X-Git-Url: https://git.kianting.info/?a=blobdiff_plain;f=src%2Findex.ts;h=26a690eb3b71659a119668c7e4c5f202faf04a39;hb=d4dfd2e99f564ca880d40172687986ea3ce757f0;hp=a8d103b5d550e29e63fe0e707a9b36b3b10bd090;hpb=6f2e788329da7702ea96dc28ae04499917ec8152;p=clo

diff --git a/src/index.ts b/src/index.ts
index a8d103b..26a690e 100644
--- a/src/index.ts
+++ b/src/index.ts
@@ -1,120 +1,265 @@
 var fs = require('fs');
-import { argv } from 'node:process';
-import * as tk from './tokenize.js';
+import jsTokens from "js-tokens";
 import * as util from 'util';
+import * as p from 'typescript-parsec';
+import { Token } from 'typescript-parsec';
+import { TokenType } from "./tokenize";
+/**
+ * 
+ * # REPRESENTATION
+ */
+/**
+ * convert a `tkTree` AST to S-expr string
+ * @param t the `tkTree`
+ * @returns S-expr String
+ */
+export function tkTreeToSExp(t: tkTree): string{
+    var str = "";
 
+    if (Array.isArray(t)){
+        let strArray = t.map((x)=>tkTreeToSExp(x));
+        str = "(" + strArray.join(" ") + ")";
+    }else{
+        if (t=== undefined){
+            str = "%undefined"
+        }else{
+            str = t;
+        }
+    }
+
+    return str;
+}
+
+/**inspect the inner of the representation. */
+let repr = (x : any)=>{return util.inspect(x, {depth: null})};
 /**
- * token tree type.
+ * 
+ * # TYPES
  */
-type tkTree = tk.Token[] | tk.Token
 
-export interface TokenMatcheePair {
-    matched: tkTree[]
-    remained: tk.Token[]
+
+type tkTree = string | tkTree[];
+
+enum TokenKind {
+    Seperator, // ---
+    Semicolon, // ;
+    Number,
+    Op,
+    ExprMark, // @
+    ExcapeAt, // \@
+    Paren,
+    SpaceNL, // \s\t\n\r
+    Id,
+    Str,
+    Comment, // /* ooo */
 }
 
 /**
- * @description
- * match one token type.
+ * 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],
+
+]);
+
+/**
  * 
- * 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`.
+ * # TEST
  */
-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) {
-            m.matched.push(ttbm);
-            return {
-                _tag: "Some", value: {
-                    matched: m.matched,
-                    remained: m.remained.slice(1)
-                }
-            };
-        }
-        else {
-            return { _tag: "None" };
-        }
+const inputTxt=
+`import a as b; /*bacourt*/
+/* ba choir 
+ipsum lorem*/
+
+import you as john;
+---
+
+臺中市\\\@
+
+公園
+@1+2==3;
+
+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];
+}
+
+function applySemiColon(value: Token<TokenKind.Semicolon>): tkTree{
+    return value.text;
+}
+
+function applyParts(first: tkTree,
+                    second: [Token<TokenKind>, tkTree]):tkTree {
+    return ["%clo", first , second[1]];
+}
+
+
+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;}
 };
 
-let toSome = tk.toSome;
-let thenDo = tk.thenDo;
-let orDo = tk.orDo;
+function applyText (input : tkTree): tkTree[]{
+    return ["%text", input];
+};
 
+function applyContent(input : tkTree[]): tkTree[]{
+    return ["%content", input];
+};
 
-argv.forEach((val, index) => {
-    console.log(`${index}=${val}`);
-});
+function applySpaceNL(value : Token<TokenKind.SpaceNL>): tkTree{
+    return value.text;
+}
 
-let commandInput = argv[2];
-let commandInputTokenized = tk.tokenize(commandInput);
-console.log(commandInputTokenized);
+/**
+ * IMPORTEE:  Number, Op, Paren, Id, Str, Comment,
+ */
+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),
+    );
 
 /**
- * matchee pair of commandInputTokenized
+ * PROG : IMPORTS '---' CONTENT;
  */
-let commandTPair : TokenMatcheePair = {matched:[],
-                                remained: commandInputTokenized};
+PROG.setPattern(
+    p.lrec_sc(IMPORTS, p.seq(p.str('---'), CONTENT), applyParts)
 
+)
 
-let tInt = m1TType(tk.TokenType.INT);
-let tFlo = m1TType(tk.TokenType.FLO);
-let tStr = m1TType(tk.TokenType.STR);
-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"};
-    }
-}
+/**
+ * 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)
+);
 
 /**
- * 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. 
+ * IMPORT :
+ * 'import' IMPORTEE* SEMICOLON |
+ * COMMENT |
  */
-function gramRHS (process: Function, arrange : number[]){
-    return (m : TokenMatcheePair)=>{
+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)
 
-    let result : tk.Maybe<TokenMatcheePair> = process(m);
-    console.log(`result ${result}`)
-    if (result._tag == "None"){
-        return result;
-    }
-    else{
-        let matched = result.value.matched;
-        let return_array : tkTree[] = Array(arrange.length);
+    )
+);
 
-        arrange.forEach((val, index) => {
-            return_array[arrange[index]] = matched[index];
-        });
+/**
+ * SEMICOLON : ';';
+ */
+SEMICOLON.setPattern(
+    p.apply(p.tok(TokenKind.Semicolon), applySemiColon)
+);
 
-        return return_array;
-    }
-    }
-}
+
+
+/**
+ * SEGMENT : '@' NOT_AT* '@' |
+ * (NOT_AT_TEXT | EXCAPE_AT)*
+ */
+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),
+    )
+);
 
 /**
- *  CONST ::= INT | STR | FLO | BOOL
+ * CONTENT : SEGMENT*
  */
-var constParser = gramRHS((x : TokenMatcheePair)=>
-    {return thenDo(toSome(x),orDo(orDo(orDo(tInt,tFlo),tStr),tBool))}, [0]);
+CONTENT.setPattern(
+    p.apply(p.rep(SEGMENT), applyContent)
+);
+
+
+let tree = p.expectSingleResult(p.expectEOF(PROG.parse(lexer.parse(inputTxt))));
+
+
 
-let tree = constParser(commandTPair);
-console.log(util.inspect(tree, { showHidden: true, depth: null })); 
+console.log("RESULT="+tkTreeToSExp(tree));