filter by vid_type for add to download queue

tubearchivist/home/src/download/queue.py

@@ -162,11 +162,11 @@ class PendingList(PendingIndex):
 
     def _process_entry(self, entry):
         """process single entry from url list"""
+        vid_type = self._get_vid_type(entry)
         if entry["type"] == "video":
-            vid_type = self._get_vid_type(entry)
             self._add_video(entry["url"], vid_type)
         elif entry["type"] == "channel":
-            self._parse_channel(entry["url"])
+            self._parse_channel(entry["url"], vid_type)
         elif entry["type"] == "playlist":
             self._parse_playlist(entry["url"])
             PlaylistSubscription().process_url_str([entry], subscribed=False)
@@ -178,21 +178,21 @@ class PendingList(PendingIndex):
         """add vid type enum if available"""
         vid_type_str = entry.get("vid_type")
         if not vid_type_str:
-            return VideoTypeEnum.VIDEOS
+            return VideoTypeEnum.UNKNOWN
 
         return VideoTypeEnum(vid_type_str)
 
-    def _add_video(self, url, vid_type=VideoTypeEnum.VIDEOS):
+    def _add_video(self, url, vid_type):
         """add video to list"""
         if url not in self.missing_videos and url not in self.to_skip:
             self.missing_videos.append((url, vid_type))
         else:
             print(f"{url}: skipped adding already indexed video to download.")
 
-    def _parse_channel(self, url):
+    def _parse_channel(self, url, vid_type):
         """add all videos of channel to list"""
         video_results = ChannelSubscription().get_last_youtube_videos(
-            url, limit=False
+            url, limit=False, query_filter=vid_type
         )
         for video_id, _, vid_type in video_results:
             self._add_video(video_id, vid_type)

tubearchivist/home/src/download/subscriptions.py

@@ -36,30 +36,15 @@ class ChannelSubscription:
 
         return all_channels
 
-    def get_last_youtube_videos(self, channel_id, limit=True):
+    def get_last_youtube_videos(
+        self, channel_id, limit=True, query_filter=VideoTypeEnum.UNKNOWN
+    ):
         """get a list of last videos from channel"""
-
-        queries = [
-            (
-                VideoTypeEnum.VIDEOS,
-                VideoTypeEnum.VIDEOS.value,
-                self.config["subscriptions"]["channel_size"],
-            ),
-            (
-                VideoTypeEnum.STREAMS,
-                VideoTypeEnum.STREAMS.value,
-                self.config["subscriptions"]["live_channel_size"],
-            ),
-            (
-                VideoTypeEnum.SHORTS,
-                VideoTypeEnum.SHORTS.value,
-                self.config["subscriptions"]["shorts_channel_size"],
-            ),
-        ]
+        queries = self._build_queries(query_filter, limit)
 
         last_videos = []
 
-        for vid_type, url, limit_amount in queries:
+        for vid_type, limit_amount in queries:
             obs = {
                 "skip_download": True,
                 "extract_flat": True,
@@ -67,8 +52,9 @@ class ChannelSubscription:
             if limit:
                 obs["playlistend"] = limit_amount
 
+            path = vid_type.value
             channel = YtWrap(obs, self.config).extract(
-                f"https://www.youtube.com/channel/{channel_id}/{url}"
+                f"https://www.youtube.com/channel/{channel_id}/{path}"
             )
             if not channel:
                 continue
@@ -78,6 +64,36 @@ class ChannelSubscription:
 
         return last_videos
 
+    def _build_queries(self, query_filter, limit):
+        """build query list for vid_type"""
+        limit_map = {
+            "videos": self.config["subscriptions"]["channel_size"],
+            "streams": self.config["subscriptions"]["live_channel_size"],
+            "shorts": self.config["subscriptions"]["shorts_channel_size"],
+        }
+
+        queries = []
+
+        if query_filter and query_filter.value != "unknown":
+            if limit:
+                query_limit = limit_map.get(query_filter.value)
+            else:
+                query_limit = False
+
+            queries.append((query_filter, query_limit))
+
+            return queries
+
+        for query_item, default_limit in limit_map.items():
+            if limit:
+                query_limit = default_limit
+            else:
+                query_limit = False
+
+            queries.append((VideoTypeEnum(query_item), query_limit))
+
+        return queries
+
     def find_missing(self):
         """add missing videos from subscribed channels to pending"""
         all_channels = self.get_channels()