media indexing and management

Pipfile

@@ -16,6 +16,7 @@ django-compressor = "*"
 httptools = "*"
 youtube-dl = "*"
 django-background-tasks = "*"
+requests = "*"
 
 [requires]
 python_version = "3"

Pipfile.lock

@@ -1,7 +1,7 @@
 {
     "_meta": {
         "hash": {
-            "sha256": "229a3783dcfd9d030ac58856ecb5d52d806db6acfa40eb6346eb6b58d79e91ab"
+            "sha256": "a4bb556fc61ee4583f9588980450b071814298ee4d1a1023fad149c14d14aaba"
         },
         "pipfile-spec": 6,
         "requires": {
@@ -23,6 +23,20 @@
             ],
             "version": "==3.3.1"
         },
+        "certifi": {
+            "hashes": [
+                "sha256:1f422849db327d534e3d0c5f02a263458c3955ec0aae4ff09b95f195c59f4edd",
+                "sha256:f05def092c44fbf25834a51509ef6e631dc19765ab8a57b4e7ab85531f0a9cf4"
+            ],
+            "version": "==2020.11.8"
+        },
+        "chardet": {
+            "hashes": [
+                "sha256:84ab92ed1c4d4f16916e05906b6b75a6c0fb5db821cc65e70cbd64a3e2a5eaae",
+                "sha256:fc323ffcaeaed0e0a02bf4d117757b98aed530d9ed4531e3e15460124c106691"
+            ],
+            "version": "==3.0.4"
+        },
         "django": {
             "hashes": [
                 "sha256:5c866205f15e7a7123f1eec6ab939d22d5bde1416635cab259684af66d8e48a2",
@@ -92,6 +106,13 @@
             "index": "pypi",
             "version": "==0.1.1"
         },
+        "idna": {
+            "hashes": [
+                "sha256:b307872f855b18632ce0c21c5e45be78c0ea7ae4c15c828c20788b26921eb3f6",
+                "sha256:b97d804b1e9b523befed77c48dacec60e6dcb0b5391d57af6a65a312a90648c0"
+            ],
+            "version": "==2.10"
+        },
         "libsass": {
             "hashes": [
                 "sha256:1521d2a8d4b397c6ec90640a1f6b5529077035efc48ef1c2e53095544e713d1b",
@@ -158,6 +179,14 @@
             ],
             "version": "==1.0.6"
         },
+        "requests": {
+            "hashes": [
+                "sha256:7f1a0b932f4a60a1a65caa4263921bb7d9ee911957e0ae4a23a6dd08185ad5f8",
+                "sha256:e786fa28d8c9154e6a4de5d46a1d921b8749f8b74e28bde23768e5e16eece998"
+            ],
+            "index": "pypi",
+            "version": "==2.25.0"
+        },
         "rjsmin": {
             "hashes": [
                 "sha256:0ab825839125eaca57cc59581d72e596e58a7a56fbc0839996b7528f0343a0a8",
@@ -190,6 +219,13 @@
             ],
             "version": "==0.4.1"
         },
+        "urllib3": {
+            "hashes": [
+                "sha256:19188f96923873c92ccb987120ec4acaa12f0461fa9ce5d3d0772bc965a39e08",
+                "sha256:d8ff90d979214d7b4f8ce956e80f4028fc6860e4431f731ea4a8c08f23f99473"
+            ],
+            "version": "==1.26.2"
+        },
         "whitenoise": {
             "hashes": [
                 "sha256:05ce0be39ad85740a78750c86a93485c40f08ad8c62a6006de0233765996e5c7",
@@ -200,11 +236,11 @@
         },
         "youtube-dl": {
             "hashes": [
-                "sha256:bc82acb0b59b25b822fad85bef0cbe78e5754ca532e3bd6899fe06386e2b8e7c",
+                "sha256:c73c79ccaabb7eabc223b36889ad9d3fbe04433d933312e8752d6a2c2bdc028d",
-                "sha256:daa514c0d36af478fe249ea93ca63ae3aaa60ac5f82aa3d5514bce9ba7ed8451"
+                "sha256:cd7a072314ee67472de75a0f770bffc1363be46921c28bfc47a786a5188c417f"
             ],
             "index": "pypi",
-            "version": "==2020.12.2"
+            "version": "==2020.12.5"
         }
     },
     "develop": {}

app/common/static/styles/_colours.scss

@@ -40,6 +40,9 @@ $collection-no-items-text-colour: $colour-near-black;
 $collection-background-hover-colour: $colour-orange;
 $collection-text-hover-colour: $colour-near-white;
 
+$mediacard-title-background-colour: $colour-white;
+$mediacard-title-text-colour: $colour-black;
+
 $box-error-background-colour: $colour-red;
 $box-error-text-colour: $colour-near-white;
 

app/common/static/styles/_template.scss

@@ -104,6 +104,23 @@ main {
         }
     }
 
+    .mediacard {
+        .card-title {
+            padding: 3px 10px 3px 5px !important;
+            margin: 5px !important;
+            line-height: 1.5rem;
+            font-size: 1.5rem;
+            left: initial !important;
+            right: 0 !important;
+            max-width: 75% !important;
+            opacity: 80%;
+            background-color: $mediacard-title-background-colour !important;
+            color: $mediacard-title-text-colour !important;
+            span {
+                font-size: 1rem;
+            }
+        }
+    }
 
     .pagination {
         width: 100%;

app/common/templates/infobox.html

@@ -3,7 +3,7 @@
   <div class="col s12">
     <div class="card infobox">
       <div class="card-content">
-        <i class="fas fa-info-circle"></i> {{ message }}
+        <i class="fas fa-info-circle"></i> {{ message|safe }}
       </div>
     </div>
   </div>

app/sync/migrations/0005_auto_20201205_0411.py

@@ -0,0 +1,23 @@
+# Generated by Django 3.1.3 on 2020-12-05 04:11
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('sync', '0004_remove_source_url'),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name='media',
+            name='thumb_height',
+            field=models.PositiveSmallIntegerField(blank=True, help_text='Height (Y) of the thumbnail', null=True, verbose_name='thumb height'),
+        ),
+        migrations.AlterField(
+            model_name='media',
+            name='thumb_width',
+            field=models.PositiveSmallIntegerField(blank=True, help_text='Width (X) of the thumbnail', null=True, verbose_name='thumb width'),
+        ),
+    ]

app/sync/migrations/0006_auto_20201205_0502.py

@@ -0,0 +1,36 @@
+# Generated by Django 3.1.3 on 2020-12-05 05:02
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('sync', '0005_auto_20201205_0411'),
+    ]
+
+    operations = [
+        migrations.RemoveField(
+            model_name='source',
+            name='output_format',
+        ),
+        migrations.RemoveField(
+            model_name='source',
+            name='source_profile',
+        ),
+        migrations.AddField(
+            model_name='source',
+            name='source_resolution',
+            field=models.CharField(choices=[('360p', '360p (SD)'), ('480p', '480p (SD)'), ('720p', '720p (HD)'), ('1080p', '1080p (Full HD)'), ('2160p', '2160p (4K)'), ('audio', 'Audio only')], db_index=True, default='1080p', help_text='Source resolution, desired video resolution to download', max_length=8, verbose_name='source resolution'),
+        ),
+        migrations.AddField(
+            model_name='source',
+            name='source_vcodec',
+            field=models.CharField(choices=[('M4A', 'M4A'), ('OPUS', 'OPUS'), ('AAC', 'AAC')], db_index=True, default='OPUS', help_text='Source audio codec, desired audio encoding format to download', max_length=8, verbose_name='source audio codec'),
+        ),
+        migrations.AlterField(
+            model_name='source',
+            name='prefer_60fps',
+            field=models.BooleanField(default=True, help_text='Where possible, prefer 60fps media for this source', verbose_name='prefer 60fps'),
+        ),
+    ]

app/sync/migrations/0007_auto_20201205_0509.py

@@ -0,0 +1,18 @@
+# Generated by Django 3.1.3 on 2020-12-05 05:09
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('sync', '0006_auto_20201205_0502'),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name='source',
+            name='source_vcodec',
+            field=models.CharField(choices=[('M4A', 'M4A'), ('OPUS', 'OPUS')], db_index=True, default='OPUS', help_text='Source audio codec, desired audio encoding format to download', max_length=8, verbose_name='source audio codec'),
+        ),
+    ]

app/sync/migrations/0008_auto_20201205_0512.py

@@ -0,0 +1,23 @@
+# Generated by Django 3.1.3 on 2020-12-05 05:12
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('sync', '0007_auto_20201205_0509'),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name='source',
+            name='source_acodec',
+            field=models.CharField(choices=[('M4A', 'M4A'), ('OPUS', 'OPUS')], db_index=True, default='OPUS', help_text='Source audio codec, desired audio encoding format to download', max_length=8, verbose_name='source audio codec'),
+        ),
+        migrations.AlterField(
+            model_name='source',
+            name='source_vcodec',
+            field=models.CharField(choices=[('AVC1', 'AVC1 (H.264)'), ('VP9', 'VP9')], db_index=True, default='VP9', help_text='Source video codec, desired video encoding format to download', max_length=8, verbose_name='source video codec'),
+        ),
+    ]

app/sync/migrations/0009_auto_20201205_0512.py

@@ -0,0 +1,18 @@
+# Generated by Django 3.1.3 on 2020-12-05 05:12
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('sync', '0008_auto_20201205_0512'),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name='source',
+            name='source_vcodec',
+            field=models.CharField(choices=[('AVC1', 'AVC1 (H.264)'), ('VP9', 'VP9')], db_index=True, default='VP9', help_text='Source video codec, desired video encoding format to download (ignored if "resolution" is audio only)', max_length=8, verbose_name='source video codec'),
+        ),
+    ]

app/sync/models.py

@@ -1,7 +1,12 @@
 import uuid
+import json
+from datetime import datetime
+from pathlib import Path
 from django.conf import settings
 from django.db import models
+from django.utils.text import slugify
 from django.utils.translation import gettext_lazy as _
+from .youtube import get_media_info as get_youtube_media_info
 
 
 class Source(models.Model):
@@ -18,36 +23,40 @@ class Source(models.Model):
         (SOURCE_TYPE_YOUTUBE_PLAYLIST, _('YouTube playlist')),
     )
 
-    SOURCE_PROFILE_360p = '360p'
+    SOURCE_RESOLUTION_360p = '360p'
-    SOURCE_PROFILE_480p = '480p'
+    SOURCE_RESOLUTION_480p = '480p'
-    SOURCE_PROFILE_720P = '720p'
+    SOURCE_RESOLUTION_720P = '720p'
-    SOURCE_PROFILE_1080P = '1080p'
+    SOURCE_RESOLUTION_1080P = '1080p'
-    SOURCE_PROFILE_2160P = '2160p'
+    SOURCE_RESOLUTION_2160P = '2160p'
-    SOURCE_PROFILE_AUDIO = 'audio'
+    SOURCE_RESOLUTION_AUDIO = 'audio'
-    SOURCE_PROFILES = (SOURCE_PROFILE_360p, SOURCE_PROFILE_480p, SOURCE_PROFILE_720P,
+    SOURCE_RESOLUTIONS = (SOURCE_RESOLUTION_360p, SOURCE_RESOLUTION_480p,
-                       SOURCE_PROFILE_1080P, SOURCE_PROFILE_2160P,
+                          SOURCE_RESOLUTION_720P, SOURCE_RESOLUTION_1080P,
-                       SOURCE_PROFILE_AUDIO)
+                          SOURCE_RESOLUTION_2160P, SOURCE_RESOLUTION_AUDIO)
-    SOURCE_PROFILE_CHOICES = (
+    SOURCE_RESOLUTION_CHOICES = (
-        (SOURCE_PROFILE_360p, _('360p (SD)')),
+        (SOURCE_RESOLUTION_360p, _('360p (SD)')),
-        (SOURCE_PROFILE_480p, _('480p (SD)')),
+        (SOURCE_RESOLUTION_480p, _('480p (SD)')),
-        (SOURCE_PROFILE_720P, _('720p (HD)')),
+        (SOURCE_RESOLUTION_720P, _('720p (HD)')),
-        (SOURCE_PROFILE_1080P, _('1080p (Full HD)')),
+        (SOURCE_RESOLUTION_1080P, _('1080p (Full HD)')),
-        (SOURCE_PROFILE_2160P, _('2160p (4K)')),
+        (SOURCE_RESOLUTION_2160P, _('2160p (4K)')),
-        (SOURCE_PROFILE_AUDIO, _('Audio only')),
+        (SOURCE_RESOLUTION_AUDIO, _('Audio only')),
     )
 
-    OUTPUT_FORMAT_MP4 = 'mp4'
+    SOURCE_VCODEC_AVC1 = 'AVC1'
-    OUTPUT_FORMAT_MKV = 'mkv'
+    SOURCE_VCODEC_VP9 = 'VP9'
-    OUTPUT_FORMAT_M4A = 'm4a'
+    SOURCE_VCODECS = (SOURCE_VCODEC_AVC1, SOURCE_VCODEC_VP9)
-    OUTPUT_FORMAT_OGG = 'ogg'
+    SOURCE_VCODECS_PRIORITY = (SOURCE_VCODEC_VP9, SOURCE_VCODEC_AVC1)
-    OUTPUT_FORMATS = (OUTPUT_FORMAT_MP4, OUTPUT_FORMAT_MKV, OUTPUT_FORMAT_M4A,
+    SOURCE_VCODEC_CHOICES = (
-                      OUTPUT_FORMAT_OGG)
+        (SOURCE_VCODEC_AVC1, _('AVC1 (H.264)')),
-    OUTPUT_FORMAT_CHOICES = (
+        (SOURCE_VCODEC_VP9, _('VP9')),
-        (OUTPUT_FORMAT_MP4, _('.mp4 container')),
+    )
-        (OUTPUT_FORMAT_MKV, _('.mkv container')),
+
-        (OUTPUT_FORMAT_MKV, _('.webm container')),
+    SOURCE_ACODEC_M4A = 'M4A'
-        (OUTPUT_FORMAT_M4A, _('.m4a container (audio only)')),
+    SOURCE_ACODEC_OPUS = 'OPUS'
-        (OUTPUT_FORMAT_OGG, _('.ogg container (audio only)')),
+    SOURCE_ACODECS = (SOURCE_ACODEC_M4A, SOURCE_ACODEC_OPUS)
+    SOURCE_ACODEC_PRIORITY = (SOURCE_ACODEC_OPUS, SOURCE_ACODEC_M4A)
+    SOURCE_ACODEC_CHOICES = (
+        (SOURCE_ACODEC_M4A, _('M4A')),
+        (SOURCE_ACODEC_OPUS, _('OPUS')),
     )
 
     FALLBACK_FAIL = 'f'
@@ -56,8 +65,8 @@ class Source(models.Model):
     FALLBACKS = (FALLBACK_FAIL, FALLBACK_NEXT_SD, FALLBACK_NEXT_HD)
     FALLBACK_CHOICES = (
         (FALLBACK_FAIL, _('Fail, do not download any media')),
-        (FALLBACK_NEXT_SD, _('Get next best SD media instead')),
+        (FALLBACK_NEXT_SD, _('Get next best SD media or codec instead')),
-        (FALLBACK_NEXT_HD, _('Get next best HD media instead')),
+        (FALLBACK_NEXT_HD, _('Get next best HD media or codec instead')),
     )
 
     ICONS = {
@@ -70,6 +79,16 @@ class Source(models.Model):
         SOURCE_TYPE_YOUTUBE_PLAYLIST: 'https://www.youtube.com/playlist?list={key}',
     }
 
+    INDEXERS = {
+        SOURCE_TYPE_YOUTUBE_CHANNEL: get_youtube_media_info,
+        SOURCE_TYPE_YOUTUBE_PLAYLIST: get_youtube_media_info,
+    }
+
+    KEY_FIELD = {  # Field returned by indexing which contains a unique key
+        SOURCE_TYPE_YOUTUBE_CHANNEL: 'id',
+        SOURCE_TYPE_YOUTUBE_PLAYLIST: 'id',
+    }
+
     uuid = models.UUIDField(
         _('uuid'),
         primary_key=True,
@@ -127,17 +146,33 @@ class Source(models.Model):
         help_text=_('If "delete old media" is ticked, the number of days after which '
                     'to automatically delete media')
     )
-    source_profile = models.CharField(
+    source_resolution = models.CharField(
-        _('source profile'),
+        _('source resolution'),
         max_length=8,
         db_index=True,
-        choices=SOURCE_PROFILE_CHOICES,
+        choices=SOURCE_RESOLUTION_CHOICES,
-        default=SOURCE_PROFILE_1080P,
+        default=SOURCE_RESOLUTION_1080P,
-        help_text=_('Source profile, the quality to attempt to download media')
+        help_text=_('Source resolution, desired video resolution to download')
+    )
+    source_vcodec = models.CharField(
+        _('source video codec'),
+        max_length=8,
+        db_index=True,
+        choices=SOURCE_VCODEC_CHOICES,
+        default=SOURCE_VCODEC_VP9,
+        help_text=_('Source video codec, desired video encoding format to download (ignored if "resolution" is audio only)')
+    )
+    source_acodec = models.CharField(
+        _('source audio codec'),
+        max_length=8,
+        db_index=True,
+        choices=SOURCE_ACODEC_CHOICES,
+        default=SOURCE_ACODEC_OPUS,
+        help_text=_('Source audio codec, desired audio encoding format to download')
     )
     prefer_60fps = models.BooleanField(
         _('prefer 60fps'),
-        default=False,
+        default=True,
         help_text=_('Where possible, prefer 60fps media for this source')
     )
     prefer_hdr = models.BooleanField(
@@ -145,21 +180,13 @@ class Source(models.Model):
         default=False,
         help_text=_('Where possible, prefer HDR media for this source')
     )
-    output_format = models.CharField(
-        _('output format'),
-        max_length=8,
-        db_index=True,
-        choices=OUTPUT_FORMAT_CHOICES,
-        default=OUTPUT_FORMAT_MKV,
-        help_text=_('Output format, the file format container in which to save media')
-    )
     fallback = models.CharField(
         _('fallback'),
         max_length=1,
         db_index=True,
         choices=FALLBACK_CHOICES,
         default=FALLBACK_FAIL,
-        help_text=_('What do do when media in your source profile is not available')
+        help_text=_('What do do when media in your source resolution and codecs is not available')
     )
 
     def __str__(self):
@@ -182,19 +209,47 @@ class Source(models.Model):
     def url(self):
         return Source.create_url(self.source_type, self.key)
 
+    @property
+    def format_summary(self):
+        if self.source_resolution == Source.SOURCE_RESOLUTION_AUDIO:
+            vc = 'none'
+        else:
+            vc = self.source_vcodec
+        ac = self.source_acodec
+        f = '60FPS' if self.prefer_60fps else ''
+        h = 'HDR' if self.prefer_hdr else ''
+        return f'{self.source_resolution} (video:{vc}, audio:{ac}) {f} {h}'.strip()
+
     @property
     def directory_path(self):
-        if self.source_profile == self.SOURCE_PROFILE_AUDIO:
+        if self.source_resolution == self.SOURCE_RESOLUTION_AUDIO:
             return settings.SYNC_AUDIO_ROOT / self.directory
         else:
             return settings.SYNC_VIDEO_ROOT / self.directory
 
+    @property
+    def key_field(self):
+        return self.KEY_FIELD.get(self.source_type, '')
+
+    def index_media(self):
+        '''
+            Index the media source returning a list of media metadata as dicts.
+        '''
+        indexer = self.INDEXERS.get(self.source_type, None)
+        if not callable(indexer):
+            raise Exception(f'Source type f"{self.source_type}" has no indexer')
+        response = indexer(self.url)
+        return response.get('entries', [])
+
 
 def get_media_thumb_path(instance, filename):
     fileid = str(instance.uuid)
-    filename = f'{fileid.lower()}.{instance.image_type.lower()}'
+    filename = f'{fileid.lower()}.jpg'
     prefix = fileid[:2]
-    return os.path.join('thumbs', prefix, filename)
+    return Path('thumbs') / prefix / filename
+
+
+_metadata_cache = {}
 
 
 class Media(models.Model):
@@ -203,6 +258,11 @@ class Media(models.Model):
         Source.
     '''
 
+    URLS = {
+        Source.SOURCE_TYPE_YOUTUBE_CHANNEL: 'https://www.youtube.com/watch?v={key}',
+        Source.SOURCE_TYPE_YOUTUBE_PLAYLIST: 'https://www.youtube.com/watch?v={key}',
+    }
+
     uuid = models.UUIDField(
         _('uuid'),
         primary_key=True,
@@ -248,11 +308,13 @@ class Media(models.Model):
     thumb_width = models.PositiveSmallIntegerField(
         _('thumb width'),
         blank=True,
+        null=True,
         help_text=_('Width (X) of the thumbnail')
     )
     thumb_height = models.PositiveSmallIntegerField(
         _('thumb height'),
         blank=True,
+        null=True,
         help_text=_('Height (Y) of the thumbnail')
     )
     metadata = models.TextField(
@@ -317,3 +379,53 @@ class Media(models.Model):
     class Meta:
         verbose_name = _('Media')
         verbose_name_plural = _('Media')
+
+    @property
+    def loaded_metadata(self):
+        if self.pk in _metadata_cache:
+            return _metadata_cache[self.pk]
+        _metadata_cache[self.pk] = json.loads(self.metadata)
+        return _metadata_cache[self.pk]
+
+    @property
+    def extension(self):
+        '''
+            The extension is also used by youtube-dl to set the output container. As
+            it is possible to quite easily pick combinations of codecs and containers
+            which are invalid (e.g. OPUS audio in an MP4 container) just set this for
+            people. All video is set to mkv containers, audio-only is set to m4a or ogg
+            depending on audio codec.
+        '''
+        if self.source.source_resolution == Source.SOURCE_RESOLUTION_AUDIO:
+            if self.source.source_acodec == Source.SOURCE_ACODEC_M4A:
+                return 'm4a'
+            elif self.source.source_acodec == Source.SOURCE_ACODEC_OPUS:
+                return 'ogg'
+            else:
+                raise ValueError('Unable to choose audio extension, uknown acodec')
+        else:
+            return 'mkv'
+
+    @property
+    def url(self):
+        url = self.URLS.get(self.source.source_type, '')
+        return url.format(key=self.key)
+
+    @property
+    def title(self):
+        return self.loaded_metadata.get('title', '').strip()
+
+    @property
+    def upload_date(self):
+        upload_date_str = self.loaded_metadata.get('upload_date', '').strip()
+        try:
+            return datetime.strptime(upload_date_str, '%Y%m%d')
+        except ValueError as e:
+            return None
+
+    @property
+    def filename(self):
+        upload_date = self.upload_date.strftime('%Y-%m-%d')
+        title = slugify(self.title.replace('&', 'and').replace('+', 'and'))
+        ext = self.extension
+        return f'{upload_date}_{title}.{ext}'

app/sync/signals.py

@@ -0,0 +1,38 @@
+from django.conf import settings
+from django.db.models.signals import post_save, post_delete
+from django.dispatch import receiver
+
+from .models import Source, Media
+from .tasks import delete_index_source_task, index_source_task, download_media_thumbnail
+
+
+@receiver(post_save, sender=Source)
+def source_post_save(sender, instance, created, **kwargs):
+    # Triggered when a source is saved
+    if created:
+        # If the source is newly created schedule its indexing
+        index_source_task(str(instance.pk), repeat=settings.INDEX_SOURCE_EVERY)
+
+
+@receiver(post_delete, sender=Source)
+def source_post_delete(sender, instance, **kwargs):
+    # Triggered when a source is deleted
+    delete_index_source_task(str(instance.pk))
+
+
+@receiver(post_save, sender=Media)
+def media_post_save(sender, instance, created, **kwargs):
+    # Triggered when media is saved
+    if created:
+        # If the media is newly created fire a task off to download its thumbnail
+        metadata = instance.loaded_metadata
+        thumbnail_url = metadata.get('thumbnail', '')
+        if thumbnail_url:
+            download_media_thumbnail(str(instance.pk), thumbnail_url)
+
+
+@receiver(post_delete, sender=Media)
+def media_post_delete(sender, instance, **kwargs):
+    # Triggered when media is deleted
+    pass
+    # TODO: delete thumbnail and media file from disk

app/sync/tasks.py

@@ -0,0 +1,84 @@
+'''
+    Start, stop and manage scheduled tasks. These are generally triggered by Django
+    signals (see signals.py).
+'''
+
+
+import json
+from io import BytesIO
+from django.conf import settings
+from django.core.files.uploadedfile import SimpleUploadedFile
+from background_task import background
+from background_task.models import Task
+from .models import Source, Media
+from .utils import get_remote_image
+
+
+def delete_index_source_task(source_id):
+    task = None
+    try:
+        # get_task currently returns a QuerySet, but catch DoesNotExist just in case
+        task = Task.objects.get_task('sync.tasks.index_source_task', args=(source_id,))
+    except Task.DoesNotExist:
+        pass
+    if task:
+        # A scheduled task exists for this Source, delete it
+        task.delete()
+
+
+@background(schedule=0)
+def index_source_task(source_id):
+    '''
+        Indexes media available from a Source object.
+    '''
+    try:
+        source = Source.objects.get(pk=source_id)
+    except Source.DoesNotExist:
+        # Task triggered but the Source has been deleted, delete the task
+        delete_index_source_task(source_id)
+        return
+    videos = source.index_media()
+    for video in videos:
+        # Create or update each video as a Media object
+        key = video.get(source.key_field, None)
+        if not key:
+            # Video has no unique key (ID), it can't be indexed
+            continue
+        try:
+            media = Media.objects.get(key=key)
+        except Media.DoesNotExist:
+            media = Media(key=key)
+        media.source = source
+        media.metadata = json.dumps(video)
+        media.save()
+
+
+@background(schedule=0)
+def download_media_thumbnail(media_id, url):
+    '''
+        Downloads an image from a URL and saves it as a local thumbnail attached to a
+        Media object.
+    '''
+    try:
+        media = Media.objects.get(pk=media_id)
+    except Media.DoesNotExist:
+        # Task triggered but the media no longer exists, ignore task
+        return
+    i = get_remote_image(url)
+    max_width, max_height = getattr(settings, 'MAX_MEDIA_THUMBNAIL_SIZE', (512, 512))
+    if i.width > max_width or i.height > max_height:
+        # Image is larger than we want to save, resize it
+        i.thumbnail(size=(max_width, max_height))
+    image_file = BytesIO()
+    i.save(image_file, 'JPEG', quality=80, optimize=True, progressive=True)
+    image_file.seek(0)
+    media.thumb.save(
+        'thumb',
+        SimpleUploadedFile(
+            'thumb',
+            image_file.read(),
+            'image/jpeg',
+        ),
+        save=True
+    )
+    return True

app/sync/templates/sync/media-item.html

@@ -0,0 +1,39 @@
+{% extends 'base.html' %}
+
+{% block headtitle %}Media - {{ media.key }}{% endblock %}
+
+{% block content %}
+<div class="row">
+  <div class="col s12">
+    <h1 class="truncate">Media <strong>{{ media.key }}</strong></h1>
+    <p class="truncate"><strong><a href="{{ media.url }}" target="_blank"><i class="fas fa-link"></i> {{ media.url }}</a></strong></p>
+    <p class="truncate">Saving to: <strong>{{ media.source.directory_path }}</strong></p>
+  </div>
+</div>
+<div class="row">
+  <div class="col s12">
+    <table class="striped">
+      <tr title="The media source">
+        <td class="hide-on-small-only">Source</td>
+        <td><span class="hide-on-med-and-up">Source<br></span><strong><a href="{% url 'sync:source' pk=media.source.pk %}">{{ media.source }}</a></strong></td>
+      </tr>
+      <tr title="The media title">
+        <td class="hide-on-small-only">Title</td>
+        <td><span class="hide-on-med-and-up">Title<br></span><strong>{{ media.title }}</strong></td>
+      </tr>
+      <tr title="The filename the media will be downloaded as">
+        <td class="hide-on-small-only">Filename</td>
+        <td><span class="hide-on-med-and-up">Filename<br></span><strong>{{ media.filename }}</strong></td>
+      </tr>
+      <tr title="The desired format">
+        <td class="hide-on-small-only">Desired format</td>
+        <td><span class="hide-on-med-and-up">Desired format<br></span><strong>{{ media.source.format_summary }}</strong></td>
+      </tr>
+      <tr title="Has the media been downloaded">
+        <td class="hide-on-small-only">Downloaded</td>
+        <td><span class="hide-on-med-and-up">Downloaded<br></span><strong>{% if media.downloaded %}<i class="fas fa-check"></i>{% else %}<i class="fas fa-times"></i>{% endif %}</strong></td>
+      </tr>
+    </table>
+  </div>
+</div>
+{% endblock %}

app/sync/templates/sync/media.html

@@ -1,11 +1,30 @@
-{% extends 'base.html' %}
+{% extends 'base.html' %}{% load static %}
 
-{% block headtitle %}Media{% endblock %}
+{% block headtitle %}Media{% if source %} - {{ source }}{% endif %}{% endblock %}
 
 {% block content %}
-<div class="row">
+{% include 'infobox.html' with message=message %}
-  <div class="col s12">
+<div class="row no-margin-bottom">
-    media    
+  {% for m in media %}
+  <div class="col s12 m6 l4 xl3">
+    <a href="{% url 'sync:media-item' pk=m.pk %}" class="collection-item">
+      <div class="card mediacard">
+        <div class="card-image">
+          <img src="{% if m.thumb %}{% url 'sync:media-thumb' pk=m.pk %}{% else %}{% static 'images/nothumb.jpg' %} {% endif %}">
+          <span class="card-title truncate">{{ m.source }}<br>
+            <span>{{ m }}</span>
+          </span>
         </div>
+      </div>
+    </a>
+  </div>
+  {% empty %}
+  <div class="col s12">
+    <div class="collection">
+      <span class="collection-item no-items"><i class="fas fa-info-circle"></i> No media has been indexed.</span>
+    </div>
+  </div>
+  {% endfor %}
 </div>
+{% include 'pagination.html' with pagination=sources.paginator %}
 {% endblock %}

app/sync/urls.py

@@ -1,7 +1,7 @@
 from django.urls import path
 from .views import (DashboardView, SourcesView, ValidateSourceView, AddSourceView,
                     SourceView, UpdateSourceView, DeleteSourceView, MediaView,
-                    TasksView, LogsView)
+                    MediaThumbView, MediaItemView, TasksView, LogsView)
 
 
 app_name = 'sync'
@@ -41,12 +41,20 @@ urlpatterns = [
          DeleteSourceView.as_view(),
          name='delete-source'),
 
-    # Media URLs (note /media/ is the static media URL, don't use that)
+    # Media URLs
 
-    path('mediafiles',
+    path('media',
          MediaView.as_view(),
          name='media'),
 
+    path('media-thumb/<uuid:pk>',
+         MediaThumbView.as_view(),
+         name='media-thumb'),
+
+    path('media-item/<uuid:pk>',
+         MediaItemView.as_view(),
+         name='media-item'),
+
     # Task URLs
 
     path('tasks',

app/sync/utils.py

@@ -1,4 +1,6 @@
 import re
+import requests
+from PIL import Image
 from urllib.parse import urlsplit, parse_qs
 from django.forms import ValidationError
 
@@ -42,3 +44,13 @@ def validate_url(url, validator):
     elif extract_from == 'qs_args':
         extract_value = url_query_parts[extract_param][0]
     return extract_value
+
+
+def get_remote_image(url):
+    headers = {
+        'user-agent': ('Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 '
+                       '(KHTML, like Gecko) Chrome/69.0.3497.64 Safari/537.36')
+    }
+    r = requests.get(url, headers=headers, stream=True, timeout=60)
+    r.raw.decode_content = True
+    return Image.open(r.raw)

app/sync/views.py

@@ -1,16 +1,19 @@
+from base64 import b64decode
 from django.conf import settings
 from django.http import Http404
 from django.views.generic import TemplateView, ListView, DetailView
 from django.views.generic.edit import (FormView, FormMixin, CreateView, UpdateView,
                                        DeleteView)
+from django.http import HttpResponse
 from django.urls import reverse_lazy
 from django.forms import ValidationError
 from django.utils.text import slugify
 from django.utils.translation import gettext_lazy as _
 from common.utils import append_uri_params
-from .models import Source
+from .models import Source, Media
 from .forms import ValidateSourceForm, ConfirmDeleteSourceForm
 from .utils import validate_url
+from . import signals
 from . import youtube
 
 
@@ -196,8 +199,8 @@ class AddSourceView(CreateView):
     template_name = 'sync/source-add.html'
     model = Source
     fields = ('source_type', 'key', 'name', 'directory', 'delete_old_media',
-              'days_to_keep', 'source_profile', 'prefer_60fps', 'prefer_hdr',
+              'days_to_keep', 'source_resolution', 'source_vcodec', 'source_acodec',
-              'output_format', 'fallback')
+              'prefer_60fps', 'prefer_hdr', 'fallback')
 
     def __init__(self, *args, **kwargs):
         self.prepopulated_data = {}
@@ -240,8 +243,8 @@ class UpdateSourceView(UpdateView):
     template_name = 'sync/source-update.html'
     model = Source
     fields = ('source_type', 'key', 'name', 'directory', 'delete_old_media',
-              'days_to_keep', 'source_profile', 'prefer_60fps', 'prefer_hdr',
+              'days_to_keep', 'source_resolution', 'source_vcodec', 'source_acodec',
-              'output_format', 'fallback')
+              'prefer_60fps', 'prefer_hdr', 'fallback')
 
     def get_success_url(self):
         url = reverse_lazy('sync:sources')
@@ -272,16 +275,76 @@ class DeleteSourceView(DeleteView, FormMixin):
         return append_uri_params(url, {'message': 'source-deleted'})
 
 
-class MediaView(TemplateView):
+class MediaView(ListView):
     '''
         A bare list of media added with their states.
     '''
 
     template_name = 'sync/media.html'
+    context_object_name = 'media'
+    paginate_by = settings.MEDIA_PER_PAGE
+    messages = {
+        'filter': _('Viewing media for source: <strong>{name}</strong>'),
+    }
+
+    def __init__(self, *args, **kwargs):
+        self.filter_source = None
+        super().__init__(*args, **kwargs)
 
     def dispatch(self, request, *args, **kwargs):
+        filter_by = request.GET.get('filter', '')
+        if filter_by:
+            try:
+                self.filter_source = Source.objects.get(pk=filter_by)
+            except Source.DoesNotExist:
+                self.filter_source = None
         return super().dispatch(request, *args, **kwargs)
 
+    def get_queryset(self):
+        if self.filter_source:
+            return Media.objects.filter(source=self.filter_source).order_by('-created')
+        else:
+            return Media.objects.all().order_by('-created')
+
+    def get_context_data(self, *args, **kwargs):
+        data = super().get_context_data(*args, **kwargs)
+        data['message'] = ''
+        data['source'] = None
+        if self.filter_source:
+            message = str(self.messages.get('filter', ''))
+            print(message)
+            data['message'] = message.format(name=self.filter_source.name)
+            data['source'] = self.filter_source
+        print(data)
+        return data
+
+
+class MediaThumbView(DetailView):
+    '''
+        Shows a media thumbnail. Whitenose doesn't support post-start media image
+        serving and the images here are pretty small, just serve them manually. This
+        isn't fast, but it's not likely to be a serious bottleneck.
+    '''
+
+    model = Media
+
+    def get(self, request, *args, **kwargs):
+        media = self.get_object()
+        if media.thumb:
+            thumb = open(media.thumb.path, 'rb').read()
+            content_type = 'image/jpeg' 
+        else:
+            thumb = b64decode('R0lGODlhAQABAIABAP///wAAACH5BAEKAAEALAA'
+                              'AAAABAAEAAAICTAEAOw==')
+            content_type = 'image/gif'
+        return HttpResponse(thumb, content_type=content_type)
+
+
+class MediaItemView(DetailView):
+
+    template_name = 'sync/media-item.html'
+    model = Media
+
 
 class TasksView(TemplateView):
     '''

app/sync/youtube.py

@@ -15,22 +15,20 @@ class YouTubeError(youtube_dl.utils.DownloadError):
     '''
         Generic wrapped error for all errors that could be raised by youtube-dl.
     '''
-
     pass
 
 
-def extract_info(url):
+def get_media_info(url):
     '''
         Extracts information from a YouTube URL and returns it as a dict. For a channel
         or playlist this returns a dict of all the videos on the channel or playlist
         as well as associated metadata.
     '''
+
     opts = _defaults.update({
         'skip_download': True,
         'forcejson': True,
         'simulate': True,
-        'extract_flat': 'in_playlist',
-        'playlist_items': 1,
     })
     response = {}
     with youtube_dl.YoutubeDL(opts) as y:

app/tubesync/settings.py

@@ -96,7 +96,7 @@ USE_TZ = True
 
 STATIC_URL = '/static/'
 STATIC_ROOT = BASE_DIR / 'static'
-MEDIA_URL = '/media/'
+#MEDIA_URL = '/media/'
 MEDIA_ROOT = BASE_DIR / 'media'
 SYNC_VIDEO_ROOT = BASE_DIR / 'downloads' / 'video'
 SYNC_AUDIO_ROOT = BASE_DIR / 'downloads' / 'audio'
@@ -121,10 +121,17 @@ BACKGROUND_TASK_PRIORITY_ORDERING = 'DESC'  # Process high priority tasks first
 
 
 SOURCES_PER_PAGE = 25
+MEDIA_PER_PAGE = 25
+
+
+INDEX_SOURCE_EVERY = 60  # Seconds between indexing sources, 21600 = every 6 hours
+
+
+MAX_MEDIA_THUMBNAIL_SIZE = (320, 240)       # Max size in pixels for media thumbnails
 
 
 YOUTUBE_DEFAULTS = {
-    'age_limit': 99,
+    'age_limit': 99,        # Age in years to spoof the client as
 }