Lyrics: switch to multiple search engines

Brave sometimes ratelimits SearXNG server

backend/config.py

@@ -13,10 +13,6 @@ class Config:
         # Cookies are in Netscape CSV format, see yt-dlp docs
         self.cookies_dir = Path(os.getenv('COOKIES_DIR') or 'cookies')
 
-        # Note: yt-dlp's path trimmer also counts album_path_tmpl, not only filename
-        # Why 235? 255 is the ext4 limit. 255 - len("/var/lib/musicdlp/") = 237, rounded down to 235
-        self.path_length = int(os.getenv('PATH_LENGTH') or 235)
-
         self.tmpl = os.path.join(
             # `artists.0` instead of `artist`, because the latter can contain "feat. ..."
             os.getenv('ALBUM_PATH_TMPL') or 'music/%(artists.0)s/%(album)s',

backend/genius.py

@@ -16,16 +16,16 @@ lyrics_xpath = etree.XPath('//div[@id="lyrics-root"][1]/div[@data-lyrics-contain
 br_xpath = etree.XPath('.//br')
 
 
-def search(title: str, artist: str) -> str:
+def search(title: str, artist: str) -> tuple[str, str]:
-    '''Searches for Genius lyrics using SearXNG + Yahoo and returns the first URL.
+    '''Searches for Genius lyrics using SearXNG + Yahoo
-    Irrelevant texts should be picked manually'''
+    and returns the first result as tuple(title, url)'''
 
     resp = http_pool.get().request(
         'GET',
         'https://searx.dc09.ru/search',
         fields={
             'q': artist + ' ' + title + ' site:genius.com',
-            'engines': 'brave',
+            'engines': 'brave,yahoo',
             'safesearch': '0',
             'format': 'json',
         },
@@ -34,7 +34,27 @@ def search(title: str, artist: str) -> str:
     result: dict[str, str] = resp.json()['results'][0]
     del resp
 
-    return result['url']
+    return (result['title'], result['url'])
+
+
+def raise_on_irrelevant_result(res_title: str, track_track: str, track_artist: str) -> None:
+    '''Raises ValueError
+    if no words from track title are present in search result track title
+    and no words from artist name are present in search result artist name'''
+
+    res_artist, res_track = res_title.lower().split(' \u2013 ', maxsplit=1)
+    if not (
+        any(
+            word.group(0).lower() in res_artist
+            for word in word_regex.finditer(track_artist)
+        )
+        and
+        any(
+            word.group(0).lower() in res_track
+            for word in word_regex.finditer(track_track)
+        )
+    ):
+        raise ValueError
 
 
 def parse(url: str) -> str:

backend/id3pp.py

@@ -63,7 +63,8 @@ class ID3TagsPP(PostProcessor):
             file['TCON'] = id3.TCON(encoding=ENC_UTF8, text=information['genre'])
 
         try:
-            lyr_url = genius.search(title, artists[0])
+            lyr_title, lyr_url = genius.search(title, artists[0])
+            genius.raise_on_irrelevant_result(lyr_title, title, artists[0])
             file['USLT'] = id3.USLT(encoding=ENC_UTF8, text=genius.parse(lyr_url))
         except:
             pass

backend/test_genius.py

@@ -23,15 +23,46 @@ LYR3 = '''you are gonna get yours
 Another day'''
 
 
+# There is no lyrics for this song on Genius
+# Maybe someday TITLE2 and ARTIST2 will need to be changed
+# (But really existing song is chosen intentionally)
+TITLE2 = 'Паруса'
+ARTIST2 = 'PIZZA'
+
+
 class TestGenius(TestCase):
 
     def setUp(self) -> None:
         http_pool.get()
 
-    def test_search_success(self) -> None:
+    def test_search(self) -> None:
-        url = genius.search(TITLE, ARTIST)
+        _, url = genius.search(TITLE, ARTIST)
         self.assertEqual(url, URL)
 
+    def test_search_success(self) -> None:
+        title, _ = genius.search(TITLE, ARTIST)
+        genius.raise_on_irrelevant_result(title, TITLE, ARTIST)
+
+    def test_search_failure(self) -> None:
+        title, _ = genius.search(TITLE2, ARTIST2)
+        with self.assertRaises(ValueError):
+            genius.raise_on_irrelevant_result(title, TITLE2, ARTIST2)
+
+    def test_relevancy_success(self) -> None:
+        genius.raise_on_irrelevant_result(
+            'ABC hEllo world!@ \u2013 sOmE artist123',
+            'Artist123',
+            'hello World',
+        )
+
+    def test_relevancy_failure(self) -> None:
+        with self.assertRaises(ValueError):
+            genius.raise_on_irrelevant_result(
+                'DEF hEllo world@!15 \u2013 anOther artist456',
+                'DEF 789',
+                'ABC irrelevant track title',
+            )
+
     def test_lyrics_parsing(self) -> None:
         lyrics = genius.parse(URL)
         self.assertTrue(lyrics.startswith(LYR1))

backend/ydl_pool.py

@@ -63,7 +63,6 @@ class Downloader:
         if ydl is None:
             ydl = create_ydl_fn[site]()
 
-            ydl.params['trim_file_name'] = cfg.path_length  # Note: not only filename, but path in outtmpl
             ydl.params['outtmpl']['default'] = cfg.tmpl
             ydl.add_post_processor(id3pp.ID3TagsPP(), when='post_process')
 

frontend/index.html

@@ -1 +1,33 @@
-<!-- TODO: simple web ui, websockets -->
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>musicdlp</title>
+    <link rel="stylesheet" href="/style.css">
+    <script src="/script.js"></script>
+  </head>
+  <body>
+    <div>
+      <input type="text" id="url" placeholder="Playlist or track URL">
+      <button type="button" id="guess-site-btn">Guess site</button>
+    </div>
+    <div>
+      <select id="site-select">
+        <option value="youtube" selected>YouTube</option>
+        <option value="yt_proxied">YT proxied</option>
+        <option value="yandex">Yandex Music</option>
+      </select>
+    </div>
+    <div>
+      <button type="button" id="items-btn">Get playlist items</button>
+    </div>
+    <div id="items-container"></div>
+    <div>
+      <button type="button">Download</button>
+    </div>
+    <div>
+      <label>Progress: <span id="progress">not implemented</span></label>
+    </div>
+  </body>
+</html>

frontend/script.js

@@ -0,0 +1,18 @@
+addEventListener('DOMContentLoaded', () => {
+  /** @type{HTMLInputElement} */
+  const urlField = document.getElementById('url')
+  /** @type{HTMLSelectElement} */
+  const site = document.getElementById('site-select')
+
+  document.getElementById('guess-site-btn').addEventListener('click', () => {
+    const url = urlField.value
+    if (url.includes('/watch?v=') || url.includes('/playlist?list=')) {
+      if (site.value == 'yt_proxied') {
+        return
+      }
+      site.value = 'youtube'
+    } else if (url.includes('://music.yandex.')) {
+      site.value = 'yandex'
+    }
+  })
+})

frontend/style.css

@@ -0,0 +1,11 @@
+body {
+  margin: 0;
+  padding: 0.5rem;
+
+  display: flex;
+  flex-direction: column;
+  align-items: center;
+  row-gap: 0.25rem;
+
+  font-family: 'Noto Sans', 'Roboto', 'Ubuntu', sans-serif;
+}