Merge branch 'main' into patch-17

.github/workflows/ci.yaml

@@ -132,6 +132,8 @@ jobs:
             pip install --system --strict --requirements requirements.txt
       - name: Set up Django environment
         run: |
+          mkdir -v -p ~/.config/TubeSync/config
+          sudo ln -v -s -f -T ~/.config/TubeSync/config /config
           cp -v -p tubesync/tubesync/local_settings.py.example tubesync/tubesync/local_settings.py
           cp -v -a -t "${Python3_ROOT_DIR}"/lib/python3.*/site-packages/background_task/ patches/background_task/*
           cp -v -a -t "${Python3_ROOT_DIR}"/lib/python3.*/site-packages/yt_dlp/ patches/yt_dlp/*

Dockerfile

@@ -25,6 +25,7 @@ ARG TARGETARCH
 
 ENV DEBIAN_FRONTEND="noninteractive" \
     APT_KEEP_ARCHIVES=1 \
+    EDITOR="editor" \
     HOME="/root" \
     LANGUAGE="en_US.UTF-8" \
     LANG="en_US.UTF-8" \
@@ -321,6 +322,8 @@ RUN --mount=type=cache,id=apt-lib-cache-${TARGETARCH},sharing=private,target=/va
   apt-get -y autoclean && \
   rm -v -f /var/cache/debconf/*.dat-old
 
+# The preference for openresty over nginx,
+# is for the newer version.
 FROM tubesync-openresty AS tubesync
 
 ARG S6_VERSION
@@ -342,20 +345,31 @@ RUN --mount=type=cache,id=apt-lib-cache-${TARGETARCH},sharing=private,target=/va
   # Install dependencies we keep
   # Install required distro packages
   apt-get -y --no-install-recommends install \
-  libjpeg62-turbo \
   libmariadb3 \
-  libpq5 \
+  libonig5 \
-  libwebp7 \
   pkgconf \
   python3 \
   python3-libsass \
   python3-pip-whl \
   python3-socks \
   curl \
+  indent \
   less \
+  lua-lpeg \
+  tre-agrep \
+  vis \
+  xxd \
   && \
   # Link to the current python3 version
   ln -v -s -f -T "$(find /usr/local/lib -name 'python3.[0-9]*' -type d -printf '%P\n' | sort -r -V | head -n 1)" /usr/local/lib/python3 && \
+  # Configure the editor alternatives
+  touch /usr/local/bin/babi /bin/nano /usr/bin/vim.tiny && \
+  update-alternatives --install /usr/bin/editor editor /usr/local/bin/babi 50 && \
+  update-alternatives --install /usr/local/bin/nano nano /bin/nano 10 && \
+  update-alternatives --install /usr/local/bin/nano nano /usr/local/bin/babi 20 && \
+  update-alternatives --install /usr/local/bin/vim vim /usr/bin/vim.tiny 15 && \
+  update-alternatives --install /usr/local/bin/vim vim /usr/bin/vis 35 && \
+  rm -v /usr/local/bin/babi /bin/nano /usr/bin/vim.tiny && \
   # Create a 'app' user which the application will run as
   groupadd app && \
   useradd -M -d /app -s /bin/false -g app app && \
@@ -407,6 +421,7 @@ RUN --mount=type=tmpfs,target=/cache \
   g++ \
   gcc \
   libjpeg-dev \
+  libonig-dev \
   libpq-dev \
   libwebp-dev \
   make \
@@ -450,6 +465,7 @@ RUN --mount=type=tmpfs,target=/cache \
   g++ \
   gcc \
   libjpeg-dev \
+  libonig-dev \
   libpq-dev \
   libwebp-dev \
   make \
@@ -459,8 +475,21 @@ RUN --mount=type=tmpfs,target=/cache \
   && \
   apt-get -y autopurge && \
   apt-get -y autoclean && \
+  LD_LIBRARY_PATH=/usr/local/lib/python3/dist-packages/pillow.libs:/usr/local/lib/python3/dist-packages/psycopg_binary.libs \
+    find /usr/local/lib/python3/dist-packages/ \
+      -name '*.so*' -print \
+      -exec du -h '{}' ';' \
+      -exec ldd '{}' ';' \
+    >| /cache/python-shared-objects 2>&1 && \
   rm -v -f /var/cache/debconf/*.dat-old && \
-  rm -v -rf /tmp/*
+  rm -v -rf /tmp/* ; \
+  if grep >/dev/null -Fe ' => not found' /cache/python-shared-objects ; \
+  then \
+      cat -v /cache/python-shared-objects ; \
+      printf -- 1>&2 '%s\n' \
+        ERROR: '    An unresolved shared object was found.' ; \
+      exit 1 ; \
+  fi
 
 # Copy root
 COPY config/root /
@@ -480,13 +509,14 @@ COPY tubesync/tubesync/local_settings.py.container /app/tubesync/local_settings.
 # Build app
 RUN set -x && \
   # Make absolutely sure we didn't accidentally bundle a SQLite dev database
-  rm -rf /app/db.sqlite3 && \
+  test '!' -e /app/db.sqlite3 && \
   # Run any required app commands
   /usr/bin/python3 -B /app/manage.py compilescss && \
   /usr/bin/python3 -B /app/manage.py collectstatic --no-input --link && \
+  rm -rf /config /downloads /run/app && \
   # Create config, downloads and run dirs
   mkdir -v -p /run/app && \
-  mkdir -v -p /config/media && \
+  mkdir -v -p /config/media /config/tasks && \
   mkdir -v -p /config/cache/pycache && \
   mkdir -v -p /downloads/audio && \
   mkdir -v -p /downloads/video && \

Pipfile

@@ -8,6 +8,7 @@ autopep8 = "*"
 
 [packages]
 django = "~=5.2.1"
+django-huey = "*"
 django-sass-processor = {extras = ["management-command"], version = "*"}
 pillow = "*"
 whitenoise = "*"
@@ -25,3 +26,4 @@ emoji = "*"
 brotli = "*"
 html5lib = "*"
 bgutil-ytdlp-pot-provider = "~=1.0"
+babi = "*"

config/root/etc/s6-overlay/s6-rc.d/background-task-workers/contents.d/tubesync-db-worker

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/background-task-workers/contents.d/tubesync-fs-worker

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/background-task-workers/contents.d/tubesync-network-worker

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/background-task-workers/type

@@ -0,0 +1 @@
+bundle

config/root/etc/s6-overlay/s6-rc.d/huey-consumers/contents.d/huey-database

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/huey-consumers/contents.d/huey-filesystem

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/huey-consumers/contents.d/huey-net-limited

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/huey-consumers/contents.d/huey-network

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/huey-consumers/type

@@ -0,0 +1 @@
+bundle

config/root/etc/s6-overlay/s6-rc.d/huey-database/dependencies.d/base

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/huey-database/dependencies.d/gunicorn

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/huey-database/down-signal

@@ -0,0 +1 @@
+SIGINT

config/root/etc/s6-overlay/s6-rc.d/huey-database/run

@@ -0,0 +1,5 @@
+#!/command/with-contenv bash
+
+exec nice -n "${TUBESYNC_NICE:-1}" s6-setuidgid app \
+    /usr/bin/python3 /app/manage.py djangohuey \
+    --queue database

config/root/etc/s6-overlay/s6-rc.d/huey-database/type

@@ -0,0 +1 @@
+longrun

config/root/etc/s6-overlay/s6-rc.d/huey-filesystem/dependencies.d/base

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/huey-filesystem/dependencies.d/gunicorn

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/huey-filesystem/down-signal

@@ -0,0 +1 @@
+SIGINT

config/root/etc/s6-overlay/s6-rc.d/huey-filesystem/run

@@ -0,0 +1,5 @@
+#!/command/with-contenv bash
+
+exec nice -n "${TUBESYNC_NICE:-1}" s6-setuidgid app \
+    /usr/bin/python3 /app/manage.py djangohuey \
+    --queue filesystem

config/root/etc/s6-overlay/s6-rc.d/huey-filesystem/type

@@ -0,0 +1 @@
+longrun

config/root/etc/s6-overlay/s6-rc.d/huey-net-limited/dependencies.d/base

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/huey-net-limited/dependencies.d/gunicorn

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/huey-net-limited/down-signal

@@ -0,0 +1 @@
+SIGINT

config/root/etc/s6-overlay/s6-rc.d/huey-net-limited/run

@@ -0,0 +1,5 @@
+#!/command/with-contenv bash
+
+exec nice -n "${TUBESYNC_NICE:-1}" s6-setuidgid app \
+    /usr/bin/python3 /app/manage.py djangohuey \
+    --queue limited

config/root/etc/s6-overlay/s6-rc.d/huey-net-limited/type

@@ -0,0 +1 @@
+longrun

config/root/etc/s6-overlay/s6-rc.d/huey-network/dependencies.d/base

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/huey-network/dependencies.d/gunicorn

@@ -0,0 +1 @@
+

config/root/etc/s6-overlay/s6-rc.d/huey-network/down-signal

@@ -0,0 +1 @@
+SIGINT

config/root/etc/s6-overlay/s6-rc.d/huey-network/run

@@ -0,0 +1,5 @@
+#!/command/with-contenv bash
+
+exec nice -n "${TUBESYNC_NICE:-1}" s6-setuidgid app \
+    /usr/bin/python3 /app/manage.py djangohuey \
+    --queue network

config/root/etc/s6-overlay/s6-rc.d/huey-network/type

@@ -0,0 +1 @@
+longrun

config/root/etc/s6-overlay/s6-rc.d/tubesync-init/run

@@ -4,6 +4,21 @@
 groupmod -o -g "${PGID:=911}" app
 usermod -o -u "${PUID:=911}" app
 
+# Ensure /config directories exist
+mkdir -v -p /config/{cache,media,tasks,tubesync}
+
+# Copy local_settings.py for the user
+if [ -f /config/tubesync/local_settings.py ]
+then
+    # from the image for an example or comparison
+    cp -v -p /app/tubesync/local_settings.py \
+        /config/tubesync/local_settings.py.image
+
+    # to the image to apply the user's settings
+    cp -v -p /config/tubesync/local_settings.py \
+        /app/tubesync/local_settings.py
+fi
+
 # Reset permissions
 chown -R app:app /run/app
 chmod -R 0700 /run/app

config/root/etc/s6-overlay/s6-rc.d/user/contents.d/huey-consumers

@@ -0,0 +1 @@
+

tubesync/common/errors.py

@@ -42,3 +42,8 @@ class DatabaseConnectionError(Exception):
         Raised when parsing or initially connecting to a database.
     '''
     pass
+
+
+class BgTaskWorkerError(Exception):
+    # Raised when the worker process is not in a normal working state.
+    pass

tubesync/common/huey.py

@@ -0,0 +1,37 @@
+
+
+def sqlite_tasks(key, /, prefix=None):
+    name_fmt = 'huey_{}'
+    if prefix is None:
+        prefix = ''
+    if prefix:
+        name_fmt = f'huey_{prefix}_' + '{}'
+    name = name_fmt.format(key)
+    return dict(
+        huey_class='huey.SqliteHuey',
+        name=name,
+        immediate=False,
+        results=True,
+        store_none=False,
+        utc=True,
+        compression=True,
+        connection=dict(
+            filename=f'/config/tasks/{name}.db',
+            fsync=True,
+            strict_fifo=True,
+        ),
+        consumer=dict(
+            workers=1,
+            worker_type='process',
+            max_delay=20.0,
+            flush_locks=True,
+            scheduler_interval=10,
+            simple_log=False,
+            # verbose has three positions:
+            # DEBUG: True
+            # INFO: None
+            # WARNING: False
+            verbose=False,
+        ),
+    )
+

tubesync/common/utils.py

@@ -7,9 +7,19 @@ import pstats
 import string
 import time
 from django.core.paginator import Paginator
+from functools import partial
+from operator import attrgetter, itemgetter
+from pathlib import Path
 from urllib.parse import urlunsplit, urlencode, urlparse
 from .errors import DatabaseConnectionError
 
+def directory_and_stem(arg_path, /, all_suffixes=False):
+    filepath = Path(arg_path)
+    stem = Path(filepath.stem)
+    while all_suffixes and stem.suffixes and '' != stem.suffix:
+        stem = Path(stem.stem)
+    return (filepath.parent, str(stem),)
+
 
 def getenv(key, default=None, /, *, integer=False, string=True):
     '''
@@ -46,6 +56,51 @@ def getenv(key, default=None, /, *, integer=False, string=True):
     return r
 
 
+def glob_quote(filestr, /):
+    _glob_specials = {
+        '?': '[?]',
+        '*': '[*]',
+        '[': '[[]',
+        ']': '[]]', # probably not needed, but it won't hurt
+    }
+
+    if not isinstance(filestr, str):
+        raise TypeError(f'expected a str, got "{type(filestr)}"')
+
+    return filestr.translate(str.maketrans(_glob_specials))
+
+
+def list_of_dictionaries(arg_list, /, arg_function=lambda x: x):
+    assert callable(arg_function)
+    if isinstance(arg_list, list):
+        _map_func = partial(lambda f, d: f(d) if isinstance(d, dict) else d, arg_function)
+        return (True, list(map(_map_func, arg_list)),)
+    return (False, arg_list,)
+
+
+def mkdir_p(arg_path, /, *, mode=0o777):
+    '''
+        Reminder: mode only affects the last directory
+    '''
+    dirpath = Path(arg_path)
+    return dirpath.mkdir(mode=mode, parents=True, exist_ok=True)
+
+
+def multi_key_sort(iterable, specs, /, use_reversed=False, *, item=False, attr=False, key_func=None):
+    result = list(iterable)
+    if key_func is None:
+        # itemgetter is the default
+        if item or not (item or attr):
+            key_func = itemgetter
+        elif attr:
+            key_func = attrgetter
+    for key, reverse in reversed(specs):
+        result.sort(key=key_func(key), reverse=reverse)
+    if use_reversed:
+        return list(reversed(result))
+    return result
+
+
 def parse_database_connection_string(database_connection_string):
     '''
         Parses a connection string in a URL style format, such as:
@@ -167,6 +222,15 @@ def clean_emoji(s):
     return emoji.replace_emoji(s)
 
 
+def seconds_to_timestr(seconds):
+    seconds = seconds % (24 * 3600)
+    hour = seconds // 3600
+    seconds %= 3600
+    minutes = seconds // 60
+    seconds %= 60
+    return '{:02d}:{:02d}:{:02d}'.format(hour, minutes, seconds)
+
+
 def time_func(func):
     def wrapper(*args, **kwargs):
         start = time.perf_counter()

tubesync/restart_services.sh

@@ -1,26 +1,51 @@
 #!/usr/bin/env sh
 
-dir='/run/service'
+_dir='/run/service'
 svc_path() (
-    cd "${dir}"
+    cd "${_dir}" &&
     realpath -e -s "$@"
 )
 
+_bundles="$(
+    find '/etc/s6-overlay/s6-rc.d' -mindepth 2 -maxdepth 2 \
+        -name 'type' \
+        -execdir grep -F -q -e bundle '{}' ';' \
+        -printf '%P\n' | \
+        sed -e 's,/type$,,' ;
+)"
+is_a_bundle() {
+    local bundle
+    for bundle in ${_bundles}
+    do
+        if [ "$1" = "${bundle}" ]
+        then
+            return 0
+        fi
+    done
+    return 1
+}
+
 if [ 0 -eq $# ]
 then
-    set -- \
+    set -- $(/command/s6-rc list user | grep -v -e '-init$')
-        $( cd "${dir}" && svc_path tubesync*-worker ) \
-        "$( svc_path gunicorn )" \
-        "$( svc_path nginx )"
 fi
 
-for service in $( svc_path "$@" )
+for arg in "$@"
 do
-    printf -- 'Restarting %-28s' "${service#${dir}/}..."
+    _svcs="${arg}"
+    if is_a_bundle "${arg}"
+    then
+        _svcs="$(/command/s6-rc list "${arg}" | grep -v -e '-init$')"
+    fi
+    for service in $(svc_path ${_svcs})
+    do
+        printf -- 'Restarting %-28s' "${service#${_dir}/}..."
         _began="$( date '+%s' )"
         /command/s6-svc -wr -r "${service}"
         _ended="$( date '+%s' )"
         printf -- '\tcompleted (in %2.1d seconds).\n' \
             "$( expr "${_ended}" - "${_began}" )"
     done
-unset -v _began _ended service
+done
+unset -v _began _ended _svcs arg service
+unset -v _bundles _dir

tubesync/sync/choices.py

@@ -165,6 +165,7 @@ class TaskQueue(models.TextChoices):
     DB = 'database', _('Database')
     FS = 'filesystem', _('Filesystem')
     NET = 'network', _('Networking')
+    LIMIT = 'limited', _('Limited Networking')
 
 
 class WeekDay(models.IntegerChoices):

tubesync/sync/matching.py

@@ -6,7 +6,7 @@
 
 
 from .choices import Val, Fallback
-from .utils import multi_key_sort
+from common.utils import multi_key_sort
 from django.conf import settings
 
 

tubesync/sync/migrations/0035_alter_metadata_unique_together_metadata_source_and_more.py

@@ -0,0 +1,58 @@
+# Generated by Django 5.2.1 on 2025-05-28 09:57
+
+import django.db.models.deletion
+import django.utils.timezone
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ('sync', '0034_source_target_schedule_and_more'),
+    ]
+
+    operations = [
+        migrations.AlterUniqueTogether(
+            name='metadata',
+            unique_together={('media', 'site', 'key')},
+        ),
+        migrations.AddField(
+            model_name='metadata',
+            name='source',
+            field=models.ForeignKey(
+                blank=True,
+                help_text='Source from which the video was retrieved',
+                null=True,
+                on_delete=django.db.models.deletion.CASCADE,
+                related_name='videos',
+                related_query_name='video',
+                to='sync.source',
+            ),
+        ),
+        migrations.AlterField(
+            model_name='metadata',
+            name='media',
+            field=models.OneToOneField(
+                blank=True,
+                help_text='Media the metadata belongs to',
+                null=True,
+                on_delete=django.db.models.deletion.SET_NULL,
+                related_name='new_metadata',
+                to='sync.media',
+            ),
+        ),
+        migrations.AlterField(
+            model_name='metadata',
+            name='retrieved',
+            field=models.DateTimeField(
+                db_index=True,
+                default=django.utils.timezone.now,
+                help_text='Date and time the metadata was retrieved',
+                verbose_name='retrieved',
+            ),
+        ),
+        migrations.AlterUniqueTogether(
+            name='metadata',
+            unique_together={('media', 'site', 'key'), ('source', 'site', 'key')},
+        ),
+    ]
+

tubesync/sync/models/_private.py

@@ -1,4 +1,3 @@
-from pathlib import Path
 from ..choices import Val, YouTube_SourceType # noqa
 
 
@@ -11,11 +10,3 @@ def _nfo_element(nfo, label, text, /, *, attrs={}, tail='\n', char=' ', indent=2
     element.tail = tail + (char * indent)
     return element
 
-def directory_and_stem(arg_path, /, all_suffixes=False):
-    filepath = Path(arg_path)
-    stem = Path(filepath.stem)
-    while all_suffixes and stem.suffixes and '' != stem.suffix:
-        stem = Path(stem.stem)
-    stem = str(stem)
-    return (filepath.parent, stem,)
-

tubesync/sync/models/media.py

@@ -17,15 +17,15 @@ from common.logger import log
 from common.errors import NoFormatException
 from common.json import JSONEncoder
 from common.utils import (
-    clean_filename, clean_emoji,
+    clean_filename, clean_emoji, directory_and_stem,
+    glob_quote, mkdir_p, multi_key_sort, seconds_to_timestr,
 )
 from ..youtube import (
     get_media_info as get_youtube_media_info,
     download_media as download_youtube_media,
 )
 from ..utils import (
-    seconds_to_timestr, parse_media_format, filter_response,
+    filter_response, parse_media_format, write_text_file,
-    write_text_file, mkdir_p, glob_quote, multi_key_sort,
 )
 from ..matching import (
     get_best_combined_format,
@@ -38,7 +38,7 @@ from ..choices import (
 from ._migrations import (
     media_file_storage, get_media_thumb_path, get_media_file_path,
 )
-from ._private import _srctype_dict, _nfo_element, directory_and_stem
+from ._private import _srctype_dict, _nfo_element
 from .media__tasks import (
     download_checklist, download_finished, wait_for_premiere,
 )
@@ -566,7 +566,11 @@ class Media(models.Model):
 
     @property
     def has_metadata(self):
-        return self.metadata is not None
+        result = self.metadata is not None
+        if not result:
+            return False
+        value = self.get_metadata_first_value(('id', 'display_id', 'channel_id', 'uploader_id',))
+        return value is not None
 
 
     def metadata_clear(self, /, *, save=False):
@@ -600,8 +604,10 @@ class Media(models.Model):
             arg_dict=data,
         )
         md_model = self._meta.fields_map.get('new_metadata').related_model
-        md, created = md_model.objects.get_or_create(
+        md, created = md_model.objects.filter(
-            media_id=self.pk,
+            source__isnull=True,
+        ).get_or_create(
+            media=self,
             site=site,
             key=self.key,
         )
@@ -697,8 +703,7 @@ class Media(models.Model):
         data = self.loaded_metadata
         metadata_seconds = data.get('epoch', None)
         if not metadata_seconds:
-            self.metadata = None
+            self.metadata_clear(save=True)
-            self.save(update_fields={'metadata'})
             return False
 
         now = timezone.now()

tubesync/sync/models/metadata.py

@@ -3,8 +3,9 @@ from common.json import JSONEncoder
 from common.timestamp import timestamp_to_datetime
 from common.utils import django_queryset_generator as qs_gen
 from django import db
+from django.utils import timezone
 from django.utils.translation import gettext_lazy as _
-from .media import Media
+from .media import Media, Source
 
 
 class Metadata(db.models.Model):
@@ -17,6 +18,7 @@ class Metadata(db.models.Model):
         verbose_name_plural = _('Metadata about Media')
         unique_together = (
             ('media', 'site', 'key'),
+            ('source', 'site', 'key', ),
         )
         get_latest_by = ["-retrieved", "-created"]
 
@@ -27,12 +29,22 @@ class Metadata(db.models.Model):
         default=uuid.uuid4,
         help_text=_('UUID of the metadata'),
     )
+    source = db.models.ForeignKey(
+        Source,
+        on_delete=db.models.CASCADE,
+        related_name="videos",
+        related_query_name="video",
+        help_text=_('Source from which the video was retrieved'),
+        blank=True,
+        null=True,
+    )
     media = db.models.OneToOneField(
         Media,
         # on_delete=models.DO_NOTHING,
         on_delete=db.models.SET_NULL,
         related_name='new_metadata',
         help_text=_('Media the metadata belongs to'),
+        blank=True,
         null=True,
         parent_link=False,
     )
@@ -62,8 +74,8 @@ class Metadata(db.models.Model):
     )
     retrieved = db.models.DateTimeField(
         _('retrieved'),
-        auto_now_add=True,
         db_index=True,
+        default=timezone.now,
         help_text=_('Date and time the metadata was retrieved'),
     )
     uploaded = db.models.DateTimeField(

tubesync/sync/models/source.py

@@ -404,7 +404,7 @@ class Source(db.models.Model):
             )
         elif Val(IndexSchedule.EVERY_7_DAYS) > self.index_schedule:
             self.target_schedule = advance_hour(
-                when.replace(hour=1+when.hour),
+                when + timezone.timedelta(hours=1),
                 self.target_schedule.hour,
             )
 

tubesync/sync/signals.py

@@ -4,12 +4,13 @@ from tempfile import TemporaryDirectory
 from django.conf import settings
 from django.db import IntegrityError
 from django.db.models.signals import pre_save, post_save, pre_delete, post_delete
-from django.db.transaction import on_commit
+from django.db.transaction import atomic, on_commit
 from django.dispatch import receiver
 from django.utils.translation import gettext_lazy as _
 from background_task.signals import task_failed
 from background_task.models import Task
 from common.logger import log
+from common.utils import glob_quote, mkdir_p
 from .models import Source, Media, Metadata
 from .tasks import (delete_task_by_source, delete_task_by_media, index_source_task,
                     download_media_thumbnail, download_media_metadata,
@@ -17,7 +18,7 @@ from .tasks import (delete_task_by_source, delete_task_by_media, index_source_ta
                     download_media, download_source_images,
                     delete_all_media_for_source, save_all_media_for_source,
                     rename_media, get_media_metadata_task, get_media_download_task)
-from .utils import delete_file, glob_quote, mkdir_p
+from .utils import delete_file
 from .filtering import filter_media
 from .choices import Val, YouTube_SourceType
 
@@ -429,17 +430,40 @@ def media_post_delete(sender, instance, **kwargs):
         # Re-use the old metadata if it exists
         instance_qs = Metadata.objects.filter(
             media__isnull=True,
+            source__isnull=True,
             site=old_metadata.get(site_field) or 'Youtube',
             key=skipped_media.key,
         )
         try:
-            instance_qs.update(media=skipped_media)
+            if instance_qs.count():
+                with atomic(durable=False):
+                    # clear the link to a media instance
+                    Metadata.objects.filter(media=skipped_media).update(media=None)
+                    # choose the oldest metadata for our key
+                    md = instance_qs.filter(
+                        key=skipped_media.key,
+                    ).order_by(
+                        'key',
+                        'created',
+                    ).first()
+                    # set the link to a media instance only on our selected metadata
+                    log.info(f'Reusing old metadata for "{skipped_media.key}": {skipped_media.name}')
+                    instance_qs.filter(uuid=md.uuid).update(media=skipped_media)
+                    # delete any metadata that we are no longer using
+                    instance_qs.exclude(uuid=md.uuid).delete()
+                    
         except IntegrityError:
-            # Delete the new metadata
+            # this probably won't happen, but try it without a transaction
-            Metadata.objects.filter(media=skipped_media).delete()
             try:
-                instance_qs.update(media=skipped_media)
+                # clear the link to a media instance
-            except IntegrityError:
+                Metadata.objects.filter(media=skipped_media).update(media=None)
-                # Delete the old metadata if it still failed
+                # keep one metadata
+                md = instance_qs.order_by('created').first()
+                instance_qs.filter(uuid=md.uuid).update(media=skipped_media)
+            except IntegrityError as e:
+                log.exception(f'media_post_delete: could not update selected metadata: {e}')
+            finally:
+                log.debug(f'Deleting metadata for "{skipped_media.key}": {skipped_media.pk}')
+                # delete the old metadata
                 instance_qs.delete()
 

tubesync/sync/tasks.py

@@ -10,6 +10,7 @@ import random
 import requests
 import time
 import uuid
+from collections import deque as queue
 from io import BytesIO
 from hashlib import sha1
 from pathlib import Path
@@ -19,7 +20,7 @@ from django import db
 from django.conf import settings
 from django.core.files.base import ContentFile
 from django.core.files.uploadedfile import SimpleUploadedFile
-from django.db import DatabaseError, IntegrityError
+from django.db import DatabaseError
 from django.db.transaction import atomic
 from django.utils import timezone
 from django.utils.translation import gettext_lazy as _
@@ -27,15 +28,15 @@ from background_task import background
 from background_task.exceptions import InvalidTaskError
 from background_task.models import Task, CompletedTask
 from common.logger import log
-from common.errors import ( NoFormatException, NoMediaException,
+from common.errors import ( BgTaskWorkerError, DownloadFailedException,
-                            NoThumbnailException,
+                            NoFormatException, NoMediaException,
-                            DownloadFailedException, )
+                            NoThumbnailException, )
 from common.utils import (  django_queryset_generator as qs_gen,
-                            remove_enclosed, )
+                            remove_enclosed, seconds_to_timestr, )
 from .choices import Val, TaskQueue
-from .models import Source, Media, MediaServer
+from .models import Source, Media, MediaServer, Metadata
 from .utils import ( get_remote_image, resize_image_to_height,
-                    write_text_file, filter_response, seconds_to_timestr, )
+                    write_text_file, filter_response, )
 from .youtube import YouTubeError
 
 db_vendor = db.connection.vendor
@@ -56,6 +57,7 @@ def map_task_to_instance(task):
         because UUID's are incompatible with background_task's "creator" feature.
     '''
     TASK_MAP = {
+        'sync.tasks.migrate_to_metadata': Media,
         'sync.tasks.index_source_task': Source,
         'sync.tasks.check_source_directory_exists': Source,
         'sync.tasks.download_media_thumbnail': Media,
@@ -302,6 +304,70 @@ def cleanup_removed_media(source, video_keys):
     schedule_media_servers_update()
 
 
+def save_db_batch(qs, objs, fields, /):
+    assert hasattr(qs, 'bulk_update')
+    assert callable(qs.bulk_update)
+    assert hasattr(objs, '__len__')
+    assert callable(objs.__len__)
+    assert isinstance(fields, (tuple, list, set, frozenset))
+
+    num_updated = 0
+    num_objs = len(objs)
+    with atomic(durable=False):
+        num_updated = qs.bulk_update(objs=objs, fields=fields)
+    if num_objs == num_updated:
+        # this covers at least: list, set, deque
+        if hasattr(objs, 'clear') and callable(objs.clear):
+            objs.clear()
+    return num_updated
+
+
+@background(schedule=dict(priority=20, run_at=60), queue=Val(TaskQueue.DB), remove_existing_tasks=True)
+def migrate_to_metadata(media_id):
+    try:
+        media = Media.objects.get(pk=media_id)
+    except Media.DoesNotExist as e:
+        # Task triggered but the media no longer exists, do nothing
+        log.error(f'Task migrate_to_metadata(pk={media_id}) called but no '
+                  f'media exists with ID: {media_id}')
+        raise InvalidTaskError(_('no such media')) from e
+
+    try:
+        data = Metadata.objects.get(
+            media__isnull=True,
+            source=media.source,
+            key=media.key,
+        )
+    except Metadata.DoesNotExist as e:
+        raise InvalidTaskError(_('no indexed data to migrate to metadata')) from e
+
+    video = data.value
+    fields = lambda f, m: m.get_metadata_field(f)
+    timestamp = video.get(fields('timestamp', media), None)
+    for key in ('epoch', 'availability', 'extractor_key',):
+        field = fields(key, media)
+        value = video.get(field)
+        existing_value = media.get_metadata_first_value(key)
+        if value is None:
+            if 'epoch' == key:
+                value = timestamp
+            elif 'extractor_key' == key:
+                value = data.site
+        if value is not None:
+            if existing_value and ('epoch' == key or value == existing_value):
+                continue
+            media.save_to_metadata(field, value)
+
+
+@background(schedule=dict(priority=0, run_at=0), queue=Val(TaskQueue.NET), remove_existing_tasks=False)
+def wait_for_database_queue():
+    worker_down_path = Path('/run/service/tubesync-db-worker/down')
+    while Task.objects.unlocked(timezone.now()).filter(queue=Val(TaskQueue.DB)).count() > 0:
+        time.sleep(5)
+        if worker_down_path.exists() and worker_down_path.is_file():
+            raise BgTaskWorkerError(_('queue worker stopped'))
+
+
 @background(schedule=dict(priority=20, run_at=30), queue=Val(TaskQueue.NET), remove_existing_tasks=True)
 def index_source_task(source_id):
     '''
@@ -322,13 +388,12 @@ def index_source_task(source_id):
     # update the target schedule column
     source.task_run_at_dt
     # Reset any errors
-    # TODO: determine if this affects anything
     source.has_failed = False
-    save_model(source)
     # Index the source
     videos = source.index_media()
     if not videos:
-        # TODO: Record this error in source.has_failed ?
+        source.has_failed = True
+        save_model(source)
         raise NoMediaException(f'Source "{source}" (ID: {source_id}) returned no '
                                f'media to index, is the source key valid? Check the '
                                f'source configuration is correct and that the source '
@@ -336,8 +401,31 @@ def index_source_task(source_id):
     # Got some media, update the last crawl timestamp
     source.last_crawl = timezone.now()
     save_model(source)
+    wait_for_database_queue(
+        priority=19, # the indexing task uses 20
+        verbose_name=_('Waiting for database tasks to complete'),
+    )
+    wait_for_database_queue(
+        priority=29, # the checking task uses 30
+        queue=Val(TaskQueue.FS),
+        verbose_name=_('Delaying checking all media for database tasks'),
+    )
+    delete_task_by_source('sync.tasks.save_all_media_for_source', source.pk)
     num_videos = len(videos)
     log.info(f'Found {num_videos} media items for source: {source}')
+    tvn_format = '{:,}' + f'/{num_videos:,}'
+    db_batch_data = queue(list(), maxlen=50)
+    db_fields_data = frozenset((
+        'retrieved',
+        'site',
+        'value',
+    ))
+    db_batch_media = queue(list(), maxlen=10)
+    db_fields_media = frozenset((
+        'duration',
+        'published',
+        'title',
+    ))
     fields = lambda f, m: m.get_metadata_field(f)
     task = get_source_index_task(source_id)
     if task:
@@ -346,7 +434,6 @@ def index_source_task(source_id):
             valid='0123456789/,',
             end=task.verbose_name.find('Index'),
         )
-    tvn_format = '{:,}' + f'/{num_videos:,}'
     vn = 0
     video_keys = set()
     while len(videos) > 0:
@@ -358,15 +445,17 @@ def index_source_task(source_id):
             # Video has no unique key (ID), it can't be indexed
             continue
         video_keys.add(key)
+        if len(db_batch_data) == db_batch_data.maxlen:
+            save_db_batch(Metadata.objects, db_batch_data, db_fields_data)
+        if len(db_batch_media) == db_batch_media.maxlen:
+            save_db_batch(Media.objects, db_batch_media, db_fields_media)
         update_task_status(task, tvn_format.format(vn))
-        # media, new_media = Media.objects.get_or_create(key=key, source=source)
+        media_defaults = dict()
-        try:
+        # create a dummy instance to use its functions
-            media = Media.objects.get(key=key, source=source)
+        media = Media(source=source, key=key)
-        except Media.DoesNotExist:
+        media_defaults['duration'] = float(video.get(fields('duration', media), None) or 0) or None
-            media = Media(key=key)
+        media_defaults['title'] = str(video.get(fields('title', media), ''))[:200]
-        media.source = source
+        site = video.get(fields('ie_key', media), None)
-        media.duration = float(video.get(fields('duration', media), None) or 0) or None
-        media.title = str(video.get(fields('title', media), ''))[:200]
         timestamp = video.get(fields('timestamp', media), None)
         try:
             published_dt = media.ts_to_dt(timestamp)
@@ -374,26 +463,40 @@ def index_source_task(source_id):
             pass
         else:
             if published_dt:
-                media.published = published_dt
+                media_defaults['published'] = published_dt
-        try:
+        # Retrieve or create the actual media instance
-            media.save()
+        media, new_media = source.media_source.only(
-        except IntegrityError as e:
+            'uuid',
-            log.error(f'Index media failed: {source} / {media} with "{e}"')
+            'source',
-        else:
+            'key',
-            log.debug(f'Indexed media: {vn}: {source} / {media}')
+            *db_fields_media,
-            # log the new media instances
+        ).get_or_create(defaults=media_defaults, source=source, key=key)
-            new_media_instance = (
+        db_batch_media.append(media)
-                # new_media or
+        data, new_data = source.videos.defer('value').filter(
-                media.created and
+            media__isnull=True,
-                source.last_crawl and
+        ).get_or_create(source=source, key=key)
-                media.created >= source.last_crawl
+        if site:
+            data.site = site
+        data.retrieved = source.last_crawl
+        data.value = video
+        db_batch_data.append(data)
+        vn_fmt = _('Updating metadata from indexing results for: "{}": {}')
+        migrate_to_metadata(
+            str(media.pk),
+            verbose_name=vn_fmt.format(media.key, media.name),
         )
-            if new_media_instance:
+        if not new_media:
+            # update the existing media
+            for key, value in media_defaults.items():
+                setattr(media, key, value)
+            log.debug(f'Indexed media: {vn}: {source} / {media}')
+        else:
+            # log the new media instances
             log.info(f'Indexed new media: {source} / {media}')
             log.info(f'Scheduling tasks to download thumbnail for: {media.key}')
             thumbnail_fmt = 'https://i.ytimg.com/vi/{}/{}default.jpg'
             vn_fmt = _('Downloading {} thumbnail for: "{}": {}')
-                for prefix in ('hq', 'sd', 'maxres',):
+            for num, prefix in enumerate(('hq', 'sd', 'maxres',)):
                 thumbnail_url = thumbnail_fmt.format(
                     media.key,
                     prefix,
@@ -401,19 +504,34 @@ def index_source_task(source_id):
                 download_media_thumbnail(
                     str(media.pk),
                     thumbnail_url,
+                    schedule=dict(run_at=10+(300*num)),
                     verbose_name=vn_fmt.format(prefix, media.key, media.name),
                 )
             log.info(f'Scheduling task to download metadata for: {media.url}')
             verbose_name = _('Downloading metadata for: "{}": {}')
             download_media_metadata(
                 str(media.pk),
+                schedule=dict(priority=35),
                 verbose_name=verbose_name.format(media.key, media.name),
             )
     # Reset task.verbose_name to the saved value
     update_task_status(task, None)
+    # Update any remaining items in the batches
+    save_db_batch(Metadata.objects, db_batch_data, db_fields_data)
+    save_db_batch(Media.objects, db_batch_media, db_fields_media)
     # Cleanup of media no longer available from the source
     cleanup_removed_media(source, video_keys)
+    # Clear references to indexed data
     videos = video = None
+    db_batch_data.clear()
+    db_batch_media.clear()
+    # Trigger any signals that we skipped with batched updates
+    vn_fmt = _('Checking all media for "{}"')
+    save_all_media_for_source(
+        str(source.pk),
+        schedule=dict(run_at=60),
+        verbose_name=vn_fmt.format(source.name),
+    )
 
 
 @background(schedule=dict(priority=0, run_at=0), queue=Val(TaskQueue.FS))

tubesync/sync/templatetags/filters.py

@@ -1,5 +1,6 @@
 from django import template
 from django.template.defaultfilters import filesizeformat
+from math import ceil
 
 
 register = template.Library()
@@ -23,3 +24,64 @@ def sub(value, arg):
         except Exception:
             return ""
 
+
+@register.filter
+def timedelta(value, arg=None, /, *, fmt_2=None):
+    if hasattr(value, 'total_seconds') and callable(value.total_seconds):
+        seconds_total = value.total_seconds()
+    elif hasattr(value, 'seconds'):
+        seconds_total = value.seconds + (value.days * 24 * 60 * 60)
+    else:
+        seconds_total = value
+
+    dynamic_arg = False
+    if arg is None:
+        if seconds_total < 1.0:
+            return f'{seconds_total:.6f} seconds'
+        dynamic_arg = True
+        arg = '{hours2}:{minutes2}:{seconds2}'
+
+    if fmt_2 is None:
+        fmt_2 = '{:02d}'
+
+    seconds_total = ceil(seconds_total)
+    seconds = seconds_total % 60
+
+    minutes_total = seconds_total // 60
+    minutes = minutes_total % 60
+
+    hours_total = minutes_total // 60
+    hours = hours_total % 24
+
+    days_total = hours_total // 24
+    days = days_total % 365
+
+    years_total = days_total // 365
+    years = years_total
+
+    if dynamic_arg:
+        prefix_years = prefix_days = ''
+        if years_total > 0:
+            prefix_years = '{years_total} years, '
+        if prefix_years and days_total > 0:
+            prefix_days = '{days} days, '
+        elif days_total > 0:
+            prefix_days = '{days_total} days, '
+        arg = prefix_years + prefix_days + arg
+
+    return arg.format(**{
+        'seconds': seconds,
+        'seconds2': fmt_2.format(seconds),
+        'minutes': minutes,
+        'minutes2': fmt_2.format(minutes),
+        'hours': hours,
+        'hours2': fmt_2.format(hours),
+        'days': days,
+        'years': years,
+        'seconds_total': seconds_total,
+        'minutes_total': minutes_total,
+        'hours_total': hours_total,
+        'days_total': days_total,
+        'years_total': years_total,
+    })
+

tubesync/sync/tests.py

@@ -354,6 +354,7 @@ class FrontEndTestCase(TestCase):
         # Add some media
         test_minimal_metadata = '''
             {
+                "channel_id":"testkey",
                 "thumbnail":"https://example.com/thumb.jpg",
                 "formats": [{
                     "format_id":"251",

tubesync/sync/utils.py

@@ -2,11 +2,11 @@ import os
 import re
 import math
 from copy import deepcopy
-from operator import attrgetter, itemgetter
 from pathlib import Path
 from tempfile import NamedTemporaryFile
 import requests
 from PIL import Image
+from common.utils import list_of_dictionaries
 from django.conf import settings
 from urllib.parse import urlsplit, parse_qs
 from django.forms import ValidationError
@@ -95,20 +95,6 @@ def resize_image_to_height(image, width, height):
     return image
 
 
-def glob_quote(filestr):
-    _glob_specials = {
-        '?': '[?]',
-        '*': '[*]',
-        '[': '[[]',
-        ']': '[]]', # probably not needed, but it won't hurt
-    }
-
-    if not isinstance(filestr, str):
-        raise TypeError(f'filestr must be a str, got "{type(filestr)}"')
-
-    return filestr.translate(str.maketrans(_glob_specials))
-
-
 def file_is_editable(filepath):
     '''
         Checks that a file exists and the file is in an allowed predefined tuple of
@@ -130,14 +116,6 @@ def file_is_editable(filepath):
     return False
 
 
-def mkdir_p(arg_path, mode=0o777):
-    '''
-        Reminder: mode only affects the last directory
-    '''
-    dirpath = Path(arg_path)
-    return dirpath.mkdir(mode=mode, parents=True, exist_ok=True)
-
-
 def write_text_file(filepath, filedata):
     if not isinstance(filedata, str):
         raise TypeError(f'filedata must be a str, got "{type(filedata)}"')
@@ -162,30 +140,6 @@ def delete_file(filepath):
     return False
 
 
-def seconds_to_timestr(seconds):
-   seconds = seconds % (24 * 3600)
-   hour = seconds // 3600
-   seconds %= 3600
-   minutes = seconds // 60
-   seconds %= 60
-   return '{:02d}:{:02d}:{:02d}'.format(hour, minutes, seconds)
-
-
-def multi_key_sort(iterable, specs, /, use_reversed=False, *, item=False, attr=False, key_func=None):
-    result = list(iterable)
-    if key_func is None:
-        # itemgetter is the default
-        if item or not (item or attr):
-            key_func = itemgetter
-        elif attr:
-            key_func = attrgetter
-    for key, reverse in reversed(specs):
-        result.sort(key=key_func(key), reverse=reverse)
-    if use_reversed:
-        return list(reversed(result))
-    return result
-
-
 def normalize_codec(codec_str):
     result = str(codec_str).upper()
     parts = result.split('.')
@@ -201,17 +155,6 @@ def normalize_codec(codec_str):
     return result
 
 
-def list_of_dictionaries(arg_list, arg_function=lambda x: x):
-    assert callable(arg_function)
-    if isinstance(arg_list, list):
-        def _call_func_with_dict(arg_dict):
-            if isinstance(arg_dict, dict):
-                return arg_function(arg_dict)
-            return arg_dict
-        return (True, list(map(_call_func_with_dict, arg_list)),)
-    return (False, arg_list,)
-
-
 def _url_keys(arg_dict, filter_func):
     result = {}
     if isinstance(arg_dict, dict):

tubesync/sync/views.py

@@ -20,13 +20,13 @@ from django.utils._os import safe_join
 from django.utils import timezone
 from django.utils.translation import gettext_lazy as _
 from common.timestamp import timestamp_to_datetime
-from common.utils import append_uri_params
+from common.utils import append_uri_params, mkdir_p, multi_key_sort
 from background_task.models import Task, CompletedTask
 from .models import Source, Media, MediaServer
 from .forms import (ValidateSourceForm, ConfirmDeleteSourceForm, RedownloadMediaForm,
                     SkipMediaForm, EnableMediaForm, ResetTasksForm, ScheduleTaskForm,
                     ConfirmDeleteMediaServerForm, SourceForm)
-from .utils import validate_url, delete_file, multi_key_sort, mkdir_p
+from .utils import delete_file, validate_url
 from .tasks import (map_task_to_instance, get_error_message,
                     get_source_completed_tasks, get_media_download_task,
                     delete_task_by_media, index_source_task,

tubesync/sync/youtube.py

@@ -7,6 +7,7 @@
 import os
 
 from common.logger import log
+from common.utils import mkdir_p
 from copy import deepcopy
 from pathlib import Path
 from tempfile import TemporaryDirectory
@@ -15,7 +16,6 @@ from urllib.parse import urlsplit, parse_qs
 from django.conf import settings
 from .choices import Val, FileExtension
 from .hooks import postprocessor_hook, progress_hook
-from .utils import mkdir_p
 import yt_dlp
 import yt_dlp.patch.check_thumbnails
 import yt_dlp.patch.fatal_http_errors

tubesync/tubesync/settings.py

@@ -1,5 +1,6 @@
 from django import VERSION as DJANGO_VERSION
 from pathlib import Path
+from common.huey import sqlite_tasks
 from common.utils import getenv
 
 
@@ -24,6 +25,7 @@ INSTALLED_APPS = [
     'django.contrib.humanize',
     'sass_processor',
     'background_task',
+    'django_huey',
     'common',
     'sync',
 ]
@@ -47,6 +49,22 @@ ROOT_URLCONF = 'tubesync.urls'
 FORCE_SCRIPT_NAME = None
 
 
+DJANGO_HUEY = {
+    'default': 'network',
+    'queues': {
+        'database': sqlite_tasks('database'),
+        'filesystem': sqlite_tasks('filesystem'),
+        'limited': sqlite_tasks('limited', prefix='net'),
+        'network': sqlite_tasks('network'),
+    },
+}
+for django_huey_queue in DJANGO_HUEY['queues'].values():
+    connection = django_huey_queue.get('connection')
+    if connection:
+        filepath = Path('/.' + connection.get('filename') or '').resolve(strict=False)
+        filepath.parent.mkdir(exist_ok=True, parents=True)
+
+
 TEMPLATES = [
     {
         'BACKEND': 'django.template.backends.django.DjangoTemplates',