now-updater/utils.js

import 'dotenv/config'
import { JSDOM } from 'jsdom'
import sanitizeHtml from 'sanitize-html'

// ========================================================================
// helper functions
// ========================================================================

// regex to match service name within an html anchor tag
// used when linkFormat is 'html' in config.js
export function htmlServiceLinkRegex(serviceName) {
	return new RegExp(
		`(<a[^>]+href=["'][^"']*["'][^>]*>)${serviceName}(<\\/a>)`,
		'i',
	)
}

// regex to match service name within a markdown link
// used when linkFormat is 'markdown' in config.js
export function markdownServiceLinkRegex(serviceName) {
	return new RegExp(`(\\[)${serviceName}(\\]\\([^)]*\\))`, 'i')
}

// escapes special markdown characters so they display literally
export function markdownCharEscape(text) {
	return text
		.replaceAll('`', '\\`')
		.replaceAll('*', '\\*')
		.replaceAll('_', '\\_')
		.replaceAll('{', '\\{')
		.replaceAll('}', '\\}')
		.replaceAll('[', '\\[')
		.replaceAll(']', '\\]')
		.replaceAll('(', '\\(')
		.replaceAll(')', '\\)')
		.replaceAll('#', '\\#')
		.replaceAll('+', '\\+')
		.replaceAll('-', '\\-')
		.replaceAll('.', '\\.')
		.replaceAll('!', '\\!')
}

// ========================================================================
// generic feed parsers
// ========================================================================

// rss feed parser - grabs the latest item from any standard rss feed
export async function getRSSItemTitle(feedUrl) {
	const res = await fetch(feedUrl)
	const data = await res.text()
	const dom = new JSDOM(data, { contentType: 'text/xml' })
	const item = dom.window.document.querySelector('item')
	const title = item.querySelector('title').textContent
	const link = item.querySelector('link').textContent.trim()
	const cleanTitle = sanitizeHtml(title, {
		allowedTags: [],
		allowedAttributes: {},
	})?.trim()
	return { text: cleanTitle, url: link }
}

// atom feed parser - grabs the latest entry from any standard atom feed
export async function getAtomFeed(url) {
	const res = await fetch(url)
	const data = await res.text()
	const dom = new JSDOM(data, { contentType: 'text/xml' })
	const entry = dom.window.document.querySelector('entry')
	if (!entry) {
		throw new Error('No recent items found')
	}
	const title = entry.querySelector('title').textContent
	const link = entry.querySelector('link').getAttribute('href')
	const cleanTitle = sanitizeHtml(title, {
		allowedTags: [],
		allowedAttributes: {},
	})?.trim()
	return { text: cleanTitle, url: link }
}

// json feed parser - grabs the latest item title
// optionally includes images separately
export async function getJsonFeedItemTitle(feedUrl, showImage = true) {
	const res = await fetch(feedUrl)
	const data = await res.json()
	const post = data.items[0]
	const { title, image, url } = post
	return { text: title, url, image: showImage ? image : null }
}

// ========================================================================
// music
// ========================================================================

// listenbrainz json api - grabs your most recent listen
// returns "track name<br />by artist name"
export async function getListenBrainzScrobble(config) {
	const { feedUrl, userId } = config

	const res = await fetch(feedUrl)
	const data = await res.json()
	if (!data.payload.listens || data.payload.listens.length === 0) {
		throw new Error('No recent listens found')
	}
	const listen = data.payload.listens[0]
	const { track_name, artist_name } = listen.track_metadata
	const text = `${track_name}<br />by ${artist_name}`
	const url = `https://listenbrainz.org/user/${userId}/`
	return { text, url }
}

// maloja music scrobbling api - grabs your most recent listen
// returns "track name<br />by artist name"
// not configured in services.js by default but kept from upstream
export async function getMalojaScrobble(malojaUrl) {
	const res = await fetch(`${malojaUrl}/apis/mlj_1/scrobbles?perpage=1`)
	const data = await res.json()
	const scrobble = data.list[0]
	const {
		title,
		album: { artists },
	} = scrobble.track
	return `${title}<br />by ${artists.join(', ')}`
}

// ========================================================================
// social media
// ========================================================================

// mastodon rss feed parser - filters out replies (posts starting with @)
// also replaces urls with 🔗 emoji to keep things tidy
// extracts images from posts if available
export async function getMastodonPost(config) {
	const { feedUrl, showImage = false } =
		typeof config === 'string' ? { feedUrl: config } : config
	const res = await fetch(feedUrl)
	const data = await res.text()
	const dom = new JSDOM(data, { contentType: 'text/xml' })

	// get all items from the rss feed
	const items = dom.window.document.querySelectorAll('item')

	// find the first item that isn't a reply (doesn't start with @)
	for (const item of items) {
		const description = item.querySelector('description')?.textContent || ''

		// skip if it is a reply
		if (description.trim().startsWith('@')) {
			continue
		}

		const link = item.querySelector('link')?.textContent.trim()

		// extract image from description html if available
		const imgMatch = description.match(/<img[^>]+src="([^"]+)"/)
		const image = imgMatch ? imgMatch[1] : null

		// use description as the text, stripping HTML
		let cleanText = sanitizeHtml(description, {
			allowedTags: [],
			allowedAttributes: {},
		})?.trim()

		// replace URLs with a link emoji
		cleanText = cleanText.replace(/https?:\/\/\S+/g, '🔗').trim()

		return { text: cleanText, url: link, image: showImage ? image : null }
	}

	// fallback to first item if no non-reply found
	const firstItem = items[0]
	const description = firstItem.querySelector('description')?.textContent || ''
	const link = firstItem.querySelector('link')?.textContent.trim()

	// extract image from description html if available
	const imgMatch = description.match(/<img[^>]+src="([^"]+)"/)
	const image = imgMatch ? imgMatch[1] : null

	let cleanText = sanitizeHtml(description, {
		allowedTags: [],
		allowedAttributes: {},
	})?.trim()
	// replace URLs with link emoji
	cleanText = cleanText.replace(/https?:\/\/\S+/g, '🔗').trim()

	return { text: cleanText, url: link, image: showImage ? image : null }
}

// ========================================================================
// movies & tv
// ========================================================================

// letterboxd rss parser - can optionally extract poster images from cdata
// images are embedded in the description field as html
export async function getLetterboxdActivity(config) {
	const { feedUrl, showImage = false } = config
	const res = await fetch(feedUrl)
	const data = await res.text()
	const dom = new JSDOM(data, { contentType: 'text/xml' })
	const item = dom.window.document.querySelector('item')
	const title = item.querySelector('title').textContent
	const link = item.querySelector('link').textContent.trim()

	let image = null
	if (showImage) {
		// Parse CDATA content as HTML to extract image
		const description = item.querySelector('description').textContent
		const imgMatch = description.match(/<img[^>]+src="([^"]+)"/)
		image = imgMatch ? imgMatch[1] : null
	}

	return { text: title, url: link, image }
}

// trakt episode history - requires TRAKT_SLURM variable (see readme for info)
// can optionally include trakt widget thumbnail image
export async function getTraktEpisode(config) {
	const { userId, traktId, showImage = false } = config
	const res = await fetch(
		`https://trakt.tv/users/${userId}/history/episodes/added/asc.atom?slurm=${process.env.TRAKT_SLURM}`,
	)
	const data = await res.text()
	const dom = new JSDOM(data)
	const entry = dom.window.document.querySelector('entry')
	const title = entry.querySelector('title').textContent
	const link = entry.querySelector('link').getAttribute('href')
	const image = showImage
		? `https://widgets.trakt.tv/users/${traktId}/watched/thumb@2x.jpg?type=episode&image_only=1`
		: null
	return { text: title, url: link, image }
}

// trakt movie history - requires TRAKT_SLURM variable (see readme for info)
// can optionally include trakt widget thumbnail image
export async function getTraktMovie(config) {
	const { userId, traktId, showImage = false } = config
	const res = await fetch(
		`https://trakt.tv/users/${userId}/history/movies/added/asc.atom?slurm=${process.env.TRAKT_SLURM}`,
	)
	const data = await res.text()
	const dom = new JSDOM(data)
	const entry = dom.window.document.querySelector('entry')
	const title = entry.querySelector('title').textContent
	const link = entry.querySelector('link').getAttribute('href')
	const image = showImage
		? `https://widgets.trakt.tv/users/${traktId}/watched/thumb@2x.jpg?type=movie&image_only=1`
		: null
	return { text: title, url: link, image }
}

// ========================================================================
// books
// ========================================================================

// hardcover graphql api - grabs your currently reading book
export async function getHardcoverActivity(config) {
	const { userId, feedUrl } = config
	const query = `
		query {
			user(username: "${userId}") {
				currently_reading_books(limit: 1) {
					title
					id
				}
			}
		}
	`

	const res = await fetch(feedUrl, {
		method: 'POST',
		headers: { 'Content-Type': 'application/json' },
		body: JSON.stringify({ query }),
	})

	const data = await res.json()
	if (
		!data.data?.user?.currently_reading_books ||
		data.data.user.currently_reading_books.length === 0
	) {
		throw new Error('Nothing!')
	}

	const book = data.data.user.currently_reading_books[0]
	return {
		text: book.title,
		url: `https://hardcover.app/books/${book.id}`,
	}
}

// ========================================================================
// gaming
// ========================================================================

// steam web api - grabs your most recently played game
// requires STEAM_WEBAPI_KEY see readme for more info
// needs your steam id 64 (not username) - convert at steamid.io
export async function getSteamRecentlyPlayed(config) {
	const { steamId } = config
	const apiKey = process.env.STEAM_WEBAPI_KEY
	if (!apiKey) {
		throw new Error('STEAM_WEBAPI_KEY environment variable not set')
	}

	const res = await fetch(
		`https://api.steampowered.com/IPlayerService/GetRecentlyPlayedGames/v1/?key=${process.env.STEAM_WEBAPI_KEY}&steamid=${steamId}&format=json`,
	)
	const data = await res.json()

	if (!data.response.games || data.response.games.length === 0) {
		throw new Error('No recent games found')
	}

	const game = data.response.games[0]
	const { name, appid } = game

	return {
		text: name,
		url: `https://store.steampowered.com/app/${appid}`,
	}
}

// ========================================================================
// code & development
// ========================================================================

// source.tube (forgejo) rss parser - strips username from the beginning
// turns "dylan pushed an update to..." into "pushed an update to..."
export async function getSourceTubeActivity(config) {
	const { feedUrl, userId } = config

	const res = await fetch(feedUrl)
	const data = await res.text()
	const dom = new JSDOM(data, { contentType: 'text/xml' })
	const item = dom.window.document.querySelector('item')
	if (!item) {
		throw new Error('No recent activity found')
	}
	const title = item.querySelector('title').textContent
	const link = item.querySelector('link').textContent.trim()
	let cleanTitle = sanitizeHtml(title, {
		allowedTags: [],
		allowedAttributes: {},
	})?.trim()

	// remove username prefix (e.g., "dylan pushed..." -> "pushed...")
	if (userId) {
		const usernamePattern = new RegExp(`^${userId}\\s+`, 'i')
		cleanTitle = cleanTitle.replace(usernamePattern, '')
	}

	return { text: cleanTitle, url: link }
}