Most releases found are missing tracks #25

Activity

1. neilanthonystuckey commented on Jul 12, 2024

   

   neilanthonystuckey

   on Jul 12, 2024

   Author

   More actions

   I let this run overnight looks like 7 out 112 albums were not missing tracks. I am using docker i had to move the download location out of my library as it was making a mess of things pretty quickly with the missing tracks.
2. dmzoneill commented on Oct 16, 2024

   

   dmzoneill

   on Oct 16, 2024

   Owner

   More actions

   PR's are welcome
3. chaquetashop commented on Sep 15, 2025

   

   chaquetashop

   on Sep 15, 2025

   Last edited by chaquetashop

   More actions

   i have got this problem because the lidar rely on ID3 tags (metadata inside the MP3 artist, album, track, etc.), not just filenames. I have fixed by creating script and run it using scheduler to fix the metadata.

   #!/usr/bin/env bash
   # Simple in-place auto-tagger for LYD downloads
   # Usage: ./lyd_auto_tagger_no_backup.sh /path/to/music
   
   set -uo pipefail
   
   BASE_DIR="$1"
   if [[ -z "${BASE_DIR:-}" ]]; then
     echo "Usage: $0 /path/to/music-downloads"
     exit 1
   fi
   
   # find eyeD3 (try user's pipx install first, fallback to PATH)
   EYE="$(command -v eyeD3 || true)"
   if [[ -z "$EYE" ]]; then
     EYE="/home/$(whoami)/.local/bin/eyeD3"
   fi
   if [[ ! -x "$EYE" ]]; then
     echo "eyeD3 not found at '$EYE'. Install it and re-run."
     exit 1
   fi
   
   echo "Base: $BASE_DIR"
   echo "eyeD3: $EYE"
   echo "Tagging in-place (no backup)."
   
   find "$BASE_DIR" -type f -iname "*.mp3" -print0 | while IFS= read -r -d '' file; do
     dir="$(dirname "$file")"
     filename="$(basename "$file")"
   
     artist=""
     album=""
     title=""
     tracknum=""
   
     # Pattern: "NN - Title.mp3"
     if [[ "$filename" =~ ^([0-9]{1,2})[[:space:]]*-[[:space:]]*(.+)\.mp3$ ]]; then
       tracknum="${BASH_REMATCH[1]}"
       title="${BASH_REMATCH[2]}"
       album="$(basename "$dir")"
       artist="$(basename "$(dirname "$dir")")"
   
     # Pattern: "Artist - Album - Title.mp3"
     elif [[ "$filename" =~ ^(.+)[[:space:]]-[[:space:]](.+)[[:space:]]-[[:space:]](.+)\.mp3$ ]]; then
       artist="${BASH_REMATCH[1]}"
       album="${BASH_REMATCH[2]}"
       title="${BASH_REMATCH[3]}"
   
     # Pattern: "Artist - Title.mp3"
     elif [[ "$filename" =~ ^(.+)[[:space:]]-[[:space:]](.+)\.mp3$ ]]; then
       artist="${BASH_REMATCH[1]}"
       title="${BASH_REMATCH[2]}"
       album="$(basename "$dir")"
     else
       echo "Skipping (unrecognized pattern): $file"
       continue
     fi
   
     # Trim whitespace
     artist="$(echo "$artist" | sed -e 's/^[[:space:]]*//' -e 's/[[:space:]]*$//')"
     album="$(echo "$album" | sed -e 's/^[[:space:]]*//' -e 's/[[:space:]]*$//')"
     title="$(echo "$title" | sed -e 's/^[[:space:]]*//' -e 's/[[:space:]]*$//')"
   
     cmd=( "$EYE" )
     [[ -n "$artist" ]] && cmd+=( -a "$artist" )
     [[ -n "$album" ]]  && cmd+=( -A "$album" )
     [[ -n "$tracknum" ]] && cmd+=( -n "$tracknum" )
     [[ -n "$title" ]]  && cmd+=( -t "$title" )
     cmd+=( "$file" )
   
     # Use sudo if file is not writable
     if [[ -w "$file" ]]; then
       "${cmd[@]}" || echo "ERROR: eyeD3 failed for $file"
     else
       sudo "${cmd[@]}" || echo "ERROR: eyeD3 failed for $file (even with sudo)"
     fi
   
     echo "Tagged: $file"
   done
   
   echo "All done!"
   

   dont forget to install eyeD3 and make your account can run without sudo

   pipx install eyed3
   sudo visudo
   user ALL=(ALL) NOPASSWD: /home/user/.local/bin/eyeD3
   
4. dmzoneill commented on Feb 19, 2026

   

   dmzoneill

   on Feb 19, 2026

   Owner

   More actions

   Thank you for reporting this issue! This is a common problem with lidarr-youtube-downloader. Let me break down what's happening and provide some solutions.

   Root Cause Analysis

   The issue occurs because YouTube videos are often partial or mismatched releases compared to what Lidarr expects. The downloader's matching logic has two limitations:

   1. Weak track count validation: The code downloads tracks based on a fuzzy string match (default 0.8 threshold) between the search query and YouTube video title, but it doesn't verify that the downloaded track actually matches the Lidarr release's metadata or track count.

   2. No pre-download validation: Before downloading from a YouTube result, the script doesn't check if that video is likely to contain all tracks from the album or if it's a partial/compilation version.

   Looking at the code in lyd.py (line 406-410), the matching is purely title-based:

   if SequenceMatcher(None, searchFor, song["title"]).ratio() > best:
       if skip_youtube_download(song["link"]) is False:
           best = SequenceMatcher(None, searchFor, song["title"]).ratio()
           bestLink = song["link"]
           bestTitle = song["title"]

   Debugging Steps

   To diagnose this on your setup:

   1. Check the downloaded files: Are the tracks actually missing from the YouTube downloads, or are they being downloaded but not recognized by Lidarr?

      # List what was downloaded
      ls -la /path/to/music/Artist/Album/

   2. Verify ID3 tags: Lidarr relies on ID3 metadata, not just filenames. Check if your downloads have proper tags:

      eyeD3 /path/to/music/Artist/Album/Song.mp3

   3. Enable verbose logging: Run with --verbose flag (if supported) or add print statements to see which YouTube results are being selected.

   4. Test specific albums: Try running the downloader on a single album to see which YouTube videos are being selected and whether they're complete versions.

   Solutions

   Option 1: Adjust Match Threshold (Strictness)

   Increase the MATCH_THRESHOLD environment variable to require closer matches:

   export MATCH_THRESHOLD="0.95"  # Higher = stricter matching
   lyd

   This will skip videos with titles that don't closely match, reducing false positives. However, this may cause more "Unable to find" results.

   Option 2: Add Keyword Filters (via BLACKLIST_KEYWORDS)

   Use the existing BLACKLIST_KEYWORDS env var to skip low-quality results:

   export BLACKLIST_KEYWORDS="playlist,compilation,mix,remix,cover,live,karaoke,instrumental"
   lyd

   Option 3: Fix Metadata Post-Download (Recommended)

   As noted in an earlier comment, the real issue is often missing/incorrect ID3 tags. Use the provided tagging script to fix metadata after download:

   pipx install eyed3
   # Then run the auto-tagger script from the issue comments
   ./lyd_auto_tagger_no_backup.sh /path/to/music

   Option 4: Feature Request - Pre-Validation

   A better long-term fix would be to:

   • Fetch YouTube video duration and compare to expected track duration
   • Skip results where the title contains obvious red flags ("full album", "playlist", etc. but treat "full album" as valid)
   • Validate against Lidarr's release track count before accepting a match

   Recommended Workaround

   For immediate relief, I'd suggest:

   1. Manually curate your first batch by checking YouTube manually for correct versions (full albums vs. compilations)
   2. Add blacklist keywords to avoid obvious mismatches
   3. Use a post-download metadata fixer to ensure Lidarr can properly index the files
   4. Consider opening a PR with enhanced validation logic (checking video duration, validating against Lidarr track counts, etc.)

   Would you be interested in contributing a PR with stricter album matching? The validation could check if a YouTube video's duration roughly matches the expected album length.
5. dmzoneill

   added

   bugSomething isn't workingSomething isn't working

   enhancementNew feature or requestNew feature or request

   on Feb 19, 2026
6. addmix commented on Mar 30, 2026

   

   addmix

   on Mar 30, 2026

   More actions

   Why all the AI generated comments, man?
7. dmzoneill commented on Oct 30, 2026

   

   dmzoneill

   on Oct 30, 2026 – with Cursor

   Owner

   More actions

   Thanks for reporting this, and sorry for the slow reply! To dig into the missing tracks, could you share:

   • The lidarr-youtube-downloader version (or commit) and your Lidarr version
   • How you run it (Docker, pip, OS)
   • One or two example albums where tracks went missing, with the track count Lidarr expects versus what was downloaded
   • The log output from a run on one of those albums

   That will show whether the matching is picking the wrong YouTube release or dropping tracks after the match.
8. dmzoneill

   removed

   enhancementNew feature or requestNew feature or request

   on Oct 30, 2026