On full post update also check media exists.

Try to download missing or broken media.

rep0st/db/post.py

@@ -227,11 +227,12 @@ def get_posts(self, type: Optional[str] = None) -> Query[Post]:
   @transactional()
   def get_posts_missing_features(self, type: Optional[Type] = None):
     session = self._get_session()
-    q = session.query(Post)
+    q = session.query(Post).join(Post.features, isouter=True)
     if type:
       q = q.filter(Post.type == type)
     return q.filter(
-        and_(Post.status == Status.NOT_INDEXED, Post.deleted == False))
+        and_(Post.status == Status.NOT_INDEXED, Post.deleted == False,
+             Feature.id == None))
 
   @transactional()
   def post_count(self) -> int:

rep0st/service/download_media_service.py

@@ -4,7 +4,7 @@
 from injector import Binder, Module, inject, singleton
 from prometheus_client import Counter
 
-from rep0st.db.post import Post, Type
+from rep0st.db.post import Post, Status, Type
 from rep0st.pr0gramm.api import APIException, Pr0grammAPI, Pr0grammAPIModule
 from rep0st.service.media_service import _MediaDirectory, _MediaFlagModule
 
@@ -44,7 +44,7 @@ def download_media(self, post: Post):
 
     def _download_media(path: str, dl_fn, dir_prefix: str = ''):
       media_file = self.media_dir / dir_prefix / path
-      if media_file.is_file():
+      if media_file.is_file() and post.status != Status.MEDIA_BROKEN:
         log.debug(
             f'Media for post {post.id} found at location {media_file.absolute()}, skipping download'
         )

rep0st/service/post_service.py

@@ -111,21 +111,19 @@ def _process_batch(self, batch_start_id: int, batch_end_id: int,
           to_save.append(post_from_db)
         continue
       # Post is in both DB and API. Potentially update it.
-      dirty = False
       if post_from_db.deleted:
         log.debug(
             f'Unmarking post as deleted since the API contains it: {post_from_db}'
         )
         post_from_db.deleted = False
-        dirty = True
       if post_from_db.flags != post_from_api.flags:
         log.debug(
             f'Updating flags of post since they changed: {post_from_db}. post_from_db.flags={post_from_db.flags}, post_from_api.flags={post_from_api.flags}'
         )
         post_from_db.flags = post_from_api.flags
-        dirty = True
-      if dirty:
-        to_save.append(post_from_db)
+      # Download media if not exists or broken.
+      self._download_media(post_from_db)
+      to_save.append(post_from_db)
     # TODO(https://github.com/ReneHollander/rep0st/issues/42): Sync updates posts to elasticsearch index.
     self.post_repository.persist_all(to_save)