epg/sites/programtv.onet.pl/programtv.onet.pl.config.js

const axios = require('axios')
const cheerio = require('cheerio')
const dayjs = require('dayjs')
const utc = require('dayjs/plugin/utc')
const timezone = require('dayjs/plugin/timezone')
dayjs.extend(utc)
dayjs.extend(timezone)

module.exports = {
  delay: 5000,
  site: 'programtv.onet.pl',
  days: 2,
  url: function ({ date, channel }) {
    const currDate = dayjs().utc().startOf('day')
    const day = date.diff(currDate, 'd')

    return `https://programtv.onet.pl/program-tv/${channel.site_id}?dzien=${day}`
  },
  async parser({ content, date }) {
    const programs = []
    const items = parseItems(content)
    for (const item of items) {
      const prev = programs[programs.length - 1]
      const $item = cheerio.load(item)
      let start = parseStart($item, date)
      if (prev) {
        if (start < prev.start) {
          start = start.add(1, 'day')
          date = date.add(1, 'day')
        }
        prev.stop = start
      }
      const stop = start.add(1, 'hour')

      const programUrl = parseProgramUrl($item)
      const details = await loadProgramDetails(programUrl)
      let image
      if (details) {
        image = details.image
      }

      programs.push({
        title: parseTitle($item),
        description: parseDescription($item),
        image,
        category: parseCategory($item),
        start,
        stop
      })
    }

    return programs
  },
  async channels() {
    const data = await axios
      .get('https://programtv.onet.pl/stacje')
      .then(r => r.data)
      .catch(console.log)

    let channels = []

    const $ = cheerio.load(data)
    $('ul.channelList a').each((i, el) => {
      const name = $(el).text()
      const url = $(el).attr('href')
      const [, site_id] = url.match(/^\/program-tv\/(.*)$/i)

      channels.push({
        lang: 'pl',
        site_id,
        name
      })
    })

    return channels
  }
}

async function loadProgramDetails(url) {
  const html = await axios
    .get(url)
    .then(r => r.data)
    .catch(console.error)
  if (!html) return

  const $ = cheerio.load(html)

  return {
    image: $('meta[property="og:image"]').attr('content')
  }
}

function parseProgramUrl($item) {
  const href = $item('.titles > a').attr('href')
  const url = new URL(href, 'https://programtv.onet.pl')

  return url.href
}

function parseStart($item, date) {
  const timeString = $item('.hours > .hour').text()
  const dateString = `${date.format('MM/DD/YYYY')} ${timeString}`

  return dayjs.tz(dateString, 'MM/DD/YYYY HH:mm', 'Europe/Warsaw').utc()
}

function parseCategory($item) {
  return $item('.titles > .type').text()
}

function parseDescription($item) {
  return $item('.titles > p').text().trim()
}

function parseTitle($item) {
  return $item('.titles > a').text().trim()
}

function parseItems(content) {
  const $ = cheerio.load(content)

  return $('#channelTV > section > div.emissions > ul > li').toArray()
}