X-Git-Url: https://git.kianting.info/?a=blobdiff_plain;f=src%2Findex.ts;h=157b1607348020eac4b7e8648093eb77ff8059c6;hb=8d03cc503c747bb974c75d39f8b9c0678a9cc91f;hp=48d2c866b1d5f622a965eb212194f267bb18722a;hpb=ec563ca30f683cf8dcfd825c697913128f2a5445;p=clo

diff --git a/src/index.ts b/src/index.ts
index 48d2c86..157b160 100644
--- a/src/index.ts
+++ b/src/index.ts
@@ -1,223 +1,145 @@
 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';
 /**
- * token tree type.
+ * 
+ * # REPRESENTATION
  */
-type tkTree = tkTree[] | tk.Token
+/**
+ * convert a `tkTree` AST to S-expr string
+ * @param t the `tkTree`
+ * @returns S-expr String
+ */
+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;
 
+type tkTree = string | tkTree[];
 
-argv.forEach((val, index) => {
-    console.log(`${index}=${val}`);
-});
+enum TokenKind {
+    Seperator,
+    Semicolon,
+    Number,
+    Op,
+    ExprMark,
+    Paren,
+    SpaceNL,
+    Id,
+    Str,
+}
 
-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.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]
+]);
 
 /**
- * matchee pair of commandInputTokenized
+ * 
+ * # TEST
  */
-let commandTPair : TokenMatcheePair = {matched:[],
-                                remained: commandInputTokenizedFiltered};
+const inputTxt=
+`import ast;
+---
+122`;
 
 
-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);
+const PROG = p.rule<TokenKind, tkTree>();
+const UNIT = p.rule<TokenKind, tkTree>();
+const IMPORTS = p.rule<TokenKind, tkTree>();
+const SEMICOLON = p.rule<TokenKind, tkTree>();
 
 
-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"};
-    }
+let doubleMinus = { type: 'Punctuator', value: '--' };
+let doubleMinus2 = p.str('--');
+const TERM = p.rule<TokenKind, tkTree>();
+
+function applyUnit(value: Token<TokenKind.Number>): tkTree{
+    return value.text;
 }
 
-/**
- * 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;
-    }
-    }
-}
 
-/**
- * typeABS ::= "'" ID
- */
-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;
-}
 
-/**
- * TypeId ::=  typeABS | ID
- */
-var typeName =  (x : TokenMatcheePair)=>
-{
-    return thenDo(toSome(x), orDo(typeABS, tId));
-}
+function applyImports(input: [Token<TokenKind>,Token<TokenKind>[], tkTree]):tkTree{
+    let importTail = input[1].map(x=>x.text);
+    return ["import"].concat(importTail);
+};
 
 /**
- *  CONST ::= INT | STR | FLO | BOOL
+ * PROG : IMPORTS '---' UNIT;
  */
+PROG.setPattern(
+    p.lrec_sc(IMPORTS, p.seq(p.str('---'), UNIT), applyParts)
 
-/**
- * TODO: 要用 debugger 檢查分析問題
- */
-var constParser = gramRHS((x : TokenMatcheePair)=>
-    {return thenDo(toSome(x),orDo(orDo(orDo(tInt,tFlo),tStr),tBool))}, [0]);
+)
 
 /**
- * storing the tree
+ * PROG : 'import' Id* SEMICOLON;
  */
-var astTree : tkTree = [];
+IMPORTS.setPattern(
+    p.apply(p.seq(p.str('import'), p.rep_sc(p.tok(TokenKind.Id)), SEMICOLON) , applyImports)
+);
 
 /**
- * TYPE_PAIR ::= TYP_ID ID
+ * SEMICOLON : ';';
  */
-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;
-    }
-} 
+SEMICOLON.setPattern(
+    p.apply(p.tok(TokenKind.Semicolon), applySemiColon)
+);
 
 /**
- * function's arguments
- * FN_ARGS = TYPE_PAIR ("," TYPE_PAIR)+
+ * UNIT : Number;
  */
+UNIT.setPattern(
+    p.apply(p.tok(TokenKind.Number), applyUnit)
+);
 
-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));