var fs = require('fs');
-import { argv, resourceUsage } from 'node:process';
-import * as tk from './tokenize.js';
+import jsTokens from "js-tokens";
import * as util from 'util';
-import { reduceRotation } from 'pdf-lib';
-
+import * as p from 'typescript-parsec';
+import { Token } from 'typescript-parsec';
+import { TokenType } from "./tokenize";
+/**
+ *
+ * # REPRESENTATION
+ */
/**
- * token tree type.
+ * convert a `tkTree` AST to S-expr string
+ * @param t the `tkTree`
+ * @returns S-expr String
*/
-type tkTree = tkTree[] | tk.Token
+export function tkTreeToSExp(t: tkTree): string{
+ var str = "";
-export interface TokenMatcheePair {
- matched: tkTree[]
- remained: tk.Token[]
-}
+ if (Array.isArray(t)){
+ let strArray = t.map((x)=>tkTreeToSExp(x));
+ str = "(" + strArray.join(" ") + ")";
+ }else{
+ if (t=== undefined){
+ str = "%undefined"
+ }else{
+ str = t;
+ }
+ }
-export interface MaybeTokensAST{
- maybeTokens: tk.Maybe<TokenMatcheePair>;
- ast: tkTree;
+ return str;
}
+/**inspect the inner of the representation. */
+let repr = (x : any)=>{return util.inspect(x, {depth: null})};
/**
- * @description
- * match one token type.
*
- * it returns a function which test if the type of first token of the `remained` part of
- * the argument of the function is `typ` , if it's true, update the `TokenMatcheePair` wrapped
- * in `Some`. Otherwise, it returns `None`.
- * * @param typ : the type to be test.
- * @returns the updated `TokenMatcheePair` wrapped in `Some(x)` or `None`.
+ * # TYPES
*/
-export function m1TType(typ: tk.TokenType):
- (m: TokenMatcheePair) => tk.Maybe<TokenMatcheePair> {
- return (m: TokenMatcheePair) => {
- if (m.remained.length == 0) {
- return { _tag: "None" };
- }
- /**
- * token to be matched
- * */
- const ttbm = m.remained[0];
-
- if (ttbm.type == typ) {
- let new_matched = m.matched.concat(ttbm);
- let result : tk.Some<TokenMatcheePair> = {
- _tag: "Some", value: {
- matched: new_matched,
- remained: m.remained.slice(1)
- }
- };
- return result;
- }
- else {
- return { _tag: "None" };
- }
- }
-};
-
-let toSome = tk.toSome;
-let thenDo = tk.thenDo;
-let zeroOrOnceDo = tk.zeroOrOnceDo;
-let orDo = tk.orDo;
-let zeroOrMoreDo = tk.zeroOrMoreDo;
-argv.forEach((val, index) => {
- console.log(`${index}=${val}`);
-});
+type tkTree = string | tkTree[];
+
+enum TokenKind {
+ Seperator, // ---
+ Semicolon, // ;
+ Number,
+ Op,
+ ExprMark, // @
+ ExcapeAt, // \@
+ Paren,
+ SpaceNL, // \s\t\n\r
+ Id,
+ Str,
+ Comment, // /* ooo */
+}
-let commandInput = "int a str b"//argv[2];
-let commandInputTokenized = tk.tokenize(commandInput);
-let commandInputTokenizedFiltered = commandInputTokenized.filter(
- (x : tk.Token)=>{return x.type != tk.TokenType.SP &&
- x.type != tk.TokenType.NL});
-console.log("aaa: "+util.inspect(commandInputTokenizedFiltered, { showHidden: true, depth: null }));
+/**
+ * Parsing
+ */
+const lexer = p.buildLexer([
+ [true, /^\d+(\.\d+)?/g, TokenKind.Number],
+ [true, /^\\\@/g, TokenKind.ExcapeAt],
+ [true, /^\/\*([^/]|\/[^*])*\*\//g, TokenKind.Comment],
+ [true, /^\;/g, TokenKind.Semicolon],
+ [true, /^[-][-][-]/g, TokenKind.Seperator],
+ [true, /^[\+\-\*\/\&\|\!\^\<\>\~\=\?]+/g, TokenKind.Op],
+ [true, /^\@/g, TokenKind.ExprMark],
+ [true, /^[()\[\]{}]/g, TokenKind.Paren],
+ [true, /^["]([\"]|[\\].)*["]/g, TokenKind.Str],
+ [true, /^[']([\']|[\\].)*[']/g, TokenKind.Str],
+ [true, /^[()\[\]{}]/g, TokenKind.Paren],
+ [true, /^[^\/\\\@\s\n\t\r;]+/g, TokenKind.Id],
+ [true, /^(\s|\n|\r|\t)+/g, TokenKind.SpaceNL],
+
+]);
/**
- * matchee pair of commandInputTokenized
+ *
+ * # TEST
*/
-let commandTPair : TokenMatcheePair = {matched:[],
- remained: commandInputTokenizedFiltered};
+const inputTxt=
+`import a as b; /*bacourt*/
+/* ba choir
+ipsum lorem*/
+import you as john;
+---
-let tInt = m1TType(tk.TokenType.INT);
-let tFlo = m1TType(tk.TokenType.FLO);
-let tStr = m1TType(tk.TokenType.STR);
-let tId = m1TType(tk.TokenType.ID);
-let tApos = m1TType(tk.TokenType.APOS);
+臺中市\\\@
+公園
+@1+2==3;
-function tBool (x : TokenMatcheePair) :tk.Maybe<TokenMatcheePair> {
- let text = x.remained[0].text
- if (text == "true" || text == "false"){
- return thenDo(toSome(x), m1TType(tk.TokenType.ID));
- }else{
- return {_tag : "None"};
- }
+console.log("122");@
+
+山頂
+`;
+
+
+const PROG = p.rule<TokenKind, tkTree>();
+const SEGMENT = p.rule<TokenKind, tkTree>();
+const IMPORT = p.rule<TokenKind, tkTree>();
+const IMPORTS = p.rule<TokenKind, tkTree>();
+const SEMICOLON = p.rule<TokenKind, tkTree>();
+const EXCAPE_AT = p.rule<TokenKind, tkTree>();
+const NOT_AT_TEXT = p.rule<TokenKind, tkTree>();
+const CONTENT = p.rule<TokenKind, tkTree>();
+
+let doubleMinus = { type: 'Punctuator', value: '--' };
+let doubleMinus2 = p.str('--');
+const TERM = p.rule<TokenKind, tkTree>();
+
+function applySegment(input: [Token<TokenKind>, Token<TokenKind>[],
+ Token<TokenKind>]): tkTree[]{
+ let unpackedInnerExprs = input[1].map((x)=>{return x.text});
+ return ["%exprs", unpackedInnerExprs];
}
-/**
- * define the right hand side of a grammar
- * eg. `LHS ::= a + b`
- * @param process the right hand side processing : eg. `a + b` in `LHS`
- * @param arrange define the order (0 starting) of the elements of the result.
- * ast. : eg. `a + c` is `1 0 2` `(+ a c)`
- * @returns the processed ast.
- */
-function gramRHS (process: Function, arrange : number[]){
- return (m : TokenMatcheePair)=>{
+function applySemiColon(value: Token<TokenKind.Semicolon>): tkTree{
+ return value.text;
+}
- let middle : tk.Maybe<TokenMatcheePair> = process(m);
+function applyParts(first: tkTree,
+ second: [Token<TokenKind>, tkTree]):tkTree {
+ return ["%clo", first , second[1]];
+}
- console.log("Middle"+util.inspect(middle, { showHidden: true, depth: null }));
- if (middle._tag == "None"){
- return middle;
- }
- else{
- let matched = middle.value.matched;
- let arrLength = arrange.length;
- let returnRrray : tkTree[] = Array(arrange.length);
-
- arrange.forEach((val, index) => {
- returnRrray[arrange[index]] = matched[index];
- });
-
- let matchedTmp1Length = matched.length-arrLength;
- console.log(matchedTmp1Length);
- var matchedTmp1 : tkTree[] = matched
- .slice(0,matchedTmp1Length);
-
- console.log("matchedTmp1"+util.inspect(matchedTmp1, { showHidden: true, depth: null }));
- console.log("returnRrray"+util.inspect(returnRrray, { showHidden: true, depth: null }));
- matchedTmp1.push(returnRrray);
-
-
- let result : tk.Some<TokenMatcheePair> = {_tag:"Some",
- value : {matched : matchedTmp1,
- remained : middle.value.remained}};
- return result;
- }
+function applyComment(value: Token<TokenKind.Comment>): tkTree[]{
+ return [value.text];
+}
+
+
+function applyImport(input: [Token<TokenKind>,Token<TokenKind>[], tkTree]) : tkTree{
+ let importTail = input[1].map(x=>x.text);
+ return ["import"].concat(importTail);
+};
+
+
+/*
+function applyImportComment(input: [Token<TokenKind>,Token<TokenKind>[],
+ tkTree, Token<TokenKind.Comment>]) : tkTree{
+ let importTail = input[1].map(x=>x.text);
+ let comment = [input[3].text];
+ return ["import"].concat(importTail).concat(comment);
+};*/
+
+function applyImports(input : [tkTree, tkTree[]]): tkTree{
+ let resultBody = [input[0]].concat(input[1]);
+ let resultWrapper = ["%import", resultBody];
+ return resultWrapper;
+};
+
+
+
+
+function applyNotAtText(value : Token<TokenKind>): tkTree{
+ if (value.text == "\\\@"){
+ return '@';
}
+ else{return value.text;}
+};
+
+function applyText (input : tkTree): tkTree[]{
+ return ["%text", input];
+};
+
+function applyContent(input : tkTree[]): tkTree[]{
+ return ["%content", input];
+};
+
+function applySpaceNL(value : Token<TokenKind.SpaceNL>): tkTree{
+ return value.text;
}
/**
- * typeABS ::= "'" ID
+ * IMPORTEE: Number, Op, Paren, Id, Str, Comment,
*/
-var typeABS = (x : TokenMatcheePair)=>
-{
- var result = thenDo(thenDo(toSome(x),tApos),tId);
- if (result._tag == "Some" && "text" in result.value.matched[1]){
- var realToken : tk.Token = result.value.matched[1];
- realToken.text = "'"+realToken.text;
- result.value.matched = [realToken];
- }
- return result;
-}
+let IMPORTEE = p.alt(p.tok(TokenKind.Number),
+ p.tok(TokenKind.Op),
+ p.tok(TokenKind.Paren),
+ p.tok(TokenKind.Id),
+ p.tok(TokenKind.Str),
+ p.tok(TokenKind.SpaceNL),
+ p.tok(TokenKind.Comment));
+
+let NOT_AT = p.alt(p.tok(TokenKind.Seperator),
+ p.tok(TokenKind.Semicolon),
+ p.tok(TokenKind.Number),
+ p.tok(TokenKind.ExcapeAt),
+ p.tok(TokenKind.Op),
+ p.tok(TokenKind.Paren),
+ p.tok(TokenKind.SpaceNL),
+ p.tok(TokenKind.Id),
+ p.tok(TokenKind.Str),
+ p.tok(TokenKind.Comment),
+ );
/**
- * TypeId ::= typeABS | ID
+ * PROG : IMPORTS '---' CONTENT;
*/
-var typeName = (x : TokenMatcheePair)=>
-{
- return thenDo(toSome(x), orDo(typeABS, tId));
-}
+PROG.setPattern(
+ p.lrec_sc(IMPORTS, p.seq(p.str('---'), CONTENT), applyParts)
+
+)
/**
- * CONST ::= INT | STR | FLO | BOOL
+ * NOT_AT_TEXT : NOT_AT
*/
+NOT_AT_TEXT.setPattern(
+ p.apply(NOT_AT, applyNotAtText)
+);
+
+IMPORTS.setPattern(
+ p.apply( p.seq(IMPORT, p.rep(IMPORT)), applyImports)
+);
/**
- * TODO: 要用 debugger 檢查分析問題
+ * IMPORT :
+ * 'import' IMPORTEE* SEMICOLON |
+ * COMMENT |
*/
-var constParser = gramRHS((x : TokenMatcheePair)=>
- {return thenDo(toSome(x),orDo(orDo(orDo(tInt,tFlo),tStr),tBool))}, [0]);
+IMPORT.setPattern(
+ p.alt(
+ p.apply(p.seq(p.str('import'), p.rep_sc(IMPORTEE), SEMICOLON),
+ applyImport),
+ p.apply(p.tok(TokenKind.Comment), applyComment),
+ p.apply(p.tok(TokenKind.SpaceNL), applySpaceNL)
+
+ )
+);
/**
- * storing the tree
+ * SEMICOLON : ';';
*/
-var astTree : tkTree = [];
+SEMICOLON.setPattern(
+ p.apply(p.tok(TokenKind.Semicolon), applySemiColon)
+);
+
+
/**
- * TYPE_PAIR ::= TYP_ID ID
+ * SEGMENT : '@' NOT_AT* '@' |
+ * (NOT_AT_TEXT | EXCAPE_AT)*
*/
-var typePair = (x : MaybeTokensAST)=>
-{
-
-
- let a = thenDo(thenDo(x.maybeTokens, typeName), tId);
- if (a._tag == "Some"){
- let matched = a.value.matched;
- let slice = matched.slice(matched.length-2);
- console.log("slice"+slice);
-
- let b : MaybeTokensAST = {maybeTokens : a, ast : slice};
- return b;
- }
- else{
- let b : MaybeTokensAST= {maybeTokens : a, ast : []};
- return b;
- }
-}
+SEGMENT.setPattern(
+ p.alt(
+ p.apply(p.rep_sc(NOT_AT_TEXT), applyText),
+ p.apply(p.seq(p.str('@'), p.rep(NOT_AT), p.str('@')), applySegment),
+ )
+);
/**
- * function's arguments
- * FN_ARGS = TYPE_PAIR ("," TYPE_PAIR)+
+ * CONTENT : SEGMENT*
*/
+CONTENT.setPattern(
+ p.apply(p.rep(SEGMENT), applyContent)
+);
+
-var fnArgs = (x : TokenMatcheePair)=>
- {
- let wrapper : MaybeTokensAST = {maybeTokens : toSome(x), ast : []};
- let a = typePair(wrapper);
- console.log("AAAAA"+util.inspect(a, { showHidden: true, depth: null }));
- let abanibi = typePair(a);
- console.log("ABNB"+util.inspect(abanibi, { showHidden: true, depth: null }));
+let tree = p.expectSingleResult(p.expectEOF(PROG.parse(lexer.parse(inputTxt))));
- return {maybeTokens : abanibi.maybeTokens, ast : [a.ast, abanibi.ast]};
-
- };
-let tree = fnArgs(commandTPair);
-console.log("CHRANN"+util.inspect(tree, { showHidden: true, depth: null }));
+console.log("RESULT="+tkTreeToSExp(tree));