Merge branch 'advplyr:master' into audible-confidence-score

server/utils/ffmpegHelpers.js

@@ -103,18 +103,39 @@ module.exports.resizeImage = resizeImage
  */
 module.exports.downloadPodcastEpisode = (podcastEpisodeDownload) => {
   return new Promise(async (resolve) => {
-    const response = await axios({
-      url: podcastEpisodeDownload.url,
-      method: 'GET',
-      responseType: 'stream',
-      headers: {
-        'User-Agent': 'audiobookshelf (+https://audiobookshelf.org)'
-      },
-      timeout: global.PodcastDownloadTimeout
-    }).catch((error) => {
-      Logger.error(`[ffmpegHelpers] Failed to download podcast episode with url "${podcastEpisodeDownload.url}"`, error)
-      return null
-    })
+    // Some podcasts fail due to user agent strings
+    // See: https://github.com/advplyr/audiobookshelf/issues/3246 (requires iTMS user agent)
+    // See: https://github.com/advplyr/audiobookshelf/issues/4401 (requires no iTMS user agent)
+    const userAgents = ['audiobookshelf (+https://audiobookshelf.org; like iTMS)', 'audiobookshelf (+https://audiobookshelf.org)']
+
+    let response = null
+    let lastError = null
+
+    for (const userAgent of userAgents) {
+      try {
+        response = await axios({
+          url: podcastEpisodeDownload.url,
+          method: 'GET',
+          responseType: 'stream',
+          headers: {
+            'User-Agent': userAgent
+          },
+          timeout: global.PodcastDownloadTimeout
+        })
+
+        Logger.debug(`[ffmpegHelpers] Successfully connected with User-Agent: ${userAgent}`)
+        break
+      } catch (error) {
+        lastError = error
+        Logger.warn(`[ffmpegHelpers] Failed to download podcast episode with User-Agent "${userAgent}" for url "${podcastEpisodeDownload.url}"`, error.message)
+
+        // If this is the last attempt, log the full error
+        if (userAgent === userAgents[userAgents.length - 1]) {
+          Logger.error(`[ffmpegHelpers] All User-Agent attempts failed for url "${podcastEpisodeDownload.url}"`, lastError)
+        }
+      }
+    }
+
     if (!response) {
       return resolve({
         success: false

server/utils/notifications.js

@@ -60,6 +60,38 @@ module.exports.notificationData = {
         errorMsg: 'Example error message'
       }
     },
+    {
+      name: 'onRSSFeedFailed',
+      requiresLibrary: true,
+      description: 'Triggered when the RSS feed request fails for an automatic episode download',
+      descriptionKey: 'NotificationOnRSSFeedFailedDescription',
+      variables: ['feedUrl', 'numFailed', 'title'],
+      defaults: {
+        title: 'RSS Feed Request Failed',
+        body: 'Failed to request RSS feed for {{title}}.\nFeed URL: {{feedUrl}}\nNumber of failed attempts: {{numFailed}}'
+      },
+      testData: {
+        title: 'Test RSS Feed',
+        feedUrl: 'https://example.com/rss',
+        numFailed: 3
+      }
+    },
+    {
+      name: 'onRSSFeedDisabled',
+      requiresLibrary: true,
+      description: 'Triggered when automatic episode downloads are disabled due to too many failed attempts',
+      descriptionKey: 'NotificationOnRSSFeedDisabledDescription',
+      variables: ['feedUrl', 'numFailed', 'title'],
+      defaults: {
+        title: 'Podcast Episode Download Schedule Disabled',
+        body: 'Automatic episode downloads for {{title}} have been disabled due to too many failed RSS feed requests.\nFeed URL: {{feedUrl}}\nNumber of failed attempts: {{numFailed}}'
+      },
+      testData: {
+        title: 'Test RSS Feed',
+        feedUrl: 'https://example.com/rss',
+        numFailed: 5
+      }
+    },
     {
       name: 'onTest',
       requiresLibrary: false,

server/utils/podcastUtils.js

@@ -1,8 +1,9 @@
 const axios = require('axios')
 const ssrfFilter = require('ssrf-req-filter')
 const Logger = require('../Logger')
-const { xmlToJSON, levenshteinDistance, timestampToSeconds } = require('./index')
+const { xmlToJSON, timestampToSeconds } = require('./index')
 const htmlSanitizer = require('../utils/htmlSanitizer')
+const Fuse = require('../libs/fusejs')
 
 /**
  * @typedef RssPodcastChapter
@@ -205,7 +206,7 @@ function extractEpisodeData(item) {
     } else if (typeof guidItem?._ === 'string') {
       episode.guid = guidItem._
     } else {
-      Logger.error(`[podcastUtils] Invalid guid ${item['guid']} for ${episode.enclosure.url}`)
+      Logger.error(`[podcastUtils] Invalid guid for ${episode.enclosure.url}`, item['guid'])
     }
   }
 
@@ -407,7 +408,7 @@ module.exports.getPodcastFeed = (feedUrl, excludeEpisodeMetadata = false) => {
     })
 }
 
-// Return array of episodes ordered by closest match (Levenshtein distance of 6 or less)
+// Return array of episodes ordered by closest match using fuse.js
 module.exports.findMatchingEpisodes = async (feedUrl, searchTitle) => {
   const feed = await this.getPodcastFeed(feedUrl).catch(() => {
     return null
@@ -420,32 +421,29 @@ module.exports.findMatchingEpisodes = async (feedUrl, searchTitle) => {
  *
  * @param {RssPodcast} feed
  * @param {string} searchTitle
- * @returns {Array<{ episode: RssPodcastEpisode, levenshtein: number }>}
+ * @param {number} [threshold=0.4] - 0.0 for perfect match, 1.0 for match anything
+ * @returns {Array<{ episode: RssPodcastEpisode }>}
  */
-module.exports.findMatchingEpisodesInFeed = (feed, searchTitle) => {
-  searchTitle = searchTitle.toLowerCase().trim()
+module.exports.findMatchingEpisodesInFeed = (feed, searchTitle, threshold = 0.4) => {
   if (!feed?.episodes) {
     return null
   }
 
+  const fuseOptions = {
+    ignoreDiacritics: true,
+    threshold,
+    keys: [
+      { name: 'title', weight: 0.7 }, // prefer match in title
+      { name: 'subtitle', weight: 0.3 }
+    ]
+  }
+  const fuse = new Fuse(feed.episodes, fuseOptions)
+
   const matches = []
-  feed.episodes.forEach((ep) => {
-    if (!ep.title) return
-    const epTitle = ep.title.toLowerCase().trim()
-    if (epTitle === searchTitle) {
-      matches.push({
-        episode: ep,
-        levenshtein: 0
-      })
-    } else {
-      const levenshtein = levenshteinDistance(searchTitle, epTitle, true)
-      if (levenshtein <= 6 && epTitle.length > levenshtein) {
-        matches.push({
-          episode: ep,
-          levenshtein
-        })
-      }
-    }
+  fuse.search(searchTitle).forEach((match) => {
+    matches.push({
+      episode: match.item
+    })
   })
-  return matches.sort((a, b) => a.levenshtein - b.levenshtein)
+  return matches
 }

server/utils/queries/libraryItemsBookFilters.js

@@ -264,9 +264,15 @@ module.exports = {
     } else if (sortBy === 'media.metadata.publishedYear') {
       return [[Sequelize.literal(`CAST(\`book\`.\`publishedYear\` AS INTEGER)`), dir]]
     } else if (sortBy === 'media.metadata.authorNameLF') {
-      return [[Sequelize.literal('`libraryItem`.`authorNamesLastFirst` COLLATE NOCASE'), dir]]
+      return [
+        [Sequelize.literal('`libraryItem`.`authorNamesLastFirst` COLLATE NOCASE'), dir],
+        [Sequelize.literal('`libraryItem`.`title` COLLATE NOCASE'), dir]
+      ]
     } else if (sortBy === 'media.metadata.authorName') {
-      return [[Sequelize.literal('`libraryItem`.`authorNamesFirstLast` COLLATE NOCASE'), dir]]
+      return [
+        [Sequelize.literal('`libraryItem`.`authorNamesFirstLast` COLLATE NOCASE'), dir],
+        [Sequelize.literal('`libraryItem`.`title` COLLATE NOCASE'), dir]
+      ]
     } else if (sortBy === 'media.metadata.title') {
       if (collapseseries) {
         return [[Sequelize.literal('display_title COLLATE NOCASE'), dir]]

server/utils/queries/libraryItemsPodcastFilters.js

@@ -149,11 +149,12 @@ module.exports = {
       libraryId
     }
     const libraryItemIncludes = []
-    if (includeRSSFeed) {
+    if (filterGroup === 'feed-open' || includeRSSFeed) {
+      const rssFeedRequired = filterGroup === 'feed-open'
       libraryItemIncludes.push({
         model: Database.feedModel,
-        required: filterGroup === 'feed-open',
-        separate: true
+        required: rssFeedRequired,
+        separate: !rssFeedRequired
       })
     }
     if (filterGroup === 'issues') {