rewrite parser

2023-10-06 01:09:28 +08:00 · 2023-10-06 01:09:28 +08:00 · 8d03cc503c
commit 8d03cc503c
parent fed01e9044
5 changed files with 167 additions and 266 deletions
--- a/README.md
+++ b/README.md
@ -23,4 +23,5 @@ License: MIT
    ```
 - 20230928：basically fix `issue1`。其他ê物件猶著做。
 - 20230929：add multi args parsing for `callee`.
- - 20230930：tîng-khí parser, using `js-token`.
+ - 20230930：tîng khí parser, using `js-token`.
+ - 20231016: tîng siá parser, using `ts-parsec`.
--- a/package-lock.json
+++ b/package-lock.json
@ -13,7 +13,8 @@
        "harfbuzzjs": "^0.3.3",
        "js-tokens": "^8.0.2",
        "npx": "^3.0.0",
-        "pdf-lib": "^1.17.1"
+        "pdf-lib": "^1.17.1",
+        "typescript-parsec": "^0.3.4"
      },
      "devDependencies": {
        "@types/chai": "^4.3.5",
@ -6411,6 +6412,11 @@
        "node": ">=14.17"
      }
    },
+    "node_modules/typescript-parsec": {
+      "version": "0.3.4",
+      "resolved": "https://registry.npmjs.org/typescript-parsec/-/typescript-parsec-0.3.4.tgz",
+      "integrity": "sha512-6RD4xOxp26BTZLopNbqT2iErqNhQZZWb5m5F07/UwGhldGvOAKOl41pZ3fxsFp04bNL+PbgMjNfb6IvJAC/uYQ=="
+    },
    "node_modules/unbox-primitive": {
      "version": "1.0.2",
      "resolved": "https://registry.npmjs.org/unbox-primitive/-/unbox-primitive-1.0.2.tgz",
@ -11275,6 +11281,11 @@
      "integrity": "sha512-mI4WrpHsbCIcwT9cF4FZvr80QUeKvsUsUvKDoR+X/7XHQH98xYD8YHZg7ANtz2GtZt/CBq2QJ0thkGJMHfqc1w==",
      "dev": true
    },
+    "typescript-parsec": {
+      "version": "0.3.4",
+      "resolved": "https://registry.npmjs.org/typescript-parsec/-/typescript-parsec-0.3.4.tgz",
+      "integrity": "sha512-6RD4xOxp26BTZLopNbqT2iErqNhQZZWb5m5F07/UwGhldGvOAKOl41pZ3fxsFp04bNL+PbgMjNfb6IvJAC/uYQ=="
+    },
    "unbox-primitive": {
      "version": "1.0.2",
      "resolved": "https://registry.npmjs.org/unbox-primitive/-/unbox-primitive-1.0.2.tgz",
--- a/package.json
+++ b/package.json
@ -41,6 +41,7 @@
    "harfbuzzjs": "^0.3.3",
    "js-tokens": "^8.0.2",
    "npx": "^3.0.0",
-    "pdf-lib": "^1.17.1"
+    "pdf-lib": "^1.17.1",
+    "typescript-parsec": "^0.3.4"
  }
 }
--- a/src/index.js
+++ b/src/index.js
@ -22,14 +22,11 @@ var __importStar = (this && this.__importStar) || function (mod) {
    __setModuleDefault(result, mod);
    return result;
 };
-var __importDefault = (this && this.__importDefault) || function (mod) {
-    return (mod && mod.__esModule) ? mod : { "default": mod };
-};
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.matchAny = exports.tkTreeToSExp = void 0;
+exports.tkTreeToSExp = void 0;
 var fs = require('fs');
-const js_tokens_1 = __importDefault(require("js-tokens"));
 const util = __importStar(require("util"));
+const p = __importStar(require("typescript-parsec"));
 /**
 *
 * # REPRESENTATION
@ -50,7 +47,7 @@ function tkTreeToSExp(t) {
            str = "%undefined";
        }
        else {
-            str = t.value;
+            str = t;
        }
    }
    return str;
@ -58,129 +55,65 @@ function tkTreeToSExp(t) {
 exports.tkTreeToSExp = tkTreeToSExp;
 /**inspect the inner of the representation. */
 let repr = (x) => { return util.inspect(x, { depth: null }); };
+var TokenKind;
+(function (TokenKind) {
+    TokenKind[TokenKind["Seperator"] = 0] = "Seperator";
+    TokenKind[TokenKind["Semicolon"] = 1] = "Semicolon";
+    TokenKind[TokenKind["Number"] = 2] = "Number";
+    TokenKind[TokenKind["Op"] = 3] = "Op";
+    TokenKind[TokenKind["ExprMark"] = 4] = "ExprMark";
+    TokenKind[TokenKind["Paren"] = 5] = "Paren";
+    TokenKind[TokenKind["SpaceNL"] = 6] = "SpaceNL";
+    TokenKind[TokenKind["Id"] = 7] = "Id";
+    TokenKind[TokenKind["Str"] = 8] = "Str";
+})(TokenKind || (TokenKind = {}));
 /**
- *
- * # PARSER UNITS
+ * Parsing
 */
-function toSome(x) {
-    return { _tag: "Some", value: x };
-}
-/**
- * like `m ==> f` in ocaml
- * @param m matchee wrapped
- * @param f matching function
- * @returns wrapped result
- */
-function thenDo(m, f) {
-    if (m._tag == "None") {
-        return m;
-    }
-    else {
-        var a = f(m.value);
-        if (a._tag == "Some") {
-            a.value.ast = m.value.ast.concat(a.value.ast);
-        }
-        return a;
-    }
-}
-/**
- *
- * @param m : the `TokenPair` to be consumed.
- * @returns if the length of `m.remained` >= 1; consumes the matchee by 1 token
- *  and wraps it in `Some`,
- * otherwise, returns `None`.
- */
-function matchAny(m) {
-    if (m.remained.length >= 1) {
-        return {
-            _tag: "Some", value: {
-                matched: m.matched.concat(m.remained[0]),
-                remained: m.remained.slice(1),
-                ast: [m.remained[0]],
-            }
-        };
-    }
-    else {
-        return { _tag: "None" };
-    }
-}
-exports.matchAny = matchAny;
-/**
- * like `f1 | f2` in regex
- * @param f1 the first tried function
- * @param f2 the second tried function
- * @returns wrapped result
- */
-function orDo(f1, f2) {
-    return (x) => {
-        let res1 = f1(x);
-        if (res1._tag == "Some") {
-            return res1;
-        }
-        else {
-            let res2 = f2(x);
-            return res2;
-        }
-    };
-}
-/**
- * like regex [^c]
- * @param f input token function. one token only.
- * @returns combined finction
- */
-function notDo(f) {
-    return (x) => {
-        let res1 = f(x);
-        if (res1._tag == "Some") {
-            return { _tag: "None" };
-        }
-        else {
-            let res2 = matchAny(x);
-            return res2;
-        }
-    };
-}
-function matchToken(typeName, value) {
-    return (t) => {
-        let headToken = t.remained[0];
-        if (headToken.type != typeName) {
-            return { _tag: "None" };
-        }
-        else {
-            if (value === undefined || value == headToken.value) {
-                let newTokenPair = {
-                    matched: t.matched.concat(headToken),
-                    remained: t.remained.slice(1),
-                    ast: [headToken]
-                };
-                return { _tag: "Some", value: newTokenPair };
-            }
-            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]
+]);
 /**
 *
 * # TEST
 */
-const tokens = Array.from((0, js_tokens_1.default)(`import; foo from\t 'bar';
-import * as util  from 'util';
-
-
-花非花，霧\\{非霧 。{{foo();}}下
-一句`));
-console.log("RESULT=" + repr(tokens));
-var mainTokenPair = {
-    matched: [],
-    remained: tokens,
-    ast: []
-};
-let a = thenDo(thenDo(toSome(mainTokenPair), matchToken('IdentifierName')), notDo(matchToken('Punctuator', ';')));
-console.log("RESULT=" + repr(a));
-if (a._tag == "Some") {
-    console.log("SEXP=" + tkTreeToSExp(a.value.ast));
+const inputTxt = `import ast;
+---
+122`;
+const PROG = p.rule();
+const UNIT = p.rule();
+const IMPORTS = p.rule();
+const SEMICOLON = p.rule();
+let doubleMinus = { type: 'Punctuator', value: '--' };
+let doubleMinus2 = p.str('--');
+const TERM = p.rule();
+function applyUnit(value) {
+    return value.text;
 }
+function applySemiColon(value) {
+    return value.text;
+}
+function applyParts(first, second) {
+    return ["%clo", first, second[1]];
+}
+PROG.setPattern(p.lrec_sc(IMPORTS, p.seq(p.str('---'), UNIT), applyParts));
+function applyImports(input) {
+    let importTail = input[1].map(x => x.text);
+    return ["import"].concat(importTail);
+}
+;
+IMPORTS.setPattern(p.apply(p.seq(p.str('import'), p.rep_sc(p.tok(TokenKind.Id)), SEMICOLON), applyImports));
+SEMICOLON.setPattern(p.apply(p.tok(TokenKind.Semicolon), applySemiColon));
+UNIT.setPattern(p.apply(p.tok(TokenKind.Number), applyUnit));
+let tree = p.expectSingleResult(p.expectEOF(PROG.parse(lexer.parse(inputTxt))));
+console.log("RESULT=" + tkTreeToSExp(tree));
--- a/src/index.ts
+++ b/src/index.ts
@ -1,7 +1,8 @@
 var fs = require('fs');
 import jsTokens from "js-tokens";
 import * as util from 'util';
-
+import * as p from 'typescript-parsec';
+import { Token } from 'typescript-parsec';
 /**
 * 
 * # REPRESENTATION
@ -21,7 +22,7 @@ export function tkTreeToSExp(t: tkTree): string{
        if (t=== undefined){
            str = "%undefined"
        }else{
-            str = t.value;
+            str = t;
        }
    }

@ -35,156 +36,110 @@ let repr = (x : any)=>{return util.inspect(x, {depth: null})};
 * # TYPES
 */

-/**
- * TokenPair for tokens' parser combinator
- *
- * matched: the matched (now and before) tokens
- * 
- * remained: tokens to be matched
- * 
- * ast: abstract syntax tree
- */
-export interface TokenPair {
-    matched: jsTokens.Token[]
-    remained: jsTokens.Token[]
-    ast : tkTree[]
-}
-export type Some<T> = { _tag: "Some"; value: T };
-export type None = { _tag: "None" };
-export type Maybe<T> = Some<T> | None;

-type Token = jsTokens.Token;
-type tkTree = Token | tkTree[];
+type tkTree = string | tkTree[];

-/**
- * 
- * # PARSER UNITS
- */
-function toSome<T>(x:T): Maybe<T>{
-    return {_tag: "Some", value: x};
+enum TokenKind {
+    Seperator,
+    Semicolon,
+    Number,
+    Op,
+    ExprMark,
+    Paren,
+    SpaceNL,
+    Id,
+    Str,
 }

 /**
- * like `m ==> f` in ocaml
- * @param m matchee wrapped
- * @param f matching function
- * @returns wrapped result
+ * Parsing
 */
-function thenDo(m : Maybe<TokenPair>, f : Function){
-    if (m._tag == "None"){
-        return m;
-    }else{
-        var a : Maybe<TokenPair> = f(m.value);
-        if (a._tag == "Some"){
-            a.value.ast = m.value.ast.concat(a.value.ast);
-        }
-
-        return a;
-    }
-}
-/**
- * 
- * @param m : the `TokenPair` to be consumed.
- * @returns if the length of `m.remained` >= 1; consumes the matchee by 1 token
- *  and wraps it in `Some`,
- * otherwise, returns `None`.
- */
-export function matchAny(m: TokenPair): Maybe<TokenPair> {
-    if (m.remained.length >= 1) {
-        return {
-            _tag: "Some", value: {
-                matched: m.matched.concat(m.remained[0]),
-                remained: m.remained.slice(1),
-                ast :  [m.remained[0]],
-            }
-        };
-    } else {
-        return { _tag: "None" };
-    }
-}
-
-/**
- * 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 : TokenPair) =>{
-        let res1 : Maybe<TokenPair> = f1(x);
-        if (res1._tag == "Some"){
-            return res1;
-        }else{
-            let res2 : Maybe<TokenPair> = f2(x);
-            return res2;
-        }
-    } 
-}
-
-/**
- * like regex [^c]
- * @param f input token function. one token only.
- * @returns combined finction
- */
-function notDo(f : Function){
-    return (x : TokenPair) =>{
-        let res1 : Maybe<TokenPair> = f(x);
-        if (res1._tag == "Some"){
-            return {_tag:"None"};
-        }else{
-            let res2 = matchAny(x);
-            return res2;
-        }
-    } 
-}
-
-function matchToken(typeName : string, value? : string):
- (t : TokenPair) => Maybe<TokenPair>{
-    return (t)=>{
-        let headToken = t.remained[0];
-        if (headToken.type != typeName){
-            return {_tag:"None"};
-        }else{
-            if (value === undefined || value == headToken.value){
-                let newTokenPair = {
-                    matched: t.matched.concat(headToken),
-                    remained: t.remained.slice(1),
-                    ast : [headToken]
-                };
-                return {_tag : "Some", value : newTokenPair};
-            }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]
+]);

 /**
 * 
 * # TEST
 */
-const tokens = Array.from(jsTokens(
-`import foo from\t 'bar';
-import * as util  from 'util';
+const inputTxt=
+`import ast;
+---
+122`;


-花非花，霧\\{非霧 。{{foo();}}下
-一句`));
-
-console.log("RESULT="+repr(tokens));
+const PROG = p.rule<TokenKind, tkTree>();
+const UNIT = p.rule<TokenKind, tkTree>();
+const IMPORTS = p.rule<TokenKind, tkTree>();
+const SEMICOLON = p.rule<TokenKind, tkTree>();


-var mainTokenPair : TokenPair = {
-    matched : [] ,
-    remained : tokens,
-    ast : []};
+let doubleMinus = { type: 'Punctuator', value: '--' };
+let doubleMinus2 = p.str('--');
+const TERM = p.rule<TokenKind, tkTree>();

-let a = thenDo(thenDo(toSome(mainTokenPair), matchToken('IdentifierName')),
-        notDo(matchToken('Punctuator', ';')));
-
-
-console.log("RESULT="+repr(a));
-if (a._tag == "Some"){
-    console.log("SEXP="+tkTreeToSExp(a.value.ast));
+function applyUnit(value: Token<TokenKind.Number>): tkTree{
+    return value.text;
 }
+
+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 applyImports(input: [Token<TokenKind>,Token<TokenKind>[], tkTree]):tkTree{
+    let importTail = input[1].map(x=>x.text);
+    return ["import"].concat(importTail);
+};
+
+/**
+ * PROG : IMPORTS '---' UNIT;
+ */
+PROG.setPattern(
+    p.lrec_sc(IMPORTS, p.seq(p.str('---'), UNIT), applyParts)
+
+)
+
+/**
+ * PROG : 'import' Id* SEMICOLON;
+ */
+IMPORTS.setPattern(
+    p.apply(p.seq(p.str('import'), p.rep_sc(p.tok(TokenKind.Id)), SEMICOLON) , applyImports)
+);
+
+/**
+ * SEMICOLON : ';';
+ */
+SEMICOLON.setPattern(
+    p.apply(p.tok(TokenKind.Semicolon), applySemiColon)
+);
+
+/**
+ * UNIT : Number;
+ */
+UNIT.setPattern(
+    p.apply(p.tok(TokenKind.Number), applyUnit)
+);
+
+let tree = p.expectSingleResult(p.expectEOF(PROG.parse(lexer.parse(inputTxt))));
+
+
+
+console.log("RESULT="+tkTreeToSExp(tree));