon),
 173     p.tok(TokenKind.Number),
 174     p.tok(TokenKind.ExcapeAt),
 175     p.tok(TokenKind.Op),
 176     p.tok(TokenKind.Paren),
 177     p.tok(TokenKind.SpaceNL),
 178     p.tok(TokenKind.Id),
 179     p.tok(TokenKind.Str),
 180     p.tok(TokenKind.Comment),
 181     );
 182
 183 /**
 184  * PROG : IMPORTS '---' NEWLINE CONTENT | '---' NEWLINE CONTNENT
 185  */
 186 PROG.setPattern(
 187     p.alt(
 188         p.lrec_sc(IMPORTS, p.seq(p.str('---'), p.str("\n"), CONTENT), applyParts),
 189         p.apply(p.seq(p.str('---'), p.str("\n"), CONTENT), applyPartsWithoutImport))
 190
 191 )
 192
 193 /**
 194  * NOT_AT_TEXT : NOT_AT
 195  */
 196 NOT_AT_TEXT.setPattern(
 197     p.apply(NOT_AT, applyNotAtText)
 198 );
 199
 200 IMPORTS.setPattern(
 201     p.apply( p.seq(IMPORT, p.rep(IMPORT)), applyImports)
 202 );
 203
 204 /**
 205  * IMPORT :
 206  * 'import' IMPORTEE* SEMICOLON |
 207  * COMMENT |
 208  */
 209 IMPORT.setPattern(
 210     p.alt(
 211         p.apply(p.seq(p.str('import'), p.rep_sc(IMPORTEE), SEMICOLON),
 212             applyImport),
 213         p.apply(p.tok(TokenKind.Comment), applyComment),
 214         p.apply(p.tok(TokenKind.SpaceNL), applySpaceNL)
 215
 216     )
 217 );
 218
 219 /**
 220  * SEMICOLON : ';';
 221  */
 222 SEMICOLON.setPattern(
 223     p.apply(p.tok(TokenKind.Semicolon), applySemiColon)
 224 );
 225
 226
 227
 228 /**
 229  * SEGMENT : '@' NOT_AT* '@' |
 230  * (NOT_AT_TEXT | EXCAPE_AT)*
 231  */
 232 SEGMENT.setPattern(
 233     p.alt(
 234         p.apply(p.rep_sc(NOT_AT_TEXT), applyText),
 235         p.apply(p.seq(p.str('@'), p.rep(NOT_AT), p.str('@')), applySegment),
 236     )
 237 );
 238
 239 /**
 240  * CONTENT : SEGMENT*
 241  */
 242 CONTENT.setPattern(
 243     p.apply(p.rep(SEGMENT), applyContent)
 244 );
 245
 246
 247
 248 /**
 249  * the head part of the output JS code : before import
 250  */
 251 export let outputHead = `
 252 /* clo, a typesetting engine, generated JS file*/
 253 /* CLO:  beginning of head*/
 254
 255 let cloLib = require("./src/libclo/index.js");
 256 let clo = new cloLib.Clo();
 257
 258 /* CLO:  end of head*/\n`
 259
 260 /**
 261  * the middle part of the output JS code : between import part and content part
 262  */
 263 export let outputMiddle =`
 264 /* CLO:  beginning of middle part*/
 265 clo.mainStream = /* CLO: end of middle part*/
 266 `
 267
 268 /**
 269  * the end part of the output JS code : after content part
 270  */
 271 export let outputEnd =`
 272 /* CLO: beginning of end part*/
 273 clo.generatePdf();
 274 /*CLO : end of end part*/
 275 `
 276
 277 export function splitText(input : tkTree): tkTree{
 278     var ret;
 279     if (!Array.isArray(input)){
 280         ret = input.split(/(\s+)/);
 281     }else{
 282         ret = input.map((x)=>splitText(x));
 283     }
 284     return ret;
 285 }
 286
 287 /**
 288  * Convert `tree` (ASTTree; `tkTree`) to JS Code.
 289  */
 290 export function treeToJS(tree : tkTree): string{
 291
 292     let head = tree[0];
 293     if (head == "%clo"){
 294         let totalResult = outputHead + treeToJS(tree[1]) +
 295             outputMiddle + treeToJS(tree[2]) + outputEnd;
 296         return totalResult;
 297     }
 298     if (head == "%import"){
 299         let imports = tree[1];
 300         if (Array.isArray(imports)){
 301             let importsText = imports.map(
 302                 (x)=>{
 303                     if (Array.isArray(x)){
 304                         return x.join('') + ';';
 305                     }
 306                     else{
 307                         return x;
 308                     }
 309                 });
 310             let importTextCombined = importsText.join('');
 311             return importTextCombined;
 312         }
 313         else{
 314             return imports;
 315         }
 316     }
 317     if (head == "%content"){
 318         let tail = tree[1];
 319         if (Array.isArray(tail)){
 320             if (tail.length == 1){
 321                 return tail.map((x)=>treeToJS(x)).join("').concat('")+ ";";
 322             }
 323             let tailStrings = tail.map((x)=>treeToJS(x));
 324             return "(" + tailStrings.join(').concat(') + ");";
 325         }else{
 326             return tail;
 327         }
 328     }
 329     if (head == "%text"){
 330         var textContents = splitText(tree[1]);
 331         if (Array.isArray(textContents)){
 332             textContents = textContents.flat().filter((x)=>{return x !== ""});
 333             let decoratedArray = textContents
 334                                 .flatMap(x=>String(x))
 335                                 .map(x=>x.replace("\'",  "\\\'"))
 336                                 .map(x=>x.replace("\`","\\\`"));
 337
 338             return "[`" + decoratedArray.join("\`, \`") + "`]";
 339         }else{
 340             let decorated = textContents.replace("\`","\\\`").replace("\'", "\\\'");
 341
 342             return "[`" + decorated + "`]";
 343         }
 344     }
 345
 346     if (head == "%exprs"){
 347         let content = tree[1];
 348         if (Array.isArray(content)){
 349             let flattenContent = content.flat();
 350             return flattenContent.join('');
 351         }
 352         else{
 353             return content;
 354         }
 355
 356     }
 357     else{
 358         if (Array.isArray(tree)){
 359             return tree.join('');
 360         }else{
 361             return tree;
 362         }
 363     }
 364 }
 365
 366
 367 /**
 368  * `inputText` to `tkTree` (ASTTree)
 369  */
 370 export function inputTextToTree(inputText : string){
 371     // force convert Windows newline to Linux newline
 372     inputText = inputText.replace("\r\n", "\n");
 373
 374     return p.expectSingleResult(
 375         p.expectEOF(PROG.parse(lexer.parse(inputText))));
 376 }