X-Git-Url: https://git.kianting.info/?a=blobdiff_plain;f=src%2Findex.ts;h=599d32a9040ef78c4127eff01698f54b0aa95ab1;hb=9784854866f652eb2118c390425a6cbe7a702ab1;hp=5ca3a12b008cf290bfdc3a21675da93324132912;hpb=cf712efc21fa3a711b8a09b953539c7a7c90f5d6;p=clo diff --git a/src/index.ts b/src/index.ts index 5ca3a12..599d32a 100644 --- a/src/index.ts +++ b/src/index.ts @@ -1,512 +1,380 @@ -import { match } from "assert"; - var fs = require('fs'); - -export type Some = { _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 { isAnyArrayBuffer, isTypedArray } from 'node:util/types'; +import { error } from 'node:console'; +import { isUndefined } from 'node:util'; /** - * wrap a x in a `Some(T)` - * @param x : variable to be wrapped. - * @returns wrapped `x`. + * debug reprensenting */ -function toSome(x: T): Some{ - 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 = { _tag: "Some", value: 12 }; - * let exam2 : Maybe = None; - * ``` + * token tree type. */ -export type Maybe = Some | 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."); + + } +} + +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."); + + } } /** - * The types of Token - * NL, // newline - * - * SP, // half-width space and tab - * - * ID, // identifier + * TokenMatcheePair for tokens' parser combinator + * + * matched: the matched (now and before) tokens * - * STR, // string + * remained: tokens to be matched * - * OP, // operator or something like it - * - * FLO, // float num - * - * INT, // integer - * - * I_* // integer manipulation - * - * F_* // float manipulation - * - * SEMI_C// semi-colon + * ast: abstract syntax tree */ -export enum TokenType{ - NL, // newlinw - SP, // half-width space and tab - ID, // identifier - STR, // string - FLO, // float num - INT, // integer - F_ADD, - F_SUB, - F_MUL, - F_DIV, - I_ADD, - I_SUB, - I_MUL, - I_DIV, - L_PAREN, // ( - R_PAREN, // ) - L_BRACK, // [ - R_BRACK, // ] - L_BRACE, // { - R_BRACE, // } - COMMA, // , - DOT, // . - COLON, // : - SEMI_C, // ; - AT, // @ - HASH, // # - EQ, // == - SET, // = - GT, // > greater than - LT, // = - LE, // <= - R_ARROW, // -> - +export interface TokenMatcheePair { + matched: tk.Token[] + remained: tk.Token[] + ast : tkTree[] } /** - * 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 + * convert a `tkTree` AST to S-expr string + * @param t the `tkTree` + * @returns S-expr String */ -export interface Token{ - text: string, - type?: TokenType, - col: number, - ln: number, +export function tkTreeToSExp(t: tkTree): string{ + var str = ""; + + if (Array.isArray(t)){ + let strArray = t.map((x)=>tkTreeToSExp(x)); + str = "(" + strArray.join(" ") + ")"; + }else{ + 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 `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 { - return (m : MatcheePair)=>{ - if (m.remained.length == 0){ +export function m1TType(typ: tk.TokenType): + (m: TokenMatcheePair) => tk.Maybe { + 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 = { + _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{ - 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 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 toSome = tk.toSome; + + +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 { - 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, f : Function){ + if (m._tag == "None"){ + return m; }else{ - return s.charCodeAt(0); + var a : tk.Maybe = 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(input : Maybe, f : Function) : Maybe{ - 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 = f1(x); + if (res1._tag == "Some"){ + return res1; + }else{ + let res2 : tk.Maybe = 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(f1 : Function, f2: Function) : (x : T ) => Maybe{ - return (x) => { - let f1x : Maybe = (f1(x)); - { - if (f1x._tag == "None"){ - return f2(x); +export function matchAny(m: TokenMatcheePair): tk.Maybe { + 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(f : Function): (x : T) => Maybe{ - return (x)=>{ - var wrapped_old_x : Maybe = {_tag: "Some", value : x}; - var wrapped_new_x : Maybe = 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 { + return (x) => { + var wrappedOldX: tk.Maybe = { _tag: "Some", value: x }; + var wrappedNewX: tk.Maybe = 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(f : Function): (x : T) => Maybe{ - return (x)=>{ - let wrapped_x : Maybe = { - _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 = f(x); + if (a._tag == "Some"){ + let inner = a.value.ast[a.value.ast.length-2]; + let ast_middle : tkTree[] = [inner]; + let new_ast = [ast_middle]; + a.value.ast = new_ast; + } + return a; } /** - * 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) + * TODO: 12(13)(14) only parsed with only 12(13) */ -export function zeroOrOnceDo(f : Function): (x : T) => Maybe{ - return (x)=>{ - var wrapped_old_x : Maybe = {_tag: "Some", value : x}; - var wrapped_new_x = thenDo(wrapped_old_x, f); +/** single1 = tInt | "(" expr ")"*/ +let single1 = circumfix((x : TokenMatcheePair) => + thenDo(thenDo(thenDo(tk.toSome(x), tLParen), expr), tRParen), "fac1"); +let single2= tInt; +let single = orDo(single1, single2); + +/** func = single | single "(" single ")" + * i.e. + * + * func = single | func_aux ( int ) + * +*/ - if (wrapped_new_x._tag != "None"){ - return wrapped_new_x; - }else{ - return wrapped_old_x; - } - }; -} +/** fac = single ["(" single ")"]? | single + * Issue1 to be fixed. + */ +let fac1Appliee = circumfix((x : TokenMatcheePair) => thenDo(thenDo(thenDo(tk.toSome(x), tLParen), tInt), tRParen), "fac1"); +let fac1 = (x : TokenMatcheePair) => + { + let raw = thenDo(thenDo(toSome(x), single), OnceOrMoreDo(fac1Appliee)); -export function tokenize(input : string){ - var input_matchee_pair : Maybe = toSome( - {matched:"", - remained: input}); - - /** - * generate a parser of a basic term (b_term) - * @param pattern : the pattern parser - * @param token_type : the returning token type - * @returns a wrapped parser. - */ - function bTerm(pattern : Function, token_type : TokenType){ - return (x : MatcheePair) =>{ - let wrapped_x = toSome(x); - let result = pattern(wrapped_x); - if (result._tag == "Some") { - result.value.matched_type = token_type; + + + if (raw._tag == "Some"){ + + + var result : tkTree = raw.value.ast[0]; + let applyToken : tk.Token = {text: '%apply', ln:0, col:0}; + for (var i=1; i)=> - thenDo(thenDo(thenDo(x, - zeroOrOnceDo(plusMinus)),d), - zeroOrMoreDo(d)), - TokenType.INT); - // space = [ \t]+ - let space = bTerm((x : Maybe)=> - thenDo(thenDo(x, s_aux), zeroOrMoreDo(s_aux)), - TokenType.INT); - - // newline = \r?\n - let newline = bTerm((x : Maybe)=> - thenDo(thenDo(x, - zeroOrOnceDo(match1Char('\r'))), - match1Char('\n')), - TokenType.NL); - - // [_A-Za-z] - let idHead = orDo(orDo(matchRange('a','z'),matchRange('A','Z')), match1Char('_')); - let idRemained = orDo(idHead, matchRange('0','9')); // [_A-Za-z0-9] - - // id = [_A-Za-z][_A-Za-z0-9]* - let id = bTerm((x : Maybe)=> - thenDo(thenDo(x, - idHead), - zeroOrMoreDo(idRemained)), - TokenType.ID); - let doublequote = match1Char("\""); - // [\\][\"] - let escapeReverseSlash = (x:MatcheePair)=> - thenDo(thenDo(toSome(x), match1Char("\\")), doublequote); - // ([\\]["]|[^\"])* - let stringInnerPattern = zeroOrMoreDo( - orDo(escapeReverseSlash, notDo(match1Char("\"")))); - - // str = ["]([\\]["]|[^"])*["] - let str = bTerm((x : Maybe)=> - thenDo(thenDo(thenDo(x,doublequote), - stringInnerPattern),doublequote), - TokenType.STR); - - // float = [+-]?\d+[.]\d+ - function floatPattern(x : Maybe){ - return thenDo(thenDo(thenDo(thenDo(thenDo(thenDo(x, - zeroOrOnceDo(plusMinus)),d), - zeroOrMoreDo(d)), - match1Char(".")),d), - zeroOrMoreDo(d))}; - let float = bTerm(floatPattern, TokenType.FLO); - - // operators - // +. - let floatAdd = bTerm((x : Maybe)=> - thenDo(thenDo(x, match1Char("+")),match1Char(".")), - TokenType.F_ADD); - // +. - let floatSub = bTerm((x : Maybe)=> - thenDo(thenDo(x, match1Char("-")),match1Char(".")), - TokenType.F_SUB); - - // *. - let floatMul = bTerm((x : Maybe)=> - thenDo(thenDo(x, match1Char("*")),match1Char(".")), - TokenType.F_MUL); - - // /. - let floatDiv = bTerm((x : Maybe)=> - thenDo(thenDo(x, match1Char("/")),match1Char(".")), - TokenType.F_DIV); - - // == - let eq = bTerm((x : Maybe)=> - thenDo(thenDo(x, match1Char("=")),match1Char("=")), - TokenType.EQ); - - // >= - let ge = bTerm((x : Maybe)=> - thenDo(thenDo(x, match1Char(">")),match1Char("=")), - TokenType.GE); - - // <= - let le = bTerm((x : Maybe)=> - thenDo(thenDo(x, match1Char("<")),match1Char("=")), - TokenType.LE); - - // -> - let rightArrow = bTerm((x : Maybe)=> - thenDo(thenDo(x, match1Char("-")),match1Char(">")), - TokenType.R_ARROW); - - /** - * unary operator : generating the pattern of basic unary operator - * @param char : uniry char for the operator - * @param token_type : the corresponding token_type - */ - function unaryOp(char: string, token_type: TokenType) { - return bTerm((x : Maybe)=>thenDo(x, match1Char(char)), - token_type);}; - - let intAdd = unaryOp('+', TokenType.I_ADD); - let intSub = unaryOp('-', TokenType.I_SUB); - let intMul = unaryOp('*', TokenType.I_MUL); - let intDiv = unaryOp('/', TokenType.I_DIV); - let lParen = unaryOp('(', TokenType.L_PAREN); - let rParen = unaryOp(')', TokenType.R_PAREN); - let lBracket = unaryOp('[', TokenType.L_BRACK); - let rBracket = unaryOp(']', TokenType.R_BRACK); - let lBrace = unaryOp('{', TokenType.L_BRACE); - let rBrace = unaryOp('}', TokenType.R_BRACE); - let comma = unaryOp(',', TokenType.COMMA); - let dot = unaryOp('.', TokenType.DOT); - let colon = unaryOp(':', TokenType.COLON); - let semicolon = unaryOp(';', TokenType.SEMI_C); - let at = unaryOp('@', TokenType.AT); - let hash = unaryOp('#', TokenType.HASH); - let set = unaryOp('=', TokenType.SET); - let greaterthan = unaryOp('>', TokenType.GT); - let lessthan = unaryOp('<', TokenType.LE); - - - let term = (token_list : Array, x : Some)=>{ - var ln = 1; - var col = 0; - var old_x = x; - let term_list = [float, newline, space, integer,str, id, - floatAdd, floatSub, floatMul, floatDiv, - intAdd, intSub, intMul, intDiv, - eq, ge, le, rightArrow, - lParen, rParen, lBracket, rBracket, lBrace, rBrace, - comma, dot, colon, semicolon, at, hash, - set,greaterthan, lessthan]; - let term_aux = term_list.reduce((x,y)=> orDo(x,y)); - - var new_x : Maybe = 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}); + + - } + return raw; + }; +let fac2 = single; +let fac = orDo(fac1, fac2); - 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)); - } +/** + * + * term1 = fac (MUL | DIV) fac + */ + +let term1 = midfix((x : TokenMatcheePair)=> + thenDo(thenDo(thenDo(tk.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(tk.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("(3(2))*2+1"); + +//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)}); - return token_list; - } - console.log(term([], input_matchee_pair)); - // TODO: id, string, space, basic operator, 3 marks: @, {, }. +let beta = expr({ + matched : [] , + remained : tokensFiltered, + ast : []}); +if (beta._tag == "Some"){ + console.log(tkTreeToSExp(beta.value.ast)); } +console.log("RESULT="+repr(beta));