micro-mdx-parser/src/utils/find-code.js

A tiny parser to convert markdown or html into JSON

const REP_SYMBOL = '◊'
const REP_SYMBOL_PATTERN = /◊/g
const ARROW_SYMBOL = '_≡►_'
const ARROW_SYMBOL_PATTERN = /_≡►_/g
// https://regex101.com/r/nIlW1U/6
const CODE_BLOCK_REGEX = /^([A-Za-z \t]*)```([A-Za-z]*)?\n([\s\S]*?)```([A-Za-z \t]*)*$/gm
// https://regex101.com/r/oPKKoC/1
const REMOVE_CODE_BLOCK_REGEX = /^(?:[A-Za-z \t]*)?(```(?:[A-Za-z]*)?\n(?:[\s\S]*?)```)([A-Za-z \t]*)*$/gm
// https://regex101.com/r/ZEcVR0/1 // results in timeouts...
// const CODE_INLINE_REGEX = /`([^\\`]*(?:(\\`*)|[^\\`]*)*)`/gm
// const CODE_INLINE_REGEX = /`([^`\n\r]*)`/gm
// const CODE_INLINE_REGEX = /`([^`][^`\n\r][^\\\s]*)`[^`]/gm
// const CODE_INLINE_REGEX = /`([^`][^`\n\r][^\s]*)`[^`]/gm
// const CODE_INLINE_REGEX = /`((?:\\`|[^`])*)`/gm
// const CODE_INLINE_REGEX = /[^\\\n]`((?:\\`|[^`])*)`/gm
// https://regex101.com/r/9bmDHe/6
// const CODE_INLINE_REGEX = /(^`((?:\\`|[^`])*)`|[^\\\n]`((?:\\`|[^`])*)`)/gm
// https://regex101.com/r/9bmDHe/7
const CODE_INLINE_REGEX = /(^`((?:\\`|[^`])*)+`|([^\\\n])`((?:\\`|[^`])*)`)/gm
const STARTS_WITH_FENCE = /^`{2,}/

/**
 * Parse code blocks out of markdown
 * @param {string} md - markdown string
 * @returns {Object}
 * @example
 * const blocks = findCodeBlocks(content)
 * console.log('blocks', blocks)
 */
function findCodeBlocks(md = '') {
  let matches
  let blocks = []
  while ((matches = CODE_BLOCK_REGEX.exec(md)) !== null) {
    if (matches.index === CODE_BLOCK_REGEX.lastIndex) {
      CODE_BLOCK_REGEX.lastIndex++ // avoid infinite loops with zero-width matches
    }
    const [ match, prefix, syntax, content, postFix ] = matches
    /* // debug
    console.log(`prefix: "${prefix}"`)
    console.log(`postFix: "${postFix}"`)
    console.log('syntax:', lang)
    console.log('Content:')
    console.log(content.trim())
    console.log('───────────────────────')
    /** */

    blocks.push({
      index: matches.index,
      syntax: syntax || '',
      block: match,
      code: content.trim()
    })
  }

  return blocks
}

/**
 * Parse inline code blocks out of markdown
 * @param {string} md - markdown string
 * @returns {Object}
 * @example
 * const blocks = findCodeBlocks(content)
 * console.log('blocks', blocks)
 */
function findInlineCode(md = '') {
  let matches
  let blocks = []
  while ((matches = CODE_INLINE_REGEX.exec(md)) !== null) {
    if (matches.index === CODE_INLINE_REGEX.lastIndex) {
      CODE_INLINE_REGEX.lastIndex++ // avoid infinite loops with zero-width matches
    }
    const [ match, _fullMatch, innerOne, space, innerTwo ] = matches
    const content = innerOne || innerTwo
    /* // debug
    console.log(matches)
    console.log('content', content)
    console.log('───────────────────────')
    /** */
    if (content && (content.indexOf('\n') === -1) && !STARTS_WITH_FENCE.test(content)) {
      blocks.push({
        index: (space) ? matches.index + 1 : matches.index,
        block: (space) ? match.substring(1) : match,
        code: content.trim()
      })
    }
  }

  return blocks
}

/**
 * Parse all code blocks out of markdown
 * @param {string} md - markdown string
 * @returns {Object}
 * @example
 * const blocks = findCodeBlocks(content)
 * console.log('blocks', blocks)
 */
function findCode(md = '') {
  // const blocks = findCodeBlocks(md)
  const inline = findInlineCode(md)

  return inline
  /*
  console.log('blocks', blocks)
  console.log('inline', inline)
  process.exit(1)
  /** */
  return blocks.concat(inline)
    .sort((a, b) => {
    if (a.index > b.index) return 1
    if (a.index < b.index) return -1
    return 0
  })
}

function fixOpenBracket(content = '') {
  return (content.indexOf(REP_SYMBOL) === -1) ? content : content.replace(REP_SYMBOL_PATTERN, '<')
}

module.exports = {
  findCode,
  findCodeBlocks,
  findInlineCode,
  fixOpenBracket,
  REP_SYMBOL,
  REP_SYMBOL_PATTERN,
  ARROW_SYMBOL,
  ARROW_SYMBOL_PATTERN,
  CODE_BLOCK_REGEX,
  REMOVE_CODE_BLOCK_REGEX
}