X-Git-Url: https://git.kianting.info/?a=blobdiff_plain;f=src%2Findex.ts;h=a8d103b5d550e29e63fe0e707a9b36b3b10bd090;hb=e8f894d994f145be9ffef38cdf4fa8fc23ae9e00;hp=8cbd145671c4508ca55759035ba2da57c1edcccb;hpb=e2668789e238707fa38ce3e724e3b64fba5d53b3;p=clo diff --git a/src/index.ts b/src/index.ts index 8cbd145..a8d103b 100644 --- a/src/index.ts +++ b/src/index.ts @@ -1,256 +1,120 @@ var fs = require('fs'); +import { argv } from 'node:process'; +import * as tk from './tokenize.js'; +import * as util from 'util'; -export type Some = { _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(x: T): Some{ - return { _tag: "Some", value: x}; -} /** - * @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 = tk.Token[] | tk.Token -/** - * @description - * the pair of the string to be matched later and the string that have been matched -*/ -export interface MatcheePair { - /** have been matched */ - matched : string - /** will be tested whether it'll be matched. */ - remained : string +export interface TokenMatcheePair { + matched: tkTree[] + remained: 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 + * 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) { + m.matched.push(ttbm); + return { + _tag: "Some", value: { + matched: m.matched, + remained: m.remained.slice(1) + } + }; } - 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`. - */ -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 toSome = tk.toSome; +let thenDo = tk.thenDo; +let orDo = tk.orDo; -/** - * @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"}; - } - } -}; + +argv.forEach((val, index) => { + console.log(`${index}=${val}`); +}); + +let commandInput = argv[2]; +let commandInputTokenized = tk.tokenize(commandInput); +console.log(commandInputTokenized); /** - * 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`. + * matchee pair of commandInputTokenized */ -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 commandTPair : TokenMatcheePair = {matched:[], + remained: commandInputTokenized}; + + +let tInt = m1TType(tk.TokenType.INT); +let tFlo = m1TType(tk.TokenType.FLO); +let tStr = m1TType(tk.TokenType.STR); +function tBool (x : TokenMatcheePair) :tk.Maybe { + let text = x.remained[0].text + if (text == "true" || text == "false"){ + return thenDo(toSome(x), m1TType(tk.TokenType.ID)); }else{ - return s.charCodeAt(0); + return {_tag : "None"}; } } /** - * @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`. + * 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. */ -export function thenDo(input : Maybe, f : Function) : Maybe{ - if (input._tag == "None"){ - return input; +function gramRHS (process: Function, arrange : number[]){ + return (m : TokenMatcheePair)=>{ + + let result : tk.Maybe = process(m); + console.log(`result ${result}`) + if (result._tag == "None"){ + return result; } else{ - let inner = input.value; - return f(inner); - } -} + let matched = result.value.matched; + let return_array : tkTree[] = Array(arrange.length); -/** - * @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 - */ -export function orDo(f1 : Function, f2: Function) : (x : T ) => Maybe{ - return (x) => { - let f1x : Maybe = (f1(x)); - { - if (f1x._tag == "None"){ - return f2(x); - } - else{ - return f1x; - } - } - }; -} - - -/** -* @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); - }; - - return wrapped_old_x; - }; -} + arrange.forEach((val, index) => { + return_array[arrange[index]] = matched[index]; + }); -/** -* @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); - - if (f_x._tag != "None"){ - return {_tag:"None"}; - }else{ - return thenDo(wrapped_x, matchAny); - } - }; + return return_array; + } + } } /** - * 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) + * CONST ::= INT | STR | FLO | BOOL */ -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); - - if (wrapped_new_x._tag != "None"){ - return wrapped_new_x; - }else{ - return wrapped_old_x; - } - }; -} - - -export function tokenize(input : string){ - var input_matchee_pair : Maybe = toSome( - {matched:"", - remained: input}); - - // 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)); - } - console.log(input+", result: "); - console.log(thenDo(input_matchee_pair, integer)); - // TODO: id, string, space, basic operator, 3 marks: @, {, }. - -} +var constParser = gramRHS((x : TokenMatcheePair)=> + {return thenDo(toSome(x),orDo(orDo(orDo(tInt,tFlo),tStr),tBool))}, [0]); -tokenize("+123"); -tokenize("123"); -tokenize("-123"); -tokenize(" 123"); -tokenize("c123"); +let tree = constParser(commandTPair); +console.log(util.inspect(tree, { showHidden: true, depth: null }));