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 { drawEllipsePath, reduceRotation } from 'pdf-lib';
-import { isTypedArray } from 'node:util/types';
-import { error } from 'node:console';
-
+import * as p from 'typescript-parsec';
+import { Token } from 'typescript-parsec';
/**
- * debug reprensenting
+ *
+ * # REPRESENTATION
*/
-let repr = (x : any)=>{return util.inspect(x, {depth: null})};
-
/**
- * 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 = "";
-/**
- * concated 2 `tkTree`s
- * @param x the array to be concated
- * @param y the item or array to ve concated
- * @returns concated tkTree array, or thrown error if can't be concated.
- */
-function concat(x: tkTree, y:tkTree): tkTree[] {
- if (Array.isArray(x)){
- return x.concat(y);
+ if (Array.isArray(t)){
+ let strArray = t.map((x)=>tkTreeToSExp(x));
+ str = "(" + strArray.join(" ") + ")";
}else{
- throw new Error("the tkTree can't be concated, because it's not an array.");
-
+ if (t=== undefined){
+ str = "%undefined"
+ }else{
+ str = t;
+ }
}
-}
-function slice(x: tkTree, index?:number, end?:number): tkTree[] {
- if (Array.isArray(x)){
- return x.slice(index,end);
- }else{
- throw new Error("the tkTree can't be concated, because it's not an array.");
-
- }
+ return str;
}
+/**inspect the inner of the representation. */
+let repr = (x : any)=>{return util.inspect(x, {depth: null})};
/**
- * TokenMatcheePair for tokens' parser combinator
- * matched: the matched (now and before) tokens
- * remained: tokens to be matched
- * ast: abstract syntax tree
+ *
+ * # TYPES
*/
-export interface TokenMatcheePair {
- matched: tk.Token[]
- remained: tk.Token[]
- ast : tkTree[]
+
+
+type tkTree = string | tkTree[];
+
+enum TokenKind {
+ Seperator,
+ Semicolon,
+ Number,
+ Op,
+ ExprMark,
+ Paren,
+ SpaceNL,
+ Id,
+ Str,
}
/**
- * @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`.
+ * Parsing
*/
-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),
- ast: ([ttbm]),
- }
- };
- return result;
- }
- else {
- return { _tag: "None" };
- }
- }
-};
+const lexer = p.buildLexer([
+ [true, /^\d+(\.\d+)?/g, TokenKind.Number],
+ [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],
+ [false, /^(\s|\n|\r|\t)+/g, TokenKind.SpaceNL]
+]);
/**
- * type int
+ *
+ * # TEST
*/
-let tInt = m1TType(tk.TokenType.INT);
-let tAdd = m1TType(tk.TokenType.I_ADD);
-let tMul = m1TType(tk.TokenType.I_MUL);
+const inputTxt=
+`import ast;
+---
+122`;
-argv.forEach((val, index) => {
- console.log(`${index}=${val}`);
-});
+const PROG = p.rule<TokenKind, tkTree>();
+const UNIT = p.rule<TokenKind, tkTree>();
+const IMPORTS = p.rule<TokenKind, tkTree>();
+const SEMICOLON = p.rule<TokenKind, tkTree>();
-/**
- * like `m ==> f` in ocaml
- * @param m matchee wrapped
- * @param f matching function
- * @returns wrapped result
- */
-function thenDo(m : tk.Maybe<TokenMatcheePair>, f : Function){
- if (m._tag == "None"){
- return m;
- }else{
- var a : tk.Maybe<TokenMatcheePair> = f(m.value);
- if (a._tag == "Some"){
- a.value.ast = concat(m.value.ast, a.value.ast);
- }
+let doubleMinus = { type: 'Punctuator', value: '--' };
+let doubleMinus2 = p.str('--');
+const TERM = p.rule<TokenKind, tkTree>();
- return a;
- }
+function applyUnit(value: Token<TokenKind.Number>): tkTree{
+ return value.text;
}
-/**
- * like `f1 | f2` in regex
- * @param f1 the first tried function
- * @param f2 the second tried function
- * @returns wrapped result
- */
-function orDo(f1 : Function, f2 : Function){
- return (x : TokenMatcheePair) =>{
- let res1 : tk.Maybe<TokenMatcheePair> = f1(x);
- if (res1._tag == "Some"){
- return res1;
- }else{
- let res2 : tk.Maybe<TokenMatcheePair> = f2(x);
- return res2;
- }
- }
-
+function applySemiColon(value: Token<TokenKind.Semicolon>): tkTree{
+ return value.text;
}
-let midfix = (f : Function, signal? : string) => (x : TokenMatcheePair)=>{
- var a = f(x);
- if (a._tag == "Some"){
- let ast_head : tkTree[] = slice(a.value.ast,0,a.value.ast.length-3);
- let ast_tail : tkTree[] = slice(a.value.ast,a.value.ast.length-3);
- let new_ast = [ast_tail];
- a.value.ast = new_ast;
-
- console.log("+"+signal+"+"+repr(a));
-
-
- }
- return a;
+function applyParts(first: tkTree,
+ second: [Token<TokenKind>, tkTree]):tkTree {
+ return ["%clo", first , second[1]];
}
-/**
- *
- * fac1 = int MUL int
- */
-//let fac1 = midfix((x : TokenMatcheePair)=>
-// thenDo(thenDo(thenDo(tk.toSome(x), tInt), tMul), tInt));
-let fac1 = (x : TokenMatcheePair) => {
- let a = midfix((x : TokenMatcheePair)=>
- thenDo(thenDo(thenDo(tk.toSome(x), tInt), tMul), tInt), "fac1")(x);
- return a;
-}
-
-/**
- *
- * fac2 = int MUL int
- */
-let fac2 = tInt;
+function applyImports(input: [Token<TokenKind>,Token<TokenKind>[], tkTree]):tkTree{
+ let importTail = input[1].map(x=>x.text);
+ return ["import"].concat(importTail);
+};
/**
- * fac = fac1 | fac2
+ * PROG : IMPORTS '---' UNIT;
*/
-let fac = orDo(fac1, fac2);
-
+PROG.setPattern(
+ p.lrec_sc(IMPORTS, p.seq(p.str('---'), UNIT), applyParts)
+
+)
/**
- *
- * expr1 = fac ADD fac
+ * PROG : 'import' Id* SEMICOLON;
*/
-let expr1 = midfix((x : TokenMatcheePair)=>
- thenDo(thenDo(thenDo(tk.toSome(x), fac), tAdd), fac), "expr1");
+IMPORTS.setPattern(
+ p.apply(p.seq(p.str('import'), p.rep_sc(p.tok(TokenKind.Id)), SEMICOLON) , applyImports)
+);
+
/**
- * expr2 = fac
+ * SEMICOLON : ';';
*/
-let expr2 = fac;
+SEMICOLON.setPattern(
+ p.apply(p.tok(TokenKind.Semicolon), applySemiColon)
+);
/**
- * expr = expr1 | expr2
+ * UNIT : Number;
*/
-let expr = orDo(expr1, expr2);
-
-
-
-
-let tokens = tk.tokenize("2+3*4");//tk.tokenize(argv[2]);
-let tokensFiltered = tokens.filter(
- (x)=>{return (x.type != tk.TokenType.NL
- && x.type != tk.TokenType.SP)});
-
-let wrappedTokens : tk.Maybe<TokenMatcheePair> =
- tk.toSome({
- matched : [] ,
- remained : tokensFiltered,
- ast : []});
+UNIT.setPattern(
+ p.apply(p.tok(TokenKind.Number), applyUnit)
+);
-let beta = expr({
- matched : [] ,
- remained : tokensFiltered,
- ast : []});
+let tree = p.expectSingleResult(p.expectEOF(PROG.parse(lexer.parse(inputTxt))));
-console.log(repr(wrappedTokens));
-console.log(repr(beta));
+console.log("RESULT="+tkTreeToSExp(tree));
projects / clo / blobdiff