Refine YouTube and Spotify music download process

akarpov/music/services/spotify.py

@@ -33,7 +33,11 @@ def get_spotdl_client():
         spot_settings = {
             "simple_tui": True,
             "log_level": "ERROR",
-            "lyrics_providers": ["genius", "musixmatch"],
+            "lyrics_providers": ["genius", "azlyrics", "musixmatch"],
+            "threads": 6,
+            "format": "mp3",
+            "ffmpeg": "ffmpeg",
+            "sponsor_block": True,
         }
         thread_local.spotdl_client = Spotdl(
             client_id=settings.MUSIC_SPOTIFY_ID,

akarpov/music/services/youtube.py

@@ -10,6 +10,7 @@
 from PIL import Image
 from pydub import AudioSegment
 from pytube import Search, YouTube
+from spotdl.providers.audio import YouTubeMusic
 
 from akarpov.music.models import Song
 from akarpov.music.services.db import load_track
@@ -18,22 +19,28 @@
 final_filename = None
 
 
-ydl_opts = {
-    "format": "m4a/bestaudio/best",
-    "postprocessors": [
-        {  # Extract audio using ffmpeg
-            "key": "FFmpegExtractAudio",
-            "preferredcodec": "m4a",
-        }
-    ],
-    "outtmpl": f"{settings.MEDIA_ROOT}/%(uploader)s_%(title)s.%(ext)s",
-}
+ytmusic = YouTubeMusic()
 
 
 def download_file(url):
+    ydl_opts = {
+        "format": "bestaudio/best",
+        "outtmpl": f"{settings.MEDIA_ROOT}/%(uploader)s_%(title)s.%(ext)s",
+        "postprocessors": [
+            {"key": "SponsorBlock"},  # Skip sponsor segments
+            {
+                "key": "FFmpegExtractAudio",
+                "preferredcodec": "mp3",
+                "preferredquality": "192",
+            },  # Extract audio
+            {"key": "EmbedThumbnail"},  # Embed Thumbnail
+            {"key": "FFmpegMetadata"},  # Apply correct metadata
+        ],
+    }
     with yt_dlp.YoutubeDL(ydl_opts) as ydl:
-        info = ydl.extract_info(url)
-    return info["requested_downloads"][0]["_filename"]
+        info = ydl.extract_info(url, download=True)
+        filename = ydl.prepare_filename(info)
+    return os.path.splitext(filename)[0] + ".mp3"
 
 
 def parse_description(description: str) -> list:
@@ -67,7 +74,7 @@ def parse_description(description: str) -> list:
 def download_from_youtube_link(link: str, user_id: int) -> Song:
     song = None
 
-    with yt_dlp.YoutubeDL(ydl_opts) as ydl:
+    with yt_dlp.YoutubeDL({"ignoreerrors": True, "extract_flat": True}) as ydl:
         info_dict = ydl.extract_info(link, download=False)
         title = info_dict.get("title", None)
         description = info_dict.get("description", None)
@@ -82,9 +89,12 @@ def download_from_youtube_link(link: str, user_id: int) -> Song:
         + slugify(orig_path.split("/")[-1].split(".")[0])
         + ".mp3"
     )
-    AudioSegment.from_file(orig_path).export(path)
-    if orig_path != path:
-        os.remove(orig_path)
+    if orig_path.endswith(".mp3"):
+        os.rename(orig_path, path)
+    else:
+        AudioSegment.from_file(orig_path).export(path)
+        if orig_path != path:
+            os.remove(orig_path)
     print(f"[processing] {title} converting to mp3: done")
 
     # split in chapters

akarpov/music/tasks.py

@@ -3,14 +3,13 @@
 import pylast
 import spotipy
 import structlog
-import yt_dlp
+import ytmusicapi
 from asgiref.sync import async_to_sync
 from celery import shared_task
 from channels.layers import get_channel_layer
 from django.conf import settings
 from django.utils import timezone
 from django.utils.timezone import now
-from pytube import Channel, Playlist
 from spotipy import SpotifyClientCredentials
 
 from akarpov.music.api.serializers import SongSerializer
@@ -40,19 +39,30 @@ def list_tracks(url, user_id):
         yandex.load_playlist(url, user_id)
     if "youtube.com" in url:
         if "channel" in url or "/c/" in url:
-            channel = Channel(url)
-            for video in channel.videos:
-                with yt_dlp.YoutubeDL({}) as ydl:
-                    info = ydl.extract_info(video, download=False)
-                    if info.get("category") == "Music":
-                        process_yb.apply_async(
-                            kwargs={"url": video.watch_url, "user_id": user_id}
-                        )
-        elif "playlist" in url or "&list=" in url:
-            playlist = Playlist(url)
-            for video in playlist.videos:
+            ytmusic = ytmusicapi.YTMusic()
+            channel_id = url.split("/")[-1]
+            channel_songs = ytmusic.get_artist(channel_id)["songs"]["results"]
+            print(channel_songs)
+
+            for song in channel_songs:
                 process_yb.apply_async(
-                    kwargs={"url": video.watch_url, "user_id": user_id}
+                    kwargs={
+                        "url": f"https://youtube.com/watch?v={song['videoId']}",
+                        "user_id": user_id,
+                    }
+                )
+
+        elif "playlist" in url or "&list=" in url:
+            ytmusic = ytmusicapi.YTMusic()
+            playlist_id = url.split("=")[-1]
+            playlist_songs = ytmusic.get_playlist(playlist_id)["tracks"]["results"]
+
+            for song in playlist_songs:
+                process_yb.apply_async(
+                    kwargs={
+                        "url": f"https://music.youtube.com/watch?v={song['videoId']}",
+                        "user_id": user_id,
+                    }
                 )
         else:
             process_yb.apply_async(kwargs={"url": url, "user_id": user_id})