parser.js

parser.js

'use strict'

const sprintf = require( 'sprintf-js' ).sprintf
const moment = require( 'moment' )

const unicode_generalCategory = require( 'unicode-13.0.0/General_Category' )

const charCategoryMultipliers = [
  0, // none
  1, // latin letters
  2  // other letters (japanese etc.)
]

const baseXPPerLetter = 0.11
const perMessageBaseXP = 0.65
const minLettersInMsg = 5

class ParsedText
{
  get type() { return 'text' }
  get text() { return this._text }
  constructor( text )
  {
    this._text = text
  }
}

class ParsedMention
{
  get type() { return 'mention' }
  get targetType() { return this._type }
  get target() { return this._target }
  constructor( type, target )
  {
    this._type = type
    this._target = target
  }
}

class ParsedEmote
{
  get type() { return 'emote' }
  get name() { return this._name }
  get code() { return this._code }
  constructor( name, code )
  {
    this._name = name
    this._code = code
  }
}

module.exports = class Parser
{
  constructor()
  {
  }
  parseMessage( fulltext )
  {
    logSprintf( 'parse', 'Parsing:' )
    console.log( fulltext )
    let i = 0
    let letters = 0
    let parsed = {
      xp: 0,
      parts: []
    }
    let buffer = ''
    let skipNext = false
    while ( i < fulltext.length )
    {
      let multiplier = 0
      const char = fulltext[i]
      if ( char === '\\' && !skipNext )
      {
        // \ to escape special handling of next character
        skipNext = true
        i++
        continue
      }
      if ( !skipNext )
      {
        if ( char === '"' && ( i + 2 < fulltext.length ) )
        {
          // quoted text (copy as-is until end)
          let close = -1
          let spos = i + 1
          while ( spos && spos < fulltext.length )
          {
            close = fulltext.indexOf( '"', spos )
            if ( ( close && fulltext[close - 1] !== '\\' ) || !close )
              spos = 0
            else if ( close )
              spos = close + 1
          }
          if ( close > i )
          {
            if ( buffer.length )
              parsed.parts.push( new ParsedText( buffer ) )
            buffer = fulltext.substr( i + 1, close - i - 1 )
            parsed.parts.push( new ParsedText( buffer ) )
            buffer = ''
            i = close + 1
            continue
          }
        }
        else if ( char === '<' && ( i + 5 < fulltext.length ) )
        {
          const close = fulltext.indexOf( '>', i + 1 )
          if ( ( fulltext[i + 1] === ':' || ( fulltext[i + 1] === 'a' && fulltext[i + 2] === ':' ) ) && close > i )
          {
            // emote
            if ( fulltext[i + 1] === 'a' )
              i++
            const emoteClose = fulltext.indexOf( ':', i + 2 )
            if ( emoteClose > i && emoteClose < close )
            {
              const emoteName = fulltext.substr( i + 1, emoteClose - i )
              const emoteNumber = fulltext.substr( emoteClose + 1, close - emoteClose - 1 )
              if ( /^\d+$/.test( emoteNumber ) )
              {
                logSprintf( 'parse', 'emote: %s (%s)', emoteName, emoteNumber )
                if ( buffer.length )
                  parsed.parts.push( new ParsedText( buffer ) )
                buffer = ''
                parsed.parts.push( new ParsedEmote( emoteName, emoteNumber ) )
                i = close + 1
                continue
              }
            }
          }
          else if ( ( fulltext[i + 1] === '@' || fulltext[i + 1] === '#' ) && close > i )
          {
            // mention (user, role, channel)
            const mentionType = ( fulltext[i + 2] === '&' ? 'role' : fulltext[i + 1] === '#' ? 'channel' : 'user' )
            const j = ( ( mentionType === 'role' || fulltext[i + 2] === '!' ) ? i + 2 : i + 1 )
            const mentionNumber = fulltext.substr( j + 1, close - j - 1 )
            if ( /^\d+$/.test( mentionNumber ) )
            {
              logSprintf( 'parse', 'mention: %s (%s)', mentionNumber, mentionType )
              if ( buffer.length )
                parsed.parts.push( new ParsedText( buffer ) )
              buffer = ''
              parsed.parts.push( new ParsedMention( mentionType, mentionNumber ) )
              i = close + 1
              continue
            }
          }
        }
      }
      const category = unicode_generalCategory.get( fulltext.codePointAt( i ) )
      if ( !skipNext && category.indexOf( 'Separator' ) >= 0 )
      {
        if ( buffer.length )
          parsed.parts.push( new ParsedText( buffer ) )
        buffer = ''
      }
      else
      {
        skipNext = false
        buffer += fulltext[i]
        if ( ['Lowercase_Letter', 'Uppercase_Letter', 'Number'].includes( category ) )
        {
          multiplier = charCategoryMultipliers[1]
          letters += 1
        }
        else if ( ['Other_Letter'].includes( category ) )
        {
          multiplier = charCategoryMultipliers[2]
          letters += 2
        }
        parsed.xp += ( baseXPPerLetter * multiplier )
      }
      i++
    }
    if ( buffer.length )
      parsed.parts.push( new ParsedText( buffer ) )
    parsed.xp = ( letters >= minLettersInMsg ? ( perMessageBaseXP + parsed.xp ) : false )
    return parsed
  }
}