curlconverter/src/shell/Word.ts

convert curl commands to Python, JavaScript, Go, PHP and more

// Words act like strings. They're lists of characters, except some
// characters can be shell variables or expressions.
// They're implemented like this:
// ["foobar", {type: "variable", value: "baz", text: "$baz"}, "qux"]
// Except for the empty string [""], there should be no empty strings in the array.

import type { Parser } from "./Parser.js";
import { CCError } from "../utils.js";

export interface ShellToken {
  type: "variable" | "command";
  value: string;
  text: string;
  syntaxNode: Parser.SyntaxNode; // For error reporting
}
export type Token = string | ShellToken;

// TODO: Words should keep a list of operations that happened to them
// like .replace() so that we can generate code that also does that operation
// on the contents of the environment variable or the output of the command.
export class Word implements Iterable<Token> {
  readonly tokens: Token[];

  constructor(tokens?: string);
  constructor(tokens?: Token[]);
  constructor(tokens?: string | Token[]) {
    if (typeof tokens === "string") {
      tokens = [tokens];
    }
    if (tokens === undefined || tokens.length === 0) {
      tokens = [""];
    }

    this.tokens = [];
    for (const t of tokens) {
      if (typeof t === "string") {
        if (
          this.tokens.length > 0 &&
          typeof this.tokens[this.tokens.length - 1] === "string"
        ) {
          // If we have 2+ strings in a row, merge them
          this.tokens[this.tokens.length - 1] += t;
        } else if (t) {
          // skip empty strings
          this.tokens.push(t);
        }
      } else {
        this.tokens.push(t);
      }
    }
    if (this.tokens.length === 0) {
      this.tokens.push("");
    }
  }

  get length(): number {
    let len = 0;
    for (const t of this.tokens) {
      if (typeof t === "string") {
        len += t.length;
      } else {
        len += 1;
      }
    }
    return len;
  }

  *[Symbol.iterator](): Iterator<Token> {
    for (const t of this.tokens) {
      if (typeof t === "string") {
        for (const c of t) {
          yield c;
        }
      } else {
        yield t;
      }
    }
  }

  // TODO: do we need this function?
  get(index: number): Token {
    let i = 0;
    for (const t of this.tokens) {
      if (typeof t === "string") {
        if (i + t.length > index) {
          return t[index - i];
        }
        i += t.length;
      } else {
        if (i === index) {
          return t;
        }
        i += 1;
      }
    }
    throw new CCError("Index out of bounds");
  }

  charAt(index = 0): Token {
    try {
      return this.get(index);
    } catch {}
    return "";
  }

  indexOf(search: string, start?: number): number {
    if (start === undefined) {
      start = 0;
    }
    let i = 0;
    for (const t of this.tokens) {
      if (typeof t === "string") {
        if (i + t.length > start) {
          const index = t.indexOf(search, start - i);
          if (index !== -1) {
            return i + index;
          }
        }
        i += t.length;
      } else {
        i += 1;
      }
    }
    return -1;
  }

  // Like indexOf() but accepts a string of characters and returns the index of the first one
  // it finds
  indexOfFirstChar(search: string): number {
    let i = 0;
    for (const t of this.tokens) {
      if (typeof t === "string") {
        for (const c of t) {
          if (search.includes(c)) {
            return i;
          }
          i += 1;
        }
      } else {
        i += 1;
      }
    }
    return -1;
  }

  removeFirstChar(c: string): Word {
    if (this.length === 0) {
      return new Word();
    }
    if (this.charAt(0) === c) {
      return this.slice(1);
    }
    return this.copy();
  }

  copy(): Word {
    return new Word(this.tokens);
  }

  slice(indexStart?: number, indexEnd?: number): Word {
    if (indexStart === undefined) {
      indexStart = this.length;
    }
    if (indexEnd === undefined) {
      indexEnd = this.length;
    }
    if (indexStart >= this.length) {
      return new Word();
    }
    if (indexStart < 0) {
      indexStart = Math.max(indexStart + this.length, 0);
    }
    if (indexEnd < 0) {
      indexEnd = Math.max(indexEnd + this.length, 0);
    }
    if (indexEnd <= indexStart) {
      return new Word();
    }

    const ret = [];
    let i = 0;
    for (const t of this.tokens) {
      if (typeof t === "string") {
        if (i + t.length > indexStart) {
          if (i < indexEnd) {
            ret.push(t.slice(Math.max(indexStart - i, 0), indexEnd - i));
          }
        }
        i += t.length;
      } else {
        if (i >= indexStart && i < indexEnd) {
          ret.push(t);
        }
        i += 1;
      }
    }
    return new Word(ret);
  }

  // TODO: check
  includes(search: string, start?: number): boolean {
    if (start === undefined) {
      start = 0;
    }
    let i = 0;
    for (const t of this.tokens) {
      if (typeof t === "string") {
        if (i + t.length > start) {
          if (t.includes(search, start - i)) {
            return true;
          }
        }
        i += t.length;
      } else {
        i += 1;
      }
    }
    return false;
  }

  test(search: RegExp): boolean {
    for (const t of this.tokens) {
      if (typeof t === "string") {
        if (search.test(t)) {
          return true;
        }
      }
    }
    return false;
  }

  prepend(c: string): Word {
    const ret = this.copy();
    if (ret.tokens.length && typeof ret.tokens[0] === "string") {
      ret.tokens[0] = c + ret.tokens[0];
    } else {
      ret.tokens.unshift(c);
    }
    return ret;
  }

  append(c: string): Word {
    const ret = this.copy();
    if (
      ret.tokens.length &&
      typeof ret.tokens[ret.tokens.length - 1] === "string"
    ) {
      ret.tokens[ret.tokens.length - 1] += c;
    } else {
      ret.tokens.push(c);
    }
    return ret;
  }

  // Merges two Words
  add(other: Word): Word {
    return new Word([...this.tokens, ...other.tokens]);
  }

  // Returns the first match, searches each string independently
  // TODO: improve this
  match(regex: RegExp): RegExpMatchArray | null {
    for (const t of this.tokens) {
      if (typeof t === "string") {
        const match = t.match(regex);
        if (match) {
          return match;
        }
      }
    }
    return null;
  }

  search(regex: RegExp): number {
    let offset = 0;
    for (const t of this.tokens) {
      if (typeof t === "string") {
        const match = t.search(regex);
        if (match !== -1) {
          return offset + match;
        }
        offset += t.length;
      }
    }
    return -1;
  }

  // .replace() is called per-string, so it won't work through shell variables
  replace(search: string | RegExp, replacement: string): Word {
    const ret: Token[] = [];
    for (const t of this.tokens) {
      if (typeof t === "string") {
        ret.push(t.replace(search, replacement));
      } else {
        ret.push(t);
      }
    }
    return new Word(ret);
  }

  // splits correctly, not like String.split()
  // The last entry can contain the separator if limit entries has been reached
  split(separator: string, limit?: number): Word[] {
    const ret: Word[] = [];
    let i = 0;
    let start = 0;
    while (i < this.length) {
      let match = true;
      for (let j = 0; j < separator.length; j++) {
        if (this.get(i + j) !== separator.charAt(j)) {
          match = false;
          break;
        }
      }
      if (match) {
        ret.push(this.slice(start, i));
        i += separator.length;
        start = i;
        if (limit !== undefined && ret.length === limit - 1) {
          break;
        }
      } else {
        i += 1;
      }
    }
    if (start <= this.length) {
      ret.push(this.slice(start));
    }
    return ret;
  }

  toLowerCase(): Word {
    return new Word(
      this.tokens.map((t) => (typeof t === "string" ? t.toLowerCase() : t)),
    );
  }

  toUpperCase(): Word {
    return new Word(
      this.tokens.map((t) => (typeof t === "string" ? t.toUpperCase() : t)),
    );
  }

  trimStart(): Word {
    const ret: Token[] = [];
    let i, t;
    for ([i, t] of this.tokens.entries()) {
      if (typeof t === "string") {
        if (i === 0) {
          t = t.trimStart();
        }
        if (t) {
          ret.push(t);
        }
      } else {
        ret.push(t);
      }
    }
    if (ret.length === 0) {
      return new Word();
    }
    return new Word(ret);
  }

  trimEnd(): Word {
    const ret: Token[] = [];
    let i, t;
    for ([i, t] of this.tokens.entries()) {
      if (typeof t === "string") {
        if (i === this.tokens.length - 1) {
          t = t.trimEnd();
        }
        if (t) {
          ret.push(t);
        }
      } else {
        ret.push(t);
      }
    }
    if (ret.length === 0) {
      return new Word();
    }
    return new Word(ret);
  }

  trim(): Word {
    const ret: Token[] = [];
    let i, t;
    for ([i, t] of this.tokens.entries()) {
      if (typeof t === "string") {
        if (i === 0) {
          t = t.trimStart();
        }
        if (i === this.tokens.length - 1) {
          t = t.trimEnd();
        }
        if (t) {
          ret.push(t);
        }
      } else {
        ret.push(t);
      }
    }
    if (ret.length === 0) {
      return new Word();
    }
    return new Word(ret);
  }

  isEmpty(): boolean {
    if (this.tokens.length === 0) {
      return true;
    }
    if (this.tokens.length === 1 && typeof this.tokens[0] === "string") {
      return this.tokens[0].length === 0;
    }
    return false;
  }

  toBool(): boolean {
    return !this.isEmpty();
  }

  // Returns true if .tokens contains no variables/commands
  isString(): boolean {
    for (const t of this.tokens) {
      if (typeof t !== "string") {
        return false;
      }
    }
    return true;
  }

  firstShellToken(): ShellToken | null {
    for (const t of this.tokens) {
      if (typeof t !== "string") {
        return t;
      }
    }
    return null;
  }

  startsWith(prefix: string): boolean {
    if (this.tokens.length === 0) {
      return false;
    }
    if (typeof this.tokens[0] === "string") {
      return this.tokens[0].startsWith(prefix);
    }
    return false;
  }

  endsWith(suffix: string): boolean {
    if (this.tokens.length === 0) {
      return false;
    }
    const lastToken = this.tokens[this.tokens.length - 1];
    if (typeof lastToken === "string") {
      return lastToken.endsWith(suffix);
    }
    return false;
  }

  // This destroys the information about the original tokenization
  toString() {
    return this.tokens
      .map((t) => (typeof t === "string" ? t : t.text))
      .join("");
  }
  valueOf = Word.toString;
}

export function eq(
  it: Word | undefined | null,
  other: string | Word | undefined | null,
): boolean {
  if (
    it === undefined ||
    it === null ||
    other === undefined ||
    other === null
  ) {
    return it === other;
  }
  if (typeof other === "string") {
    return (
      it.tokens.length === 1 &&
      typeof it.tokens[0] === "string" &&
      it.tokens[0] === other
    );
  }
  return (
    it.tokens.length === other.tokens.length &&
    it.tokens.every((itToken, i) => {
      const otherToken = other.tokens[i];
      if (typeof itToken === "string") {
        return itToken === otherToken;
      } else if (typeof otherToken !== "string") {
        return itToken.text === otherToken.text;
      }
      return false;
    })
  );
}
export function firstShellToken(word: string | Word): ShellToken | null {
  if (typeof word === "string") {
    return null;
  }
  return word.firstShellToken();
}
export function mergeWords(...words: (Word | string)[]): Word {
  const ret: Token[] = [];
  for (const w of words) {
    if (w instanceof Word) {
      ret.push(...w.tokens);
    } else {
      ret.push(w);
    }
  }
  return new Word(ret);
}
export function joinWords(words: Word[], joinChar: string): Word {
  const ret: Token[] = [];
  for (const w of words) {
    if (ret.length) {
      ret.push(joinChar);
    }
    ret.push(...w.tokens);
  }
  return new Word(ret);
}