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';
/**
- * token tree type.
+ *
+ * # REPRESENTATION
+ */
+/**
+ * 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.value;
+ }
+ }
-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}`);
-});
-
-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 }));
/**
- * matchee pair of commandInputTokenized
+ * TokenPair for tokens' parser combinator
+ *
+ * matched: the matched (now and before) tokens
+ *
+ * remained: tokens to be matched
+ *
+ * ast: abstract syntax tree
*/
-let commandTPair : TokenMatcheePair = {matched:[],
- remained: commandInputTokenizedFiltered};
-
-
-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);
+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[];
-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"};
- }
+/**
+ *
+ * # PARSER UNITS
+ */
+function toSome<T>(x:T): Maybe<T>{
+ return {_tag: "Some", value: x};
}
/**
- * 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.
+ * like `m ==> f` in ocaml
+ * @param m matchee wrapped
+ * @param f matching function
+ * @returns wrapped result
*/
-function gramRHS (process: Function, arrange : number[]){
- return (m : TokenMatcheePair)=>{
-
- let middle : tk.Maybe<TokenMatcheePair> = process(m);
-
- console.log("Middle"+util.inspect(middle, { showHidden: true, depth: null }));
+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);
+ }
- 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;
- }
+ return a;
}
}
-
/**
- * typeABS ::= "'" ID
+ *
+ * @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`.
*/
-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];
+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" };
}
- return result;
}
/**
- * TypeId ::= typeABS | ID
+ * like `f1 | f2` in regex
+ * @param f1 the first tried function
+ * @param f2 the second tried function
+ * @returns wrapped result
*/
-var typeName = (x : TokenMatcheePair)=>
-{
- return thenDo(toSome(x), orDo(typeABS, tId));
+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;
+ }
+ }
}
/**
- * CONST ::= INT | STR | FLO | BOOL
+ * 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;
+ }
+ }
+}
-/**
- * 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]);
+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"};
+ }
+ };
+ }
+};
-/**
- * storing the tree
- */
-var astTree : tkTree = [];
/**
- * TYPE_PAIR ::= TYP_ID ID
+ *
+ * # TEST
*/
-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;
- }
-}
+const tokens = Array.from(jsTokens(
+`import foo from\t 'bar';
+import * as util from 'util';
-/**
- * function's arguments
- * FN_ARGS = TYPE_PAIR ("," TYPE_PAIR)+
- */
-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 }));
+花非花,霧\\{非霧 。{{foo();}}下
+一句`));
+
+console.log("RESULT="+repr(tokens));
- return {maybeTokens : abanibi.maybeTokens, ast : [a.ast, abanibi.ast]};
-
- };
+var mainTokenPair : TokenPair = {
+ matched : [] ,
+ remained : tokens,
+ ast : []};
-let tree = fnArgs(commandTPair);
-console.log("CHRANN"+util.inspect(tree, { showHidden: true, depth: null }));
+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));
+}