var fs = require('fs');
+import { argv, resourceUsage } from 'node:process';
+import * as tk from './tokenize.js';
+import * as util from 'util';
+import { drawEllipsePath, reduceRotation } from 'pdf-lib';
+import { isAnyArrayBuffer, isTypedArray } from 'node:util/types';
+import { error } from 'node:console';
+import { isUndefined } from 'node:util';
-export type Some<T> = { _tag: "Some"; value: T };
-export type None = {_tag: "None"};
-
-
-/**
- * wrap a x in a `Some(T)`
- * @param x : variable to be wrapped.
- * @returns wrapped `x`.
- */
-function toSome<T>(x: T): Some<T>{
- return { _tag: "Some", value: x};
-}
/**
- * @description Like the `Some(a)` and `None` in Rust.
- *
- * @example
- * ```ts
- * let exam1 : Maybe<Number> = { _tag: "Some", value: 12 };
- * let exam2 : Maybe<Number> = None;
- * ```
+ * debug reprensenting
*/
-export type Maybe<T> = Some<T> | None;
-
+let repr = (x : any)=>{return util.inspect(x, {depth: null})};
/**
- * @description
- * the pair of the string to be matched later and the string that have been matched
- * @param matched : string have been matched
- * @param remained : string will be tested whether it'll be matched.
+ * token tree type.
*/
-export type MatcheePair = {matched : string; remained : string};
+type tkTree = tkTree[] | tk.Token
/**
- * @description
- * it returns a function which test if the first char of the `remained` part of
- * the argument of the function is `c`, if it's true, update the `MatchedPair` wrapped
- * in `Some`. Otherwise, it returns `None`.
- * * @param c : the char to be test.
- * @returns the updated `MatchedPair` wrapped in `Some(x)` or `None`.
+ * 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.
*/
-export function match1Char(c : string) : (m: MatcheePair) => Maybe<MatcheePair> {
- return (m : MatcheePair)=>{
- if (m.remained.length == 0){
- return { _tag: "None" };
- }
- const charToBeMatched = m.remained[0];
- if (charToBeMatched === c){
- return {_tag: "Some", value :{
- matched : m.matched + charToBeMatched,
- remained : m.remained.substring(1)}};
- }
- else{
- return {_tag: "None"};
- }
+function concat(x: tkTree, y:tkTree): tkTree[] {
+ if (Array.isArray(x)){
+ return x.concat(y);
+ }else{
+ throw new Error("the tkTree can't be concated, because it's not an array.");
+
}
-};
+}
+
+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.");
+
+ }
+}
/**
+ * TokenMatcheePair for tokens' parser combinator
+ *
+ * matched: the matched (now and before) tokens
*
- * @param m : the `MatcheePair` to be consumed.
- * @returns if the length of `m.remained` >= 1; consumes the matchee by 1 char and wraps it in `Some`,
- * otherwise, returns `None`.
+ * remained: tokens to be matched
+ *
+ * ast: abstract syntax tree
*/
-export function matchAny(m : MatcheePair) : Maybe<MatcheePair>{
- if (m.remained.length >= 1){
- return {_tag: "Some", value :{
- matched : m.matched + m.remained[0],
- remained : m.remained.substring(1)}};
+export interface TokenMatcheePair {
+ matched: tk.Token[]
+ remained: tk.Token[]
+ ast : tkTree[]
+}
+
+/**
+ * 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{
- return {_tag: "None"};
+ if (t=== undefined){
+ str = "%undefined"
+ }else{
+ str = t.text;
+ }
}
+
+ return str;
}
/**
* @description
- * it returns a function which test if the first char of the `remained` part of
- * the argument of the function is between `l` and `u`, if it's true, update the `MatchedPair` wrapped
+ * 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 l : lower bound char, 1-char string
- * * @param u : upper bound char, 1-char string
- * @returns the updated `MatchedPair` wrapped in `Some(x)` or `None`.
- */
-export function matchRange(l : string, u : string) : (m: MatcheePair) => Maybe<MatcheePair> {
- let lCodepoint = charToCodepoint(l);
- let uCodepoint = charToCodepoint(u);
- if (l > u){
- throw new Error("Error: the codepoint of `"+l+"` is not smaller than `"+u+"`)");
- }
- return (m : MatcheePair)=>{
- if (m.remained.length < 1){
- return {_tag : "None"};
+ * * @param typ : the type to be test.
+ * @returns the updated `TokenMatcheePair` wrapped in `Some(x)` or `None`.
+ */
+export function m1TType(typ: tk.TokenType):
+ (m: TokenMatcheePair) => tk.Maybe<TokenMatcheePair> {
+ return (m: TokenMatcheePair) => {
+ if (m.remained.length == 0) {
+ return { _tag: "None" };
}
- const charToBeMatched = m.remained[0];
- const codePointToBeMatched = charToCodepoint(charToBeMatched);
- if (codePointToBeMatched >= lCodepoint && codePointToBeMatched <= uCodepoint){
- return {_tag: "Some", value :{
- matched : m.matched + charToBeMatched,
- remained : m.remained.substring(1)}};
+ /**
+ * 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"};
+ else {
+ return { _tag: "None" };
}
}
};
/**
- * convert the one-char string to codepoint.
- * @param s : the string to code point.
- * @returns if `s.length > 1` return error; otherwise, return the codepoint of `s`.
+ * type int
*/
-export function charToCodepoint(s : string): number{
- if (s.length > 1){
- throw new Error("Error: the length of input string for "+s+ "is "+s.length+`,
- however, it should be 1.`);
+let tInt = m1TType(tk.TokenType.INT);
+let tId = m1TType(tk.TokenType.ID);
+
+
+let tAdd = m1TType(tk.TokenType.I_ADD);
+let tSub = m1TType(tk.TokenType.I_SUB);
+let tMul = m1TType(tk.TokenType.I_MUL);
+let tDiv = m1TType(tk.TokenType.I_DIV);
+let tLParen = m1TType(tk.TokenType.L_PAREN);
+let tRParen = m1TType(tk.TokenType.R_PAREN);
+let tComma = m1TType(tk.TokenType.COMMA);
+let toSome = tk.toSome;
+
+
+argv.forEach((val, index) => {
+ console.log(`${index}=${val}`);
+});
+
+
+/**
+ * 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{
- return s.charCodeAt(0);
+ var a : tk.Maybe<TokenMatcheePair> = f(m.value);
+ if (a._tag == "Some"){
+ a.value.ast = concat(m.value.ast, a.value.ast);
+ }
+
+ return a;
}
}
/**
- * @description thendo(input, f, ...) like
- * a ==> f
- * @param input: the wrapped input.
- * @param f: the function to be applied.
- *
- * @returns:the applied wrapped result `MatcheePair`.
+ * like `f1 | f2` in regex
+ * @param f1 the first tried function
+ * @param f2 the second tried function
+ * @returns wrapped result
*/
-export function thenDo<T>(input : Maybe<T>, f : Function) : Maybe<T>{
- if (input._tag == "None"){
- return input;
- }
- else{
- let inner = input.value;
- return f(inner);
- }
+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;
+ }
+ }
}
+
/**
- * @description "or", like the regex `( f1 | f2 )` .
- * It returns a function `f` of which the argument is`x`.
- * if `f1(x)` is None, then `f` returns `f2(x)`. Otherwise,
- * `F` returns `f1(x)`.
- * @param f1 : 1st function to be compared
- * @param f2 : 2nd function to be compared
- * @returns:the combined function
+ *
+ * @param m : the `MatcheePair` 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 orDo<T>(f1 : Function, f2: Function) : (x : T ) => Maybe<T>{
- return (x) => {
- let f1x : Maybe<T> = (f1(x));
- {
- if (f1x._tag == "None"){
- return f2(x);
+export function matchAny(m: TokenMatcheePair): tk.Maybe<TokenMatcheePair> {
+ 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 f1x;
- }
- }
- };
+ };
+ } else {
+ return { _tag: "None" };
+ }
}
-
/**
+ * Danger : Maybe it's not enough to work.
* @description repeating matching function `f`
* zero or more times, like the asterisk `*` in regex `f*` .
* @param f : the function to be repeated 0+ times.
* @returns:the combined function
*/
-export function zeroOrMoreDo<T>(f : Function): (x : T) => Maybe<T>{
- return (x)=>{
- var wrapped_old_x : Maybe<T> = {_tag: "Some", value : x};
- var wrapped_new_x : Maybe<T> = wrapped_old_x;
-
- while (wrapped_new_x._tag != "None"){
- wrapped_old_x = wrapped_new_x;
- wrapped_new_x = thenDo(wrapped_old_x, f);
+export function OnceOrMoreDo(f: Function): (x: TokenMatcheePair) =>
+ tk.Maybe<TokenMatcheePair> {
+ return (x) => {
+ var wrappedOldX: tk.Maybe<TokenMatcheePair> = { _tag: "Some", value: x };
+ var wrappedNewX: tk.Maybe<TokenMatcheePair> = wrappedOldX;
+
+ var counter = -1;
+
+ while (wrappedNewX._tag != "None") {
+ wrappedOldX = wrappedNewX;
+ wrappedNewX = thenDo(wrappedOldX, f);
+ counter += 1;
+
};
- return wrapped_old_x;
- };
+ if (counter <= 0){
+ return { _tag: "None"};
+ }
+ let ast = wrappedOldX.value.ast ;
+ wrappedOldX.value.ast =ast.slice(ast.length-counter);
+ console.log(repr(wrappedOldX.value.ast));
+
+ return wrappedOldX; };
}
/**
-* @description Not. like the `^` inside regex of [^f].
-* returns a function `F(x)` such that if `f(x)` is `None`,
-* returns the x consuming a char; if `f(x)` is not None, F(x)
-* returns `None`.
-* @param f: the function forbidden to be matched.
-* @returns: combined function `F`.
-*/
-export function notDo<T>(f : Function): (x : T) => Maybe<T>{
- return (x)=>{
- let wrapped_x : Maybe<T> = {
- _tag : "Some",
- value : x
- };
- let f_x = thenDo(wrapped_x, f);
+ * aux function for midfix operator
+ * @param f function
+ * @param signal the rule name
+ * @returns
+ */
+let midfix = (f : Function, signal? : string) => (x : TokenMatcheePair)=>{
+ var a = f(x);
+ if (a._tag == "Some"){
+ 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));
+
- if (f_x._tag != "None"){
- return {_tag:"None"};
- }else{
- return thenDo(wrapped_x, matchAny);
+ }
+ return a;
+}
+
+let circumfix = (f : Function, signal? : string) => (x : TokenMatcheePair)=>{
+ var a : tk.Maybe<TokenMatcheePair> = f(x);
+ if (a._tag == "Some"){
+ console.log("$$$"+repr(a.value.ast));
+ let inner = a.value.ast[a.value.ast.length-2];
+ var ast_middle : tkTree[];
+ // the list should not be (%list) (%apply) (%lambda) etc.
+ if (Array.isArray(inner)){
+ if ('text' in inner[0] && (inner[0].text.slice(0,1) != "%")){
+ ast_middle = inner;
+ }
+ else{
+ ast_middle = [inner];
+ }
}
- };
+ else{
+ ast_middle = [inner];
+ }
+ let new_ast = [ast_middle];
+ a.value.ast = new_ast;
+ }
+ return a;
}
+/** single1 = tInt | "(" expr ")"*/
+let single1 = circumfix((x : TokenMatcheePair) =>
+ thenDo(thenDo(thenDo(toSome(x), tLParen), expr), tRParen), "fac1");
+let single2= tInt;
+let single = orDo(single1, single2);
+
+/** args = single "," args | single */
+let args1 = (x: TokenMatcheePair)=>{
+ var ret = thenDo(thenDo(thenDo(toSome(x), single), tComma), args);
+ if (ret._tag == "Some"){
+ let retLength = ret.value.ast.length;
+ ret.value.ast = [[ret.value.ast[retLength-3]].concat(ret.value.ast[retLength-1])];
+ console.log("$$"+repr(ret.value.ast));
+ }
+ return ret;
+};
+
+let args2 = single;
+
+let args = orDo(args1, args2);
+
+
+/** callees = "(" args ")" | "(" ")" */
+
+
+let callees1 = circumfix((x : TokenMatcheePair) =>
+ thenDo(thenDo(thenDo(toSome(x), tLParen), args), tRParen), "callees1");
+let callees2 = (x: TokenMatcheePair)=>{
+ let ret = thenDo(thenDo(toSome(x), tLParen), tRParen);
+ if (ret._tag == "Some"){
+ let new_ast : tkTree[] = [[]];
+ ret.value.ast = new_ast;
+ }
+
+ return ret};
+
+let callees = orDo(callees1, callees2);
+
+
+
+/** %apply R combinating token */
+let applyToken = {
+ text: "%apply",
+ type: tk.TokenType.ID,
+ col: 0,
+ ln: 0,
+}
+
+/** facAux = callees facAux | callees */
+let facAux1 = (x: TokenMatcheePair)=>{
+ var ret = thenDo(thenDo(toSome(x), callees), facAux);
+ if (ret._tag == "Some"){
+ console.log("1232345"+repr(tkTreeToSExp(ret.value.ast[ret.value.ast.length-1])));
+ let last1 = ret.value.ast[ret.value.ast.length-1];
+ let last2 = ret.value.ast[ret.value.ast.length-2];
+
+
+ let b : tkTree[] = [applyToken];
+ ret.value.ast = [b.concat([last2, last1])];
+ console.log("11111"+repr(tkTreeToSExp(ret.value.ast)));
+
+ };
+
+return ret;}
+let facAux2 = callees;
+let facAux = orDo(facAux1, facAux2);
+
+
+
+/** fac = single facAux | single
+ * Issue1 to be fixed.
+ */
+let fac1 = (x: TokenMatcheePair)=>{
+ var ret = thenDo(thenDo(toSome(x), single),facAux);
+ if(ret._tag == "Some"){
+ console.log("777"+repr(tkTreeToSExp(ret.value.ast)));
+ ret.value.ast = [applyToken, ret.value.ast[ret.value.ast.length-2],
+ ret.value.ast[ret.value.ast.length-1]];
+ ret.value.ast;
+ rearrangeTree(ret.value.ast);
+ console.log("888"+repr(tkTreeToSExp(ret.value.ast)));
+
+ }
+
+ return ret;};
+let fac2 = single;
+let fac = orDo(fac1, fac2);
+
+
/**
- * if `x` is matched by `f` once, returns `f(x)`. Otherwise,
- * returns x
- * similar to `?` in regex `f?`.
- * @param f : the function to be matched
- * @returns return wrapped f(x)
+ * rearrangeTree : for applyToken subtree from right-combination to
+ * left-combination
+ * @input x a ast
+ * @return another ast
*/
-export function zeroOrOnceDo<T>(f : Function): (x : T) => Maybe<T>{
- return (x)=>{
- var wrapped_old_x : Maybe<T> = {_tag: "Some", value : x};
- var wrapped_new_x = thenDo(wrapped_old_x, f);
+function rearrangeTree(x: any) : any {
+
+ if (x !== undefined){
+ for (var i=1;i<x.length;i++){
+ rearrangeTree(x[i]);
+ }
+ console.log("@@"+repr(x[0]));
- if (wrapped_new_x._tag != "None"){
- return wrapped_new_x;
- }else{
- return wrapped_old_x;
- }
- };
-}
+ if (x[0] == applyToken){
+ if (Array.isArray(x[2]) && x[2][0] == applyToken){
+ let rl = rearrangeTree(x[2][1]);
+ let rr = rearrangeTree(x[2][2]);
+ let l = rearrangeTree(x[1]);
+ x[0] = applyToken;
+ x[1] = [applyToken, l, rl];
+ x[2] = rr;
+ console.log("@@=="+repr(x));
+
+ return x;
+ }
+ else{
+ x[0] = applyToken;
+ x[1] = rearrangeTree(x[1]);
+ x[2] = rearrangeTree(x[2]);
+ console.log("@@=="+repr(x));
-export function tokenize(input : string){
- var input_matchee_pair : Maybe<MatcheePair> = toSome(
- {matched:"",
- remained: input});
+ return x;
+ }
+ }
- // integer = ([+]|[-])\d\d?
- let integer = (x : MatcheePair) =>
- { let wrapped_x = toSome(x);
- let plusMinus = orDo(match1Char('+'), match1Char('-')); // ([+]|[-])
- let d = matchRange('0','9'); // \d
- return thenDo(thenDo(thenDo(wrapped_x,
- zeroOrOnceDo(plusMinus)),d),
- zeroOrMoreDo(d));
+ return x;
+ }
}
- console.log(input+", result: ");
- console.log(thenDo(input_matchee_pair, integer));
- // TODO: id, string, space, basic operator, 3 marks: @, {, }.
+
+
+
+
+/**
+ *
+ * term1 = fac (MUL | DIV) fac
+ */
+
+let term1 = midfix((x : TokenMatcheePair)=>
+ thenDo(thenDo(thenDo(toSome(x), fac), orDo(tMul,tDiv)), fac), "term1");
+
+
+/**
+ *
+ * term2 = int MUL int
+ */
+let term2 = fac;
+
+/**
+ * term = term1 | term2
+ */
+let term = orDo(term1, term2);
+
+
+/**
+ *
+ * expr1 = term ADD term
+ */
+let expr1 = midfix((x : TokenMatcheePair)=>
+ thenDo(thenDo(thenDo(toSome(x), term), orDo(tAdd,tSub)), term), "expr1");
+/**
+ * expr2 = term
+ */
+let expr2 = term;
+
+/**
+ * expr = expr1 | expr2
+ */
+let expr = orDo(expr1, expr2);
+
+
+
+let tokens = tk.tokenize("1");
+let tokens2 = tk.tokenize("1(2)");
+let tokens3 = tk.tokenize("1(2)(3)");
+let tokens4 = tk.tokenize("2()(4)(5,6)(7,8,9,10)");
+
+//let tokens = tk.tokenize("(4-(3/4))");
+//tk.tokenize(argv[2]);
+
+let tokensFiltered = tokens4.filter(
+ (x)=>{return (x.type != tk.TokenType.NL
+ && x.type != tk.TokenType.SP)});
+
+
+
+let beta = expr({
+ matched : [] ,
+ remained : tokensFiltered,
+ ast : []});
+
+
+
+if (beta._tag == "Some"){
+ beta.value.ast = rearrangeTree(beta.value.ast);
+ console.log(tkTreeToSExp(beta.value.ast));
}
-tokenize("+123");
-tokenize("123");
-tokenize("-123");
-tokenize(" 123");
-tokenize("c123");
\ No newline at end of file
+console.log("RESULT="+repr(beta));
+
projects / clo / blobdiff