[BUGFIX] Fix extractor IDs in yt-dlp download archive (#528)

src/ytdl_sub/entries/base_entry.py

@@ -14,6 +14,7 @@ from yt_dlp.utils import sanitize_filename
 from ytdl_sub.entries.variables.kwargs import DESCRIPTION
 from ytdl_sub.entries.variables.kwargs import EPOCH
 from ytdl_sub.entries.variables.kwargs import EXTRACTOR
+from ytdl_sub.entries.variables.kwargs import IE_KEY
 from ytdl_sub.entries.variables.kwargs import TITLE
 from ytdl_sub.entries.variables.kwargs import UID
 from ytdl_sub.entries.variables.kwargs import UPLOADER
@@ -101,7 +102,10 @@ class BaseEntryVariables:
         str
             The ytdl extractor name
         """
-        return self.kwargs(EXTRACTOR)
+        # pylint: disable=line-too-long
+        # Taken from https://github.com/yt-dlp/yt-dlp/blob/e6ab678e36c40ded0aae305bbb866cdab554d417/yt_dlp/YoutubeDL.py#L3514
+        # pylint: enable=line-too-long
+        return self.kwargs_get(EXTRACTOR) or self.kwargs(IE_KEY)
 
     @property
     def epoch(self: "BaseEntry") -> int:

src/ytdl_sub/entries/variables/kwargs.py

@@ -50,6 +50,7 @@ SPLIT_BY_CHAPTERS_PARENT_ENTRY = _("split_by_chapters_parent_entry", backend=Tru
 COMMENTS = _("comments", backend=True)
 UID = _("id")
 EXTRACTOR = _("extractor")
+IE_KEY = _("ie_key")
 EPOCH = _("epoch")
 CHANNEL = _("channel")
 EXT = _("ext")

src/ytdl_sub/ytdl_additions/enhanced_download_archive.py

@@ -11,6 +11,7 @@ from typing import Optional
 from typing import Set
 
 from yt_dlp import DateRange
+from yt_dlp.utils import make_archive_id
 
 from ytdl_sub.entries.entry import Entry
 from ytdl_sub.entries.variables.kwargs import SPLIT_BY_CHAPTERS_PARENT_ENTRY
@@ -313,7 +314,7 @@ class DownloadMappings:
         """
         lines: List[str] = []
         for entry_id, metadata in self._entry_mappings.items():
-            lines.append(f"{metadata.extractor} {entry_id}")
+            lines.append(make_archive_id(ie=metadata.extractor, video_id=entry_id))
 
         return DownloadArchive(download_archive_lines=lines)
 

tests/e2e/bandcamp/test_bandcamp.py

@@ -1,7 +1,9 @@
 import pytest
+from conftest import assert_logs
 from expected_download import assert_expected_downloads
 from expected_transaction_log import assert_transaction_log_matches
 
+import ytdl_sub.downloaders.downloader
 from ytdl_sub.subscriptions.subscription import Subscription
 
 
@@ -59,3 +61,19 @@ class TestBandcamp:
             dry_run=dry_run,
             expected_download_summary_file_name="bandcamp/test_artist_url.json",
         )
+
+        # Ensure another invocation will hit ExistingVideoReached
+        if not dry_run:
+            with assert_logs(
+                logger=ytdl_sub.downloaders.downloader.download_logger,
+                expected_message="ExistingVideoReached, stopping additional downloads",
+                log_level="debug",
+            ):
+                transaction_log = discography_subscription.download()
+
+            assert transaction_log.is_empty
+            assert_expected_downloads(
+                output_directory=output_directory,
+                dry_run=dry_run,
+                expected_download_summary_file_name="bandcamp/test_artist_url.json",
+            )