Files
vrobbler/vrobbler/apps/music/models.py

728 lines
25 KiB
Python

import logging
from dataclasses import dataclass
from typing import Optional
from uuid import uuid4
import musicbrainzngs
import requests
from django.conf import settings
from django.core.files.base import ContentFile
from django.db import models
from django.urls import reverse
from django.utils.translation import gettext_lazy as _
from django_extensions.db.models import TimeStampedModel
from imagekit.models import ImageSpecField
from imagekit.processors import ResizeToFit
from music.allmusic import get_allmusic_slug, scrape_data_from_allmusic
from music.bandcamp import get_bandcamp_slug
from music.musicbrainz import (
get_album_metadata_with_artist,
get_recording_mbid_exact,
get_track_metadata_with_artist,
)
from music.theaudiodb import lookup_album_from_tadb, lookup_artist_from_tadb
from music.utils import clean_artist_name
from scrobbles.dataclasses import BaseLogData
from scrobbles.mixins import ScrobblableConstants, ScrobblableMixin
logger = logging.getLogger(__name__)
BNULL = {"blank": True, "null": True}
@dataclass
class TrackLogData(BaseLogData):
mopidy_source: Optional[str] = None
rockbox_info: Optional[str] = None
rating: Optional[int] = None
class Artist(TimeStampedModel):
"""Represents a music artist.
# Lookup or create by title alone
>>> Artist.find_or_create(name="Bon Iver")
# Lookup or create by MB id alone
>>> Artist.find_or_create(musicbrainz_id="0307edfc-437c-4b48-8700-80680e66a228")
"""
uuid = models.UUIDField(default=uuid4, editable=False, **BNULL)
name = models.CharField(max_length=255)
biography = models.TextField(**BNULL)
theaudiodb_id = models.CharField(max_length=255, unique=True, **BNULL)
theaudiodb_genre = models.CharField(max_length=255, **BNULL)
theaudiodb_mood = models.CharField(max_length=255, **BNULL)
musicbrainz_id = models.CharField(max_length=255, **BNULL)
allmusic_id = models.CharField(max_length=100, **BNULL)
bandcamp_id = models.CharField(max_length=100, **BNULL)
thumbnail = models.ImageField(upload_to="artist/", **BNULL)
thumbnail_small = ImageSpecField(
source="thumbnail",
processors=[ResizeToFit(100, 100)],
format="JPEG",
options={"quality": 60},
)
thumbnail_medium = ImageSpecField(
source="thumbnail",
processors=[ResizeToFit(300, 300)],
format="JPEG",
options={"quality": 75},
)
alt_names = models.TextField(**BNULL)
class Meta:
unique_together = [["name", "musicbrainz_id"]]
def __str__(self):
return self.name
@property
def primary_image_url(self) -> str:
if self.thumbnail:
return self.thumbnail.url
if self.album_set.first().cover_image:
return self.album_set.first().cover_image.url
return ""
@property
def mb_link(self) -> str:
if self.musicbrainz_id:
return f"https://musicbrainz.org/artist/{self.musicbrainz_id}"
return ""
@property
def allmusic_link(self):
if self.allmusic_id:
return f"https://www.allmusic.com/artist/{self.allmusic_id}"
return ""
@property
def bandcamp_link(self):
if self.bandcamp_id:
return f"https://{self.bandcamp_id}.bandcamp.com/"
return ""
def get_absolute_url(self):
return reverse("music:artist_detail", kwargs={"slug": self.uuid})
def scrobbles(self):
from scrobbles.models import Scrobble
return Scrobble.objects.filter(
track__in=self.track_set.all()
).order_by("-timestamp")
@property
def tracks(self):
return (
self.track_set.all()
.annotate(scrobble_count=models.Count("scrobble"))
.order_by("-scrobble_count")
)
def charts(self):
from scrobbles.models import ChartRecord
return ChartRecord.objects.filter(track__artist=self).order_by("-year")
def scrape_allmusic(self, force=False) -> None:
if not self.allmusic_id or force:
slug = get_allmusic_slug(self.name)
if not slug:
logger.info(
"No allmusic link found", extra={"track_id": self.id}
)
return
self.allmusic_id = slug
self.save(update_fields=["allmusic_id"])
def scrape_bandcamp(self, force=False) -> None:
if not self.bandcamp_id or force:
slug = get_bandcamp_slug(self.name)
if not slug:
logger.info(
"No bandcamp link found", extra={"track_id": self.id}
)
return
self.bandcamp_id = slug
self.save(update_fields=["bandcamp_id"])
def fix_metadata(self, force_update=False):
tadb_info = {}
if self.theaudiodb_id and force_update:
tadb_info = lookup_artist_from_tadb(self.theaudiodb_id)
if not self.theaudiodb_id or (force_update and not tadb_info):
tadb_info = lookup_artist_from_tadb(self.name)
if not tadb_info:
logger.warn(
f"No response from TADB for artist {self.name}, try force_update=True"
)
return
self.biography = tadb_info["biography"]
self.theaudiodb_genre = tadb_info["genre"]
self.theaudiodb_mood = tadb_info["mood"]
thumb_url = tadb_info.get("thumb_url", "")
if thumb_url:
r = requests.get(thumb_url)
if r.status_code == 200:
fname = f"{self.name}_{self.uuid}.jpg"
self.thumbnail.save(fname, ContentFile(r.content), save=True)
@property
def rym_link(self):
artist_slug = self.name.lower().replace(" ", "-").replace(",", "")
return f"https://rateyourmusic.com/artist/{artist_slug}/"
@property
def bandcamp_search_link(self):
artist = self.name.lower()
return f"https://bandcamp.com/search?q={artist}&item_type=b"
@classmethod
def find_or_create(
cls, name: str, album_name: str = "", track_name: str = ""
) -> "Artist":
"""The biggest challenge to finding artists is that the search often
fails miserably unless you can look it up along with an album or a track name.
Thus, when we find or create an artist, we should always provide an optional
album name or track name, but probably not both."""
if album_name:
logger.info(
f"Looking for artist with name {name} and album {album_name}"
)
if track_name:
logger.info(
f"Looking for artist with name {name} and track {track_name}"
)
keys = {}
name = clean_artist_name(name)
keys["name"] = name
artist = cls.objects.filter(name=name).first()
if artist:
return artist
# alt_name = None
artist_dict = {}
if album_name:
album_dict = get_album_metadata_with_artist(album_name, name)
if album_dict:
artist_dict = album_dict.get("primary_artist")
if track_name:
track_dict = get_track_metadata_with_artist(track_name, name)
if track_dict:
artist_dict = track_dict.get("primary_artist")
if not artist_dict:
artist, created = cls.objects.get_or_create(name=name)
if created:
artist.fix_metadata()
return artist
musicbrainz_id = artist_dict.get("mbid")
found_name = artist_dict.get("name", name)
if found_name and name != found_name:
alt_name = found_name
artist = cls.objects.filter(
name=found_name, musicbrainz_id=musicbrainz_id
).first()
if not artist:
artist = cls.objects.create(
name=found_name,
musicbrainz_id=musicbrainz_id,
)
artist.fix_metadata()
# TODO: See if this alt_names stuff actually works or causes hard to debug problems
# If we did find our artist, but the found name is slightly differnt, record that
# if artist and alt_name:
# if not artist.alt_names:
# artist.alt_names = alt_name
# else:
# artist.alt_names += f"\\{alt_name}"
# logger.info(
# f"Add alt_name {alt_name} to artist {artist}",
# extra={"alt_name": alt_name, "artist_id": artist.id},
# )
# artist.save(update_fields=["alt_names"])
return artist
class Album(TimeStampedModel):
uuid = models.UUIDField(default=uuid4, editable=False, **BNULL)
name = models.CharField(max_length=255)
album_artist = models.ForeignKey(
Artist, related_name="albums", on_delete=models.DO_NOTHING, **BNULL
)
artists = models.ManyToManyField(Artist)
year = models.IntegerField(**BNULL)
musicbrainz_id = models.CharField(max_length=255, unique=True, **BNULL)
musicbrainz_releasegroup_id = models.CharField(max_length=255, **BNULL)
musicbrainz_albumartist_id = models.CharField(max_length=255, **BNULL)
cover_image = models.ImageField(upload_to="albums/", **BNULL)
cover_image_small = ImageSpecField(
source="cover_image",
processors=[ResizeToFit(100, 100)],
format="JPEG",
options={"quality": 60},
)
cover_image_medium = ImageSpecField(
source="cover_image",
processors=[ResizeToFit(300, 300)],
format="JPEG",
options={"quality": 75},
)
theaudiodb_id = models.CharField(max_length=255, unique=True, **BNULL)
theaudiodb_description = models.TextField(**BNULL)
theaudiodb_year_released = models.IntegerField(**BNULL)
theaudiodb_score = models.FloatField(**BNULL)
theaudiodb_score_votes = models.IntegerField(**BNULL)
theaudiodb_genre = models.CharField(max_length=255, **BNULL)
theaudiodb_style = models.CharField(max_length=255, **BNULL)
theaudiodb_mood = models.CharField(max_length=255, **BNULL)
theaudiodb_speed = models.CharField(max_length=255, **BNULL)
theaudiodb_theme = models.CharField(max_length=255, **BNULL)
allmusic_id = models.CharField(max_length=255, **BNULL)
allmusic_rating = models.IntegerField(**BNULL)
allmusic_review = models.TextField(**BNULL)
bandcamp_id = models.CharField(max_length=100, **BNULL)
rateyourmusic_id = models.CharField(max_length=255, **BNULL)
wikipedia_slug = models.CharField(max_length=255, **BNULL)
discogs_id = models.CharField(max_length=255, **BNULL)
wikidata_id = models.CharField(max_length=255, **BNULL)
alt_names = models.TextField(**BNULL)
def __str__(self) -> str:
return "{} by {}".format(self.name, self.album_artist)
def get_absolute_url(self):
return reverse("music:album_detail", kwargs={"slug": self.uuid})
def scrobbles(self):
from scrobbles.models import Scrobble
return Scrobble.objects.filter(
track__in=self.track_set.all()
).order_by("-timestamp")
@property
def primary_image_url(self) -> str:
if self.cover_image:
return self.cover_image_medium.url
return ""
@property
def tracks(self):
return (
self.track_set.all()
.annotate(scrobble_count=models.Count("scrobble"))
.order_by("-scrobble_count")
)
def fix_album_artist(self):
from music.utils import get_or_create_various_artists
multiple_artists = self.artists.count() > 1
if multiple_artists:
self.album_artist = get_or_create_various_artists()
else:
self.album_artist = self.artists.first()
self.save(update_fields=["album_artist"])
def scrape_allmusic(self, force=False) -> None:
if not self.name:
logger.warning(
"Album without a name cannot be scraped",
extra={"album_id": self.id},
)
return
if self.album_artist and (not self.allmusic_id or force):
slug = get_allmusic_slug(self.album_artist.name, self.name)
if not slug:
logger.info(
f"No allmsuic link for {self} by {self.album_artist}"
)
return
self.allmusic_id = slug
self.save(update_fields=["allmusic_id"])
allmusic_data = None
if self.allmusic_link:
allmusic_data = scrape_data_from_allmusic(self.allmusic_link)
if not allmusic_data:
logger.info(f"No allmsuic data for {self} by {self.album_artist}")
return
self.allmusic_review = allmusic_data["review"]
self.allmusic_rating = allmusic_data["rating"]
self.save(update_fields=["allmusic_review", "allmusic_rating"])
def scrape_theaudiodb(self) -> None:
artist = "Various Artists"
if self.album_artist:
artist = self.album_artist.name
album_data = lookup_album_from_tadb(self.name, artist)
if not album_data.get("theaudiodb_id"):
logger.info(f"No data for {self} found in TheAudioDB")
return
try:
Album.objects.filter(pk=self.pk).update(**album_data)
except:
logger.info(
f"Could not save info for album {self} with data {album_data}"
)
def scrape_bandcamp(self, force=False) -> None:
if not self.bandcamp_id or force:
slug = get_bandcamp_slug(self.album_artist.name, self.name)
if not slug:
logger.info(f"No bandcamp link for {self}")
return
self.bandcamp_id = slug
self.save(update_fields=["bandcamp_id"])
def fix_metadata(self):
if (
not self.musicbrainz_albumartist_id
or not self.year
or not self.musicbrainz_releasegroup_id
):
musicbrainzngs.set_useragent("vrobbler", "0.3.0")
mb_data = musicbrainzngs.get_release_by_id(
self.musicbrainz_id, includes=["artists", "release-groups"]
)
if not self.musicbrainz_releasegroup_id:
self.musicbrainz_releasegroup_id = mb_data["release"][
"release-group"
]["id"]
if not self.musicbrainz_albumartist_id:
self.musicbrainz_albumartist_id = mb_data["release"][
"artist-credit"
][0]["artist"]["id"]
if not self.year:
try:
self.year = mb_data["release"]["date"][0:4]
except KeyError:
pass
except IndexError:
pass
self.save(
update_fields=[
"musicbrainz_albumartist_id",
"musicbrainz_releasegroup_id",
"year",
]
)
new_artist = Artist.objects.filter(
musicbrainz_id=self.musicbrainz_albumartist_id
).first()
if self.musicbrainz_albumartist_id and new_artist:
self.artists.add(new_artist)
if not new_artist:
for t in self.track_set.all():
self.artists.add(t.artist)
if (
not self.cover_image
or self.cover_image == "default-image-replace-me"
):
self.fetch_artwork()
self.fix_album_artist()
self.scrape_theaudiodb()
self.scrape_allmusic()
def fetch_artwork(self, force=False):
if not self.cover_image and not force:
if self.musicbrainz_id:
try:
img_data = musicbrainzngs.get_image_front(
self.musicbrainz_id
)
name = f"{self.name}_{self.uuid}.jpg"
self.cover_image = ContentFile(img_data, name=name)
logger.info(f"Setting image to {name}")
except musicbrainzngs.ResponseError:
logger.warning(
f"No cover art found for {self.name} by release"
)
if (
not self.cover_image
or self.cover_image == "default-image-replace-me"
) and self.musicbrainz_releasegroup_id:
try:
img_data = musicbrainzngs.get_release_group_image_front(
self.musicbrainz_releasegroup_id
)
name = f"{self.name}_{self.uuid}.jpg"
self.cover_image = ContentFile(img_data, name=name)
logger.info(f"Setting image to {name}")
except musicbrainzngs.ResponseError:
logger.warning(
f"No cover art found for {self.name} by release group"
)
if not self.cover_image:
logger.debug(
f"No cover art found for release or release group for {self.name}, setting to default"
)
self.save()
@property
def mb_link(self) -> str:
return f"https://musicbrainz.org/release/{self.musicbrainz_id}"
@property
def allmusic_link(self) -> str:
if self.allmusic_id:
return f"https://www.allmusic.com/album/{self.allmusic_id}"
return ""
@property
def wikipedia_link(self):
if self.wikipedia_slug:
return f"https://www.wikipedia.org/en/{self.wikipedia_slug}"
return ""
@property
def tadb_link(self):
if self.theaudiodb_id:
return f"https://www.theaudiodb.com/album/{self.theaudiodb_id}"
return ""
@property
def rym_link(self):
artist_slug = self.album_artist.name.lower().replace(" ", "-")
album_slug = self.name.lower().replace(" ", "-")
return f"https://rateyourmusic.com/release/album/{artist_slug}/{album_slug}/"
@property
def bandcamp_link(self):
if self.bandcamp_id and self.album_artist.bandcamp_id:
return f"https://{self.album_artist.bandcamp_id}.bandcamp.com/album/{self.bandcamp_id}"
return ""
@property
def bandcamp_search_link(self):
artist = self.album_artist.name.lower()
album = self.name.lower()
return f"https://bandcamp.com/search?q={album} {artist}&item_type=a"
@classmethod
def find_or_create(cls, name: str, artist_name: str) -> "Album":
logger.info(
f"Looking for album with name {name} and artist_name {artist_name}"
)
artist = Artist.find_or_create(artist_name, album_name=name)
album_dict = get_album_metadata_with_artist(name, artist.name)
if not album_dict:
logger.info(
f"Could not find album {name} with artist {artist.name} on musicbrainz"
)
album = Album.objects.filter(name=name).first()
if not album:
album, created = Album.objects.get_or_create(
name=name,
)
if created:
# album.fix_metadata()
# album.fetch_artwork()
...
return album
if not artist:
artist_dict = album_dict.get("primary_artist", {})
if artist_dict:
artist = Artist.objects.filter(
musicbrainz_id=artist_dict.get("mbid"),
).first()
if not artist:
artist = Artist.objects.create(
musicbrainz_id=artist_dict.get("mbid"),
)
extra_artists = []
if not artist and len(album_dict.get("all_artists")) > 1:
artist = Artist.objects.filter(name="Various Artists").first()
extra_artists.append(artist)
if not artist:
raise Exception("No album artist found, and not a compliation")
album = cls.objects.filter(
models.Q(name=name) | models.Q(alt_names__icontains=name),
album_artist=artist,
).first()
alt_name = None
found_name = album_dict.get("album_title", name)
if found_name and name != found_name:
alt_name = name
album = Album.objects.filter(
musicbrainz_id=album_dict.get("mbid")
).first()
if not album:
year = None
if album_dict.get("release_date"):
year = album_dict.get("release_date", "").split("-")[0]
album = Album.objects.create(
name=found_name,
musicbrainz_id=album_dict.get("mbid"),
musicbrainz_releasegroup_id=album_dict.get(
"release_group_mbid"
),
year=year,
album_artist=artist,
alt_names=alt_name,
)
album.artists.add(*extra_artists)
album.fetch_artwork()
return album
class Track(ScrobblableMixin):
COMPLETION_PERCENT = getattr(settings, "MUSIC_COMPLETION_PERCENT", 100)
artist = models.ForeignKey(Artist, on_delete=models.DO_NOTHING)
albums = models.ManyToManyField(Album, related_name="tracks")
album = models.ForeignKey(Album, on_delete=models.DO_NOTHING, **BNULL)
musicbrainz_id = models.CharField(max_length=255, **BNULL)
class Meta:
unique_together = [["album", "musicbrainz_id"]]
def __str__(self):
return f"{self.title} by {self.artist}"
def logdata_cls(self):
return TrackLogData
@property
def primary_album(self):
if self.album:
return self.album
return self.albums.order_by("year").first()
def get_absolute_url(self):
return reverse("music:track_detail", kwargs={"slug": self.uuid})
@property
def subtitle(self) -> str:
return str(self.artist)
@property
def strings(self) -> ScrobblableConstants:
return ScrobblableConstants(verb="Listening", tags="notes")
@property
def mb_link(self):
return f"https://musicbrainz.org/recording/{self.musicbrainz_id}"
@property
def info_link(self):
return self.mb_link
@property
def primary_image_url(self) -> str:
url = ""
if self.artist.thumbnail:
url = self.artist.thumbnail_medium.url
if self.primary_album and self.primary_album.cover_image:
url = self.primary_album.cover_image_medium.url
return url
@classmethod
def find_or_create(
cls,
title: str = "",
artist_name: str = "",
album_name: str = "",
run_time_seconds: int | None = None,
enrich: bool = False,
commit: bool = True,
) -> "Track":
"""Given a name, try to find the track by the artist from Musicbrainz.
As a basic conceit we trust the source for giving us the track and artist
name
Optionally, we can update any found artists with overwrite."""
album = None
if album_name:
logger.info("Looking up album for: {album_name}")
album = Album.find_or_create(
name=album_name, artist_name=artist_name
)
artist = album.album_artist
else:
artist = Artist.find_or_create(artist_name, track_name=title)
if not artist:
artist = Artist.find_or_create(artist_name)
lookup_keys = {"title": title, "artist": artist}
if run_time_seconds:
lookup_keys["base_run_time_seconds"] = run_time_seconds
logger.info(f"Looking up track using: {lookup_keys}")
track = cls.objects.filter(**lookup_keys).first()
if track:
logger.info(
"Found match for track by name and artist, not going to musicbrainz ",
extra={
"track_id": track.id,
"title": title,
"artist_name": artist_name,
"run_time_seconds": run_time_seconds,
},
)
return track
track = cls.objects.filter(title=title, artist=artist).first()
if not track:
track, _ = cls.objects.get_or_create(title=title, artist=artist)
if album:
track.albums.add(album)
if enrich or not track.base_run_time_seconds:
logger.info(
f"Enriching track {track}",
extra={
"title": title,
"artist_name": artist_name,
"track_id": track.id,
},
)
try:
mbid, length = get_recording_mbid_exact(
title, artist_name, album_name
)
except Exception:
print("No musicbrainz result found, cannot enrich")
return track
track.base_run_time_seconds = run_time_seconds or int(length / 1000)
track.musicbrainz_id = mbid
if commit:
track.save()
return track
def fix_metadata(self, force_update=False):
...