fixed youtube track download, improved search

akarpov/music/documents.py

@@ -39,24 +39,70 @@ class SongDocument(Document):
         },
     )
 
-    meta = fields.ObjectField(dynamic=True)  # Added meta field here as dynamic object
+    meta = fields.ObjectField(dynamic=True)
 
     class Index:
         name = "songs"
-        settings = {"number_of_shards": 1, "number_of_replicas": 0}
+        settings = {
-        # settings = {
+            "number_of_shards": 1,
-        #     "number_of_shards": 1,
+            "number_of_replicas": 0,
-        #     "number_of_replicas": 0,
+            "analysis": {
-        #     "analysis": {
+                "filter": {
-        #         "analyzer": {
+                    "russian_stop": {
-        #             "russian_icu": {
+                        "type": "stop",
-        #                 "type": "custom",
+                        "stopwords": "_russian_",
-        #                 "tokenizer": "icu_tokenizer",
+                    },
-        #                 "filter": ["icu_folding","icu_normalizer"]
+                    "russian_keywords": {
-        #             }
+                        "type": "keyword_marker",
-        #         }
+                        "keywords": ["пример"],
-        #     }
+                    },
-        # } TODO
+                    "russian_stemmer": {
+                        "type": "stemmer",
+                        "language": "russian",
+                    },
+                    "autocomplete_filter": {
+                        "type": "edge_ngram",
+                        "min_gram": 1,
+                        "max_gram": 20,
+                    },
+                    "synonym_filter": {
+                        "type": "synonym",
+                        "synonyms": [
+                            "бит,трек,песня,музыка,песня,мелодия,композиция",
+                            "певец,исполнитель,артист,музыкант",
+                            "альбом,диск,пластинка,сборник,коллекция",
+                        ],
+                    },
+                },
+                "analyzer": {
+                    "russian": {
+                        "tokenizer": "standard",
+                        "filter": [
+                            "russian_stop",
+                            "russian_keywords",
+                            "russian_stemmer",
+                        ],
+                    },
+                    "russian_icu": {
+                        "tokenizer": "icu_tokenizer",
+                        "filter": [
+                            "russian_stop",
+                            "russian_keywords",
+                            "russian_stemmer",
+                        ],
+                    },
+                    "autocomplete": {
+                        "type": "custom",
+                        "tokenizer": "standard",
+                        "filter": [
+                            "lowercase",
+                            "autocomplete_filter",
+                            "synonym_filter",
+                        ],
+                    },
+                },
+            },
+        }
 
     class Django:
         model = Song

akarpov/music/services/db.py

@@ -72,6 +72,16 @@ def load_track(
             name = search_info["title"]
         elif not name:
             name = process_track_name(" ".join(p_name.strip().split("-")))
+            clear_name = [
+                "(Official HD Video)",
+                "(Official Music Video)",
+                "(Official Video)",
+                "Official Video",
+                "Official Music Video",
+                "Official HD Video",
+            ]
+            for c in clear_name:
+                name = name.replace(c, "")
 
     if not name:
         name = orig_name

akarpov/music/services/info.py

@@ -360,6 +360,7 @@ def update_author_info(author: Author) -> None:
 
 
 def search_all_platforms(track_name: str) -> dict:
+    print(track_name)
     session = spotipy.Spotify(
         auth_manager=spotipy.SpotifyClientCredentials(
             client_id=settings.MUSIC_SPOTIFY_ID,

akarpov/music/services/search.py

@@ -13,9 +13,9 @@ def search_song(query):
             ES_Q(
                 "multi_match",
                 query=query,
-                fields=["name^3", "authors.name^2", "album.name"],
+                fields=["name^5", "authors.name^3", "album.name^3"],
                 fuzziness="AUTO",
-            ),  # Change here
+            ),
             ES_Q("wildcard", name__raw=f"*{query.lower()}*"),
             ES_Q(
                 "nested",
@@ -27,6 +27,7 @@ def search_song(query):
                 path="album",
                 query=ES_Q("wildcard", album__name__raw=f"*{query.lower()}*"),
             ),
+            ES_Q("wildcard", meta__raw=f"*{query.lower()}*"),
         ],
         minimum_should_match=1,
     )

akarpov/music/services/youtube.py

@@ -157,6 +157,7 @@ def download_from_youtube_link(link: str, user_id: int) -> Song:
         print(f"[processing] loading {title}")
 
         info = search_all_platforms(title)
+        if "album_image" in info and info["album_image"]:
             if not info["album_image"].startswith("/"):
                 r = requests.get(info["album_image"])
                 img_pth = str(
@@ -173,6 +174,8 @@ def download_from_youtube_link(link: str, user_id: int) -> Song:
                 img_pth = f"{img_pth}.png"
             else:
                 img_pth = info["album_image"]
+        else:
+            img_pth = None
         if "genre" in info:
             song = load_track(
                 path,

akarpov/music/tasks.py

@@ -1,4 +1,5 @@
 from datetime import timedelta
+from urllib.parse import parse_qs, urlparse
 
 import pylast
 import spotipy
@@ -54,9 +55,19 @@ def list_tracks(url, 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"]
 
+            # Parse the URL and the query string
+            parsed_url = urlparse(url)
+            parsed_qs = parse_qs(parsed_url.query)
+
+            # Get the playlist ID from the parsed query string
+            playlist_id = parsed_qs.get("list", [None])[0]
+
+            if playlist_id:
+                playlist_songs = ytmusic.get_playlist(playlist_id)["tracks"]
+
+            else:
+                raise ValueError("No playlist ID found in the URL.")
             for song in playlist_songs:
                 process_yb.apply_async(
                     kwargs={

compose/production/elasticsearch/Dockerfile

@@ -0,0 +1,4 @@
+FROM elasticsearch:8.11.1
+
+# Install the ICU plugin
+RUN bin/elasticsearch-plugin install https://artifacts.elastic.co/downloads/elasticsearch-plugins/analysis-icu/analysis-icu-8.11.1.zip

local.yml

@@ -101,7 +101,9 @@ services:
     command: /start-flower
 
   elasticsearch:
-    image: elasticsearch:8.11.1
+    build:
+      context: .
+      dockerfile: ./compose/production/elasticsearch/Dockerfile
     ports:
       - "9200:9200"
       - "9300:9300"