Append source to the lyrics

Author: snejus

beetsplug/lyrics.py

@@ -255,7 +255,7 @@ def __init__(self, config, log):
 
     def fetch(
         self, artist: str, title: str, album: str, length: int
-    ) -> str | None:
+    ) -> tuple[str, str] | None:
         raise NotImplementedError
 
 
@@ -266,6 +266,7 @@ class LRCLyrics:
     DURATION_DIFF_TOLERANCE = 0.05
 
     target_duration: float
+    id: int
     duration: float
     instrumental: bool
     plain: str
@@ -281,6 +282,7 @@ def make(
     ) -> LRCLyrics:
         return cls(
             target_duration,
+            candidate["id"],
             candidate["duration"] or 0.0,
             candidate["instrumental"],
             candidate["plainLyrics"],
@@ -360,18 +362,20 @@ def pick_best_match(cls, lyrics: Iterable[LRCLyrics]) -> LRCLyrics | None:
 
     def fetch(
         self, artist: str, title: str, album: str, length: int
-    ) -> str | None:
+    ) -> tuple[str, str] | None:
         """Fetch lyrics text for the given song data."""
         fetch = partial(self.fetch_candidates, artist, title, album, length)
         make = partial(LRCLyrics.make, target_duration=length)
         pick = self.pick_best_match
         try:
-            return next(
+            item = next(
                 filter(None, map(pick, (map(make, x) for x in fetch())))
-            ).get_text(self.config["synced"])
+            )
         except StopIteration:
             return None
 
+        return item.get_text(self.config["synced"]), f"{self.GET_URL}/{item.id}"
+
 
 class DirectBackend(Backend):
     """A backend for fetching lyrics directly."""
@@ -407,7 +411,7 @@ def encode(cls, text: str) -> str:
 
         return quote(unidecode(text))
 
-    def fetch(self, artist: str, title: str, *_) -> str | None:
+    def fetch(self, artist: str, title: str, *_) -> tuple[str, str] | None:
         url = self.build_url(artist, title)
 
         html = self.fetch_text(url)
@@ -429,7 +433,7 @@ def fetch(self, artist: str, title: str, *_) -> str | None:
         # sometimes there are non-existent lyrics with some content
         if "Lyrics | Musixmatch" in lyrics:
             return None
-        return lyrics
+        return lyrics, url
 
 
 class Html:
@@ -530,13 +534,13 @@ def get_results(self, artist: str, title: str) -> Iterable[SearchResult]:
             if check_match(candidate):
                 yield candidate
 
-    def fetch(self, artist: str, title: str, *_) -> str | None:
+    def fetch(self, artist: str, title: str, *_) -> tuple[str, str] | None:
         """Fetch lyrics for the given artist and title."""
         for result in self.get_results(artist, title):
             if (html := self.fetch_text(result.url)) and (
                 lyrics := self.scrape(html)
             ):
-                return lyrics
+                return lyrics, result.url
 
         return None
 
@@ -594,11 +598,15 @@ class Tekstowo(SoupMixin, DirectBackend):
     def encode(cls, text: str) -> str:
         return cls.non_alpha_to_underscore(unidecode(text.lower()))
 
-    def fetch(self, artist: str, title: str, *_) -> str | None:
+    def fetch(self, artist: str, title: str, *_) -> tuple[str, str] | None:
+        url = self.build_url(artist, title)
         # We are expecting to receive a 404 since we are guessing the URL.
         # Thus suppress the error so that it does not end up in the logs.
         with suppress(NotFoundError):
-            return self.scrape(self.fetch_text(self.build_url(artist, title)))
+            if lyrics := self.scrape(self.fetch_text(url)):
+                return lyrics, url
+
+        return None
 
         return None
 
@@ -1004,8 +1012,9 @@ def get_lyrics(self, artist: str, title: str, *args) -> str | None:
         self.info("Fetching lyrics for {} - {}", artist, title)
         for backend in self.backends:
             with backend.handle_request():
-                if lyrics := backend.fetch(artist, title, *args):
-                    return lyrics
+                if lyrics_info := backend.fetch(artist, title, *args):
+                    lyrics, url = lyrics_info
+                    return f"{lyrics}\n\nSource: {url}"
 
         return None
 

test/plugins/test_lyrics.py

@@ -279,11 +279,12 @@ def _patch_google_search(self, requests_mock, lyrics_page):
 
     def test_backend_source(self, lyrics_plugin, lyrics_page: LyricsPage):
         """Test parsed lyrics from each of the configured lyrics pages."""
-        lyrics = lyrics_plugin.get_lyrics(
+        lyrics_info = lyrics_plugin.get_lyrics(
             lyrics_page.artist, lyrics_page.track_title, "", 186
         )
 
-        assert lyrics
+        assert lyrics_info
+        lyrics, _ = lyrics_info.split("\n\nSource: ")
         assert lyrics == lyrics_page.lyrics
 
 
@@ -400,6 +401,7 @@ def test_scrape(self, backend, lyrics_html, expecting_lyrics):
 
 def lyrics_match(**overrides):
     return {
+        "id": 1,
         "instrumental": False,
         "duration": LYRICS_DURATION,
         "syncedLyrics": "synced",
@@ -428,7 +430,9 @@ def fetch_lyrics(self, backend, requests_mock, response_data):
         [({"synced": True}, "synced"), ({"synced": False}, "plain")],
     )
     def test_synced_config_option(self, fetch_lyrics, expected_lyrics):
-        assert fetch_lyrics() == expected_lyrics
+        lyrics, _ = fetch_lyrics()
+
+        assert lyrics == expected_lyrics
 
     @pytest.mark.parametrize(
         "response_data, expected_lyrics",
@@ -475,4 +479,10 @@ def test_synced_config_option(self, fetch_lyrics, expected_lyrics):
     )
     @pytest.mark.parametrize("plugin_config", [{"synced": True}])
     def test_fetch_lyrics(self, fetch_lyrics, expected_lyrics):
-        assert fetch_lyrics() == expected_lyrics
+        lyrics_info = fetch_lyrics()
+        if lyrics_info is None:
+            assert expected_lyrics is None
+        else:
+            lyrics, _ = fetch_lyrics()
+
+            assert lyrics == expected_lyrics