Merge pull request #406 from pacoccino/subtitles

Subtitles

.gitignore

@@ -1,3 +1,4 @@
+.DS_Store
 # Byte-compiled / optimized / DLL files
 __pycache__/
 *.py[cod]

tubesync/sync/migrations/0018_source_subtitles.py

@@ -0,0 +1,27 @@
+# Generated by pac
+
+from django.db import migrations, models
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('sync', '0017_alter_source_sponsorblock_categories'),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name='source',
+            name='write_subtitles',
+            field=models.BooleanField(default=False, help_text='Download video subtitles', verbose_name='write subtitles'),
+        ),
+        migrations.AddField(
+            model_name='source',
+            name='auto_subtitles',
+            field=models.BooleanField(default=False, help_text='Accept auto-generated subtitles', verbose_name='accept auto subtitles'),
+        ),
+        migrations.AddField(
+            model_name='source',
+            name='sub_langs',
+            field=models.CharField(default='en', help_text='List of subtitles langs to download comma-separated. Example: en,fr',max_length=30),
+        ),
+    ]

tubesync/sync/models.py

@@ -8,6 +8,7 @@ from pathlib import Path
 from django.conf import settings
 from django.db import models
 from django.core.files.storage import FileSystemStorage
+from django.core.validators import RegexValidator
 from django.utils.text import slugify
 from django.utils import timezone
 from django.utils.translation import gettext_lazy as _
@@ -349,6 +350,30 @@ class Source(models.Model):
         help_text=_('Source has failed to index media')
     )
 
+    write_subtitles = models.BooleanField(
+        _('write subtitles'),
+        default=False,
+        help_text=_('Download video subtitles')
+    )
+
+    auto_subtitles = models.BooleanField(
+        _('accept auto-generated subs'),
+        default=False,
+        help_text=_('Accept auto-generated subtitles')
+    )
+    sub_langs = models.CharField(
+        _('subs langs'),
+        max_length=30,
+        default='en',
+        help_text=_('List of subtitles langs to download, comma-separated. Example: en,fr or all,-fr,-live_chat'),
+        validators=[
+            RegexValidator(
+                regex=r"^(\-?[\_\.a-zA-Z]+,)*(\-?[\_\.a-zA-Z]+){1}$",
+                message=_('Subtitle langs must be a comma-separated list of langs. example: en,fr or all,-fr,-live_chat')
+            )
+        ]
+    )
+
     def __str__(self):
         return self.name
 
@@ -1333,7 +1358,8 @@ class Media(models.Model):
         download_youtube_media(self.url, format_str, self.source.extension,
                                str(self.filepath), self.source.write_json, 
                                self.source.sponsorblock_categories, self.source.embed_thumbnail,
-                               self.source.embed_metadata, self.source.enable_sponsorblock)
+                               self.source.embed_metadata, self.source.enable_sponsorblock,
+                              self.source.write_subtitles, self.source.auto_subtitles,self.source.sub_langs )
         # Return the download paramaters
         return format_str, self.source.extension
 

tubesync/sync/templates/sync/source.html

@@ -162,6 +162,22 @@
         </strong></td>
       </tr>
       {% endif %}
+
+      <tr title="{{ _('Are Subtitles downloaded?') }}">
+        <td class="hide-on-small-only">{{ _("Download subtitles?") }}:</td>
+        <td><span class="hide-on-med-and-up">{{ _("Download subtitles?") }}:</span><strong><i class="fas {% if source.write_subtitles %}fa-check{% else %}fa-times{% endif %}"></i></strong></td>
+      </tr>
+
+      {% if source.write_subtitles %}
+      <tr title="{{ _('Are auto subs accepted?') }}">
+        <td class="hide-on-small-only">{{ _("Auto-generated subtitles?") }}:</td>
+        <td><span class="hide-on-med-and-up">{{ _("Auto-generated subtitles?") }}:</span><strong><i class="fas {% if source.auto_subtitles %}fa-check{% else %}fa-times{% endif %}"></i></strong></td>
+      </tr>
+      <tr title="{{ _('Subs langs?') }}">
+        <td class="hide-on-small-only">{{ _("Subs langs?") }}:</td>
+        <td><span class="hide-on-med-and-up">{{ _("Subs langs?") }}:</span><strong>{{source.sub_langs}}</strong></td>
+      </tr>
+      {% endif %}
       
     </table>
   </div>

tubesync/sync/views.py

@@ -299,7 +299,7 @@ class EditSourceMixin:
               'days_to_keep', 'source_resolution', 'source_vcodec', 'source_acodec',
               'prefer_60fps', 'prefer_hdr', 'fallback', 'copy_thumbnails', 'write_nfo', 
               'write_json', 'embed_metadata', 'embed_thumbnail', 'enable_sponsorblock',
-              'sponsorblock_categories')
+              'sponsorblock_categories', 'write_subtitles', 'auto_subtitles', 'sub_langs')
     errors = {
         'invalid_media_format': _('Invalid media format, the media format contains '
                                   'errors or is empty. Check the table at the end of '

tubesync/sync/youtube.py

@@ -66,13 +66,18 @@ def get_media_info(url):
 
 def download_media(url, media_format, extension, output_file, info_json, 
                    sponsor_categories="all", 
-                   embed_thumbnail=False, embed_metadata=False, skip_sponsors=True):
+                   embed_thumbnail=False, embed_metadata=False, skip_sponsors=True, 
+                   write_subtitles=False, auto_subtitles=False, sub_langs='en'):
     '''
         Downloads a YouTube URL to a file on disk.
     '''
 
     def hook(event):
         filename = os.path.basename(event['filename'])
+    
+        if event.get('downloaded_bytes') is None or event.get('total_bytes') is None:
+            return None
+
         if event['status'] == 'error':
             log.error(f'[youtube-dl] error occured downloading: {filename}')
         elif event['status'] == 'downloading':
@@ -110,8 +115,12 @@ def download_media(url, media_format, extension, output_file, info_json,
         'quiet': True,
         'progress_hooks': [hook],
         'writeinfojson': info_json,
-        'postprocessors': []
+        'postprocessors': [],
+        'writesubtitles': write_subtitles,
+        'writeautomaticsub': auto_subtitles,
+        'subtitleslangs': sub_langs.split(','),
     }
+    
     sbopt = {
         'key': 'SponsorBlock',
         'categories': [sponsor_categories]