var fs = require('fs');
-
-export type Some<T> = { _tag: "Some"; value: T };
-export type None = {_tag: "None"};
-
+import { argv, resourceUsage } from 'node:process';
+import * as tk from './tokenize.js';
+import * as util from 'util';
+import { drawEllipsePath, reduceRotation } from 'pdf-lib';
+import { isTypedArray } from 'node:util/types';
+import { error } from 'node:console';
/**
- * wrap a x in a `Some(T)`
- * @param x : variable to be wrapped.
- * @returns wrapped `x`.
+ * debug reprensenting
*/
-function toSome<T>(x: T): Some<T>{
- return { _tag: "Some", value: x};
-}
+let repr = (x : any)=>{return util.inspect(x, {depth: null})};
+
/**
- * @description Like the `Some(a)` and `None` in Rust.
- *
- * @example
- * ```ts
- * let exam1 : Maybe<Number> = { _tag: "Some", value: 12 };
- * let exam2 : Maybe<Number> = None;
- * ```
+ * token tree type.
*/
-export type Maybe<T> = Some<T> | None;
-
+type tkTree = tkTree[] | tk.Token
/**
- * @description
- * the pair of the string to be matched later and the string that have been matched
- * @var matched : have been matched
- * @var remained : will be tested whether it'll be matched.
- * @var matched_type (optional): the type of the matched string
-*/
-export interface MatcheePair {
- matched : string
- remained : string
- matched_type?: TokenType
+ * 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.
+ */
+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.");
+
+ }
}
-/**
- * The types of Token
- * NL, // newline
- *
- * SP, // half-width space and tab
- *
- * ID, // identifier
- *
- * STR, // string
- *
- * OP, // operator or something like it
- *
- * FLO, // float num
- *
- * INT, // Integer
- */
-export enum TokenType{
- NL, // newlinw
- SP, // half-width space and tab
- ID, // identifier
- STR, // string
- OP, // operator
- FLO, // float num
- INT, // integer
+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.");
+
+ }
}
/**
- * tokenized token.
- * @var text : the content text
- * @var type (optional): the type of the token
- * @var col : the column number
- * @var ln : the line number
+ * TokenMatcheePair for tokens' parser combinator
+ * matched: the matched (now and before) tokens
+ * remained: tokens to be matched
+ * ast: abstract syntax tree
*/
-export interface Token{
- text: string,
- type?: TokenType,
- col: number,
- ln: number,
+export interface TokenMatcheePair {
+ matched: tk.Token[]
+ remained: tk.Token[]
+ ast : tkTree[]
}
/**
* @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
+ * 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 c : the char to be test.
- * @returns the updated `MatchedPair` wrapped in `Some(x)` or `None`.
+ * * @param typ : the type to be test.
+ * @returns the updated `TokenMatcheePair` wrapped in `Some(x)` or `None`.
*/
-export function match1Char(c : string) : (m: MatcheePair) => Maybe<MatcheePair> {
- return (m : MatcheePair)=>{
- if (m.remained.length == 0){
+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];
- if (charToBeMatched === c){
- 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" };
}
}
};
/**
- *
- * @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`.
+ * type int
*/
-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)}};
- }else{
- return {_tag: "None"};
- }
-}
+let tInt = m1TType(tk.TokenType.INT);
+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);
+
+
+
+argv.forEach((val, index) => {
+ console.log(`${index}=${val}`);
+});
-/**
- * @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
- * 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"};
- }
- 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)}};
- }
- 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`.
+ * like `m ==> f` in ocaml
+ * @param m matchee wrapped
+ * @param f matching function
+ * @returns wrapped result
*/
-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.`);
+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
+ * aux function for midfix operator
+ * @param f function
+ * @param signal the rule name
+ * @returns
*/
-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);
- }
- else{
- return f1x;
- }
- }
- };
-}
+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));
-/**
-* @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);
- };
-
- return wrapped_old_x;
- };
+
+ }
+ return a;
}
-/**
-* @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);
+let circumfix = (f : Function, signal? : string) => (x : TokenMatcheePair)=>{
+ var a = f(x);
+ if (a._tag == "Some"){
+ let inner = a.value.ast[a.value.ast.length-2];
+ console.log("AST===="+repr(a.value.ast));
+ let ast_middle : tkTree[] = [inner];
+ let new_ast = [ast_middle];
+ 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;
}
+/** fac1 = "(" expr ")" */
+let fac1 = circumfix((x : TokenMatcheePair)=>
+thenDo(thenDo(thenDo(tk.toSome(x), tLParen), expr), tRParen), "fac1");
+
+let fac2 = tInt;
+
+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)
+ *
+ * term1 = fac (MUL | DIV) fac
*/
-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);
- if (wrapped_new_x._tag != "None"){
- return wrapped_new_x;
- }else{
- return wrapped_old_x;
- }
- };
-}
+let term1 = midfix((x : TokenMatcheePair)=>
+ thenDo(thenDo(thenDo(tk.toSome(x), fac), orDo(tMul,tDiv)), fac), "term1");
+
+/**
+ *
+ * term2 = int MUL int
+ */
+let term2 = fac;
-export function tokenize(input : string){
- var input_matchee_pair : Maybe<MatcheePair> = toSome(
- {matched:"",
- remained: input});
+/**
+ * term = term1 | term2
+ */
+let term = orDo(term1, term2);
+
- // integer = ([+]|[-])?\d\d*
- let integer = (x : MatcheePair) =>
- { let wrapped_x = toSome(x);
- let plusMinus = orDo(match1Char('+'), match1Char('-')); // ([+]|[-])
- let d = matchRange('0','9'); // \d
- var result = thenDo(thenDo(thenDo(wrapped_x,
- zeroOrOnceDo(plusMinus)),d),
- zeroOrMoreDo(d));
-
- if (result._tag == "Some"){
- result.value.matched_type = TokenType.INT;
- }
- return result;
- }
- let space = (x : MatcheePair) =>{
- let wrapped_x = toSome(x);
- let s_aux = orDo(match1Char(' '), match1Char('\t')); // (" " | "\t")
- var result = thenDo(thenDo(wrapped_x, s_aux), zeroOrMoreDo(s_aux));
- if (result._tag == "Some"){
- result.value.matched_type = TokenType.SP;
- }
- return result;
- }
- let newline = (x : MatcheePair) =>{
- let wrapped_x = toSome(x);
- // nl = \r?\n
- let result = thenDo(thenDo(wrapped_x,
- zeroOrOnceDo(match1Char('\r'))), match1Char('\n'));
- if (result._tag == "Some"){
- result.value.matched_type = TokenType.NL;
- }
- return result;
- }
+/**
+ *
+ * expr1 = term ADD term
+ */
+let expr1 = midfix((x : TokenMatcheePair)=>
+ thenDo(thenDo(thenDo(tk.toSome(x), term), orDo(tAdd,tSub)), term), "expr1");
+/**
+ * expr2 = term
+ */
+let expr2 = term;
- let term = (token_list : Array<Token>, x : Some<MatcheePair>)=>{
- var ln = 1;
- var col = 0;
- var old_x = x;
- let term_list = [newline, space, integer];
- let term_aux = term_list.reduce((x,y)=> orDo(x,y));
-
- var new_x : Maybe<MatcheePair> = thenDo(old_x, term_aux);
- while (new_x._tag != "None"){
- if (new_x.value.matched_type != TokenType.NL){
- col += new_x.value.matched.length;
- token_list.push({text : new_x.value.matched,
- type: new_x.value.matched_type,
- ln : ln,
- col : col});
-
- }
- else{
- col = 0;
- ln += 1;
-
- token_list.push({text : new_x.value.matched,
- type: new_x.value.matched_type,
- ln : ln,
- col : col});
-
- }
+/**
+ * expr = expr1 | expr2
+ */
+let expr = orDo(expr1, expr2);
- old_x = toSome({matched : "",
- remained : new_x.value.remained});
- new_x = thenDo(old_x, term_aux);
- }
- if (old_x.value.remained.length){
- console.log(token_list);
- throw new Error("the code can't be tokenized is near Ln. "+ln+", Col."+col
- +", starting with "+ old_x.value.remained.substring(0,10));
- }
- return token_list;
- }
+let tokens = tk.tokenize("(4-(3/4))");//tk.tokenize(argv[2]);
+let tokensFiltered = tokens.filter(
+ (x)=>{return (x.type != tk.TokenType.NL
+ && x.type != tk.TokenType.SP)});
- console.log(term([], input_matchee_pair));
+let wrappedTokens : tk.Maybe<TokenMatcheePair> =
+ tk.toSome({
+ matched : [] ,
+ remained : tokensFiltered,
+ ast : []});
- // TODO: id, string, space, basic operator, 3 marks: @, {, }.
+let beta = expr({
+ matched : [] ,
+ remained : tokensFiltered,
+ ast : []});
-}
+console.log(repr(beta));