Upgraded celery to 4.1, added endpoint logic for fingerprinting audio files

api/funkwhale_api/music/migrations/0016_trackfile_acoustid_track_id.py

@@ -0,0 +1,18 @@
+# Generated by Django 2.0 on 2017-12-26 16:39
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('music', '0015_bind_track_file_to_import_job'),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name='trackfile',
+            name='acoustid_track_id',
+            field=models.UUIDField(blank=True, null=True),
+        ),
+    ]

api/funkwhale_api/music/models.py

@@ -15,11 +15,9 @@ from django.utils import timezone
 from taggit.managers import TaggableManager
 from versatileimagefield.fields import VersatileImageField
 
-from funkwhale_api.taskapp import celery
 from funkwhale_api import downloader
 from funkwhale_api import musicbrainz
 from . import importers
-from . import lyrics as lyrics_utils
 
 
 class APIModelMixin(models.Model):
@@ -255,14 +253,6 @@ class Lyrics(models.Model):
     url = models.URLField(unique=True)
     content = models.TextField(null=True, blank=True)
 
-    @celery.app.task(name='Lyrics.fetch_content', filter=celery.task_method)
-    def fetch_content(self):
-        html = lyrics_utils._get_html(self.url)
-        content = lyrics_utils.extract_content(html)
-        cleaned_content = lyrics_utils.clean_content(content)
-        self.content = cleaned_content
-        self.save()
-
     @property
     def content_rendered(self):
         return markdown.markdown(
@@ -362,6 +352,7 @@ class TrackFile(models.Model):
     audio_file = models.FileField(upload_to='tracks/%Y/%m/%d', max_length=255)
     source = models.URLField(null=True, blank=True)
     duration = models.IntegerField(null=True, blank=True)
+    acoustid_track_id = models.UUIDField(null=True, blank=True)
 
     def download_file(self):
         # import the track file, since there is not any
@@ -429,26 +420,3 @@ class ImportJob(models.Model):
 
     class Meta:
         ordering = ('id', )
-    @celery.app.task(name='ImportJob.run', filter=celery.task_method)
-    def run(self, replace=False):
-        try:
-            track, created = Track.get_or_create_from_api(mbid=self.mbid)
-            track_file = None
-            if replace:
-                track_file = track.files.first()
-            elif track.files.count() > 0:
-                return
-
-            track_file = track_file or TrackFile(
-                track=track, source=self.source)
-            track_file.download_file()
-            track_file.save()
-            self.status = 'finished'
-            self.track_file = track_file
-            self.save()
-            return track.pk
-
-        except Exception as exc:
-            if not settings.DEBUG:
-                raise ImportJob.run.retry(args=[self], exc=exc, countdown=30, max_retries=3)
-            raise

api/funkwhale_api/music/tasks.py

@@ -0,0 +1,56 @@
+from funkwhale_api.taskapp import celery
+from funkwhale_api.providers.acoustid import get_acoustid_client
+
+from django.conf import settings
+from . import models
+from . import lyrics as lyrics_utils
+
+
+@celery.app.task(name='acoustid.set_on_track_file')
+@celery.require_instance(models.TrackFile, 'track_file')
+def set_acoustid_on_track_file(track_file):
+    client = get_acoustid_client()
+    result = client.get_best_match(track_file.audio_file.path)
+
+    def update(id):
+        track_file.acoustid_track_id = id
+        track_file.save(update_fields=['acoustid_track_id'])
+        return id
+    if result:
+        return update(result['id'])
+
+
+@celery.app.task(name='ImportJob.run', bind=True)
+@celery.require_instance(models.ImportJob, 'import_job')
+def import_job_run(self, import_job, replace=False):
+    try:
+        track, created = models.Track.get_or_create_from_api(mbid=import_job.mbid)
+        track_file = None
+        if replace:
+            track_file = track.files.first()
+        elif track.files.count() > 0:
+            return
+
+        track_file = track_file or models.TrackFile(
+            track=track, source=import_job.source)
+        track_file.download_file()
+        track_file.save()
+        import_job.status = 'finished'
+        import_job.track_file = track_file
+        import_job.save()
+        return track.pk
+
+    except Exception as exc:
+        if not settings.DEBUG:
+            raise import_job_run.retry(args=[self], exc=exc, countdown=30, max_retries=3)
+        raise
+
+
+@celery.app.task(name='Lyrics.fetch_content')
+@celery.require_instance(models.Lyrics, 'lyrics')
+def fetch_content(lyrics):
+    html = lyrics_utils._get_html(lyrics.url)
+    content = lyrics_utils.extract_content(html)
+    cleaned_content = lyrics_utils.clean_content(content)
+    lyrics.content = cleaned_content
+    lyrics.save(update_fields=['content'])

api/funkwhale_api/music/views.py

@@ -22,6 +22,7 @@ from . import models
 from . import serializers
 from . import importers
 from . import filters
+from . import tasks
 from . import utils
 
 
@@ -129,7 +130,8 @@ class TrackViewSet(TagViewSetMixin, SearchMixin, viewsets.ReadOnlyModelViewSet):
         lyrics = work.fetch_lyrics()
         try:
             if not lyrics.content:
-                lyrics.fetch_content()
+                tasks.fetch_content(lyrics_id=lyrics.pk)
+                lyrics.refresh_from_db()
         except AttributeError:
             return Response({'error': 'unavailable lyrics'}, status=404)
         serializer = serializers.LyricsSerializer(lyrics)
@@ -244,7 +246,7 @@ class SubmitViewSet(viewsets.ViewSet):
             pass
         batch = models.ImportBatch.objects.create(submitted_by=request.user)
         job = models.ImportJob.objects.create(mbid=request.POST['mbid'], batch=batch, source=request.POST['import_url'])
-        job.run.delay()
+        tasks.import_job_run.delay(import_job_id=job.pk)
         serializer = serializers.ImportBatchSerializer(batch)
         return Response(serializer.data)
 
@@ -272,7 +274,7 @@ class SubmitViewSet(viewsets.ViewSet):
                 models.TrackFile.objects.get(track__mbid=row['mbid'])
             except models.TrackFile.DoesNotExist:
                 job = models.ImportJob.objects.create(mbid=row['mbid'], batch=batch, source=row['source'])
-                job.run.delay()
+                tasks.import_job_run.delay(import_job_id=job.pk)
         serializer = serializers.ImportBatchSerializer(batch)
         return serializer.data, batch
 

api/funkwhale_api/providers/acoustid/__init__.py

@@ -0,0 +1,27 @@
+import acoustid
+
+from dynamic_preferences.registries import global_preferences_registry
+
+
+class Client(object):
+    def __init__(self, api_key):
+        self.api_key = api_key
+
+    def match(self, file_path):
+        return acoustid.match(self.api_key, file_path, parse=False)
+
+    def get_best_match(self, file_path):
+        results = self.match(file_path=file_path)
+        MIN_SCORE_FOR_MATCH = 0.8
+        try:
+            rows = results['results']
+        except KeyError:
+            return
+        for row in rows:
+            if row['score'] >= MIN_SCORE_FOR_MATCH:
+                return row
+
+
+def get_acoustid_client():
+    manager = global_preferences_registry.manager()
+    return Client(api_key=manager['providers_acoustid__api_key'])

api/funkwhale_api/providers/acoustid/dynamic_preferences_registry.py

@@ -0,0 +1,13 @@
+from dynamic_preferences.types import StringPreference, Section
+from dynamic_preferences.registries import global_preferences_registry
+
+acoustid = Section('providers_acoustid')
+
+
+@global_preferences_registry.register
+class APIKey(StringPreference):
+    section = acoustid
+    name = 'api_key'
+    default = ''
+    verbose_name = 'Acoustid API key'
+    help_text = 'The API key used to query AcoustID. Get one at https://acoustid.org/new-application.'

api/funkwhale_api/providers/audiofile/tasks.py

@@ -1,20 +1,20 @@
+import acoustid
 import os
 import datetime
 from django.core.files import File
 
 from funkwhale_api.taskapp import celery
+from funkwhale_api.providers.acoustid import get_acoustid_client
 from funkwhale_api.music import models, metadata
 
 
-@celery.app.task(name='audiofile.from_path')
-def from_path(path):
+def import_metadata_without_musicbrainz(path):
     data = metadata.Metadata(path)
     artist = models.Artist.objects.get_or_create(
         name__iexact=data.get('artist'),
         defaults={
             'name': data.get('artist'),
             'mbid': data.get('musicbrainz_artistid', None),
-
         },
     )[0]
 
@@ -39,11 +39,33 @@ def from_path(path):
             'mbid': data.get('musicbrainz_recordingid', None),
         },
     )[0]
+    return track
+
+
+def import_metadata_with_musicbrainz(path):
+    pass
+
+@celery.app.task(name='audiofile.from_path')
+def from_path(path):
+    acoustid_track_id = None
+    try:
+        client = get_acoustid_client()
+        result = client.get_best_match(path)
+        acoustid_track_id = result['id']
+    except acoustid.WebServiceError:
+        track = import_metadata_without_musicbrainz(path)
+    except (TypeError, KeyError):
+        track = import_metadata_without_musicbrainz(path)
+    else:
+        track, created = models.Track.get_or_create_from_api(
+            mbid=result['recordings'][0]['id']
+        )
 
     if track.files.count() > 0:
         raise ValueError('File already exists for track {}'.format(track.pk))
 
-    track_file = models.TrackFile(track=track)
+    track_file = models.TrackFile(
+        track=track, acoustid_track_id=acoustid_track_id)
     track_file.audio_file.save(
         os.path.basename(path),
         File(open(path, 'rb'))

api/funkwhale_api/taskapp/celery.py

@@ -1,10 +1,12 @@
 
 from __future__ import absolute_import
 import os
+import functools
+
 from celery import Celery
 from django.apps import AppConfig
 from django.conf import settings
-from celery.contrib.methods import task_method
+
 
 if not settings.configured:
     # set the default Django settings module for the 'celery' program.
@@ -21,12 +23,20 @@ class CeleryConfig(AppConfig):
     def ready(self):
         # Using a string here means the worker will not have to
         # pickle the object when using Windows.
-        app.config_from_object('django.conf:settings')
+        app.config_from_object('django.conf:settings', namespace='CELERY')
         app.autodiscover_tasks(lambda: settings.INSTALLED_APPS, force=True)
 
 
-
-
-@app.task(bind=True)
-def debug_task(self):
-    print('Request: {0!r}'.format(self.request))  # pragma: no cover
+def require_instance(model_or_qs, parameter_name):
+    def decorator(function):
+        @functools.wraps(function)
+        def inner(*args, **kwargs):
+            pk = kwargs.pop('_'.join([parameter_name, 'id']))
+            try:
+                instance = model_or_qs.get(pk=pk)
+            except AttributeError:
+                instance = model_or_qs.objects.get(pk=pk)
+            kwargs[parameter_name] = instance
+            return function(*args, **kwargs)
+        return inner
+    return decorator