bumps auto-archiver to 0.13.4

app/tests/worker/test_worker_main.py

@@ -1,7 +1,6 @@
 from datetime import datetime
-from unittest import mock
 
-from unittest.mock import MagicMock, patch
+from unittest.mock import patch
 
 import pytest
 
@@ -24,7 +23,7 @@ class Test_create_archive_task():
         from app.worker.main import create_archive_task
 
         m_req.id = "this-just-in"
-        m_orchestrator.return_value.run.return_value = iter([Metadata().set_url(self.URL).success()])
+        m_orchestrator.return_value.feed.return_value = iter([Metadata().set_url(self.URL).success()])
 
         task = create_archive_task(self.archive.model_dump_json())
 
@@ -32,7 +31,8 @@ class Test_create_archive_task():
         m_store.assert_called_once_with("interstellar")
         m_insert.assert_called_once()
         m_urls.assert_called_once()
-        m_orchestrator.return_value.run.assert_called_once()
+        m_orchestrator.return_value.feed.assert_called_once()
+        m_orchestrator.return_value.setup.assert_called_once()
 
         assert task["status"] == "success"
         assert task["metadata"]["url"] == self.URL
@@ -47,25 +47,25 @@ class Test_create_archive_task():
     @patch("app.worker.main.get_orchestrator_args")
     def test_raise_db_error(self, m_args, m_orchestrator):
         from app.worker.main import create_archive_task
-        m_orchestrator.return_value.run.side_effect = Exception("Orchestrator failed")
+        m_orchestrator.return_value.feed.side_effect = Exception("Orchestrator failed")
 
         with pytest.raises(Exception) as e:
             create_archive_task(self.archive.model_dump_json())
         assert str(e.value) == "Orchestrator failed"
         m_args.assert_called_once()
-        m_orchestrator.return_value.run.assert_called_once()
+        m_orchestrator.return_value.feed.assert_called_once()
 
     @patch("app.worker.main.ArchivingOrchestrator")
     @patch("app.worker.main.insert_result_into_db", return_value=None)
     @patch("app.worker.main.get_orchestrator_args")
     def test_raise_empty_result(self, m_args, m_insert, m_orchestrator):
         from app.worker.main import create_archive_task
-        m_orchestrator.return_value.run.return_value = iter([None])
+        m_orchestrator.return_value.feed.return_value = iter([None])
 
         with pytest.raises(Exception) as e:
             create_archive_task(self.archive.model_dump_json())
         assert str(e.value) == "UNABLE TO archive: https://example-live.com"
-        m_orchestrator.return_value.run.assert_called_once()
+        m_orchestrator.return_value.feed.assert_called_once()
 
 
 class Test_create_sheet_task():
@@ -85,12 +85,13 @@ class Test_create_sheet_task():
         mock_metadata = Metadata().set_url(self.URL).success()
         mock_metadata.add_media(Media("fn1.txt", urls=["outcome1.com"]))
 
-        m_orchestrator.return_value.run.return_value = iter([False, mock_metadata, mock_metadata])
+        m_orchestrator.return_value.feed.return_value = iter([False, mock_metadata, mock_metadata])
 
         res = create_sheet_task(self.sheet.model_dump_json())
 
         m_args.assert_called_once_with("interstellar", True, ["--gsheet_feeder.sheet_id", "123"])
-        m_orchestrator.return_value.run.assert_called_once()
+        m_orchestrator.return_value.setup.assert_called_once()
+        m_orchestrator.return_value.feed.assert_called_once()
         m_store.assert_called_with("interstellar")
         m_store.call_count == 2
         m_uuid.call_count == 2

app/worker/main.py

@@ -33,10 +33,10 @@ def create_archive_task(self, archive_json: str):
 
     # call auto-archiver
     args = get_orchestrator_args(archive.group_id, False, [archive.url])
-    # args = get_orchestrator_args(archive.group_id, False, [archive.url, "--extractors", "generic_extractor"])
-    logger.debug(args)
     try:
-        result = next(ArchivingOrchestrator().run(args), None)
+        orchestrator = ArchivingOrchestrator()
+        orchestrator.setup(args)
+        result = next(orchestrator.feed())
     except SystemExit as e:
         log_error(e, f"create_archive_task: SystemExit from AA")
     except Exception as e:
@@ -61,11 +61,12 @@ def create_sheet_task(self, sheet_json: str):
     logger.info(f"[queue={queue_name}] SHEET START {sheet=}")
 
     args = get_orchestrator_args(sheet.group_id, True, ["--gsheet_feeder.sheet_id", sheet.sheet_id])
-    logger.info(f"[queue={queue_name}] {args=}")
+    orchestrator = ArchivingOrchestrator()
+    orchestrator.setup(args)
 
     stats = {"archived": 0, "failed": 0, "errors": []}
     try:
-        for result in ArchivingOrchestrator().run(args):
+        for result in orchestrator.feed():
             try:
                 assert result, f"ERROR archiving URL for sheet {sheet.sheet_id}"
                 archive = schemas.ArchiveCreate(

poetry.lock

@@ -154,14 +154,14 @@ cryptography = "*"
 
 [[package]]
 name = "auto-archiver"
-version = "0.13.2"
+version = "0.13.4"
 description = "Automatically archive links to videos, images, and social media content from Google Sheets (and more)."
 optional = false
 python-versions = "<3.13,>=3.10"
 groups = ["main"]
 files = [
-    {file = "auto_archiver-0.13.2-py3-none-any.whl", hash = "sha256:672671080bdc2e4cd50792b3521a8a1d70aabb50ee3f779ed30879162b0c352b"},
-    {file = "auto_archiver-0.13.2.tar.gz", hash = "sha256:b0d5505206bdb02f2ddb1b3a3a622780cc06ace0f3440b272f1d9bc9c314c9e2"},
+    {file = "auto_archiver-0.13.4-py3-none-any.whl", hash = "sha256:490ee0dbc86e3481ee06cdbfbbaf397cbc9733b4aaac8cac233f29af5dc4ba53"},
+    {file = "auto_archiver-0.13.4.tar.gz", hash = "sha256:dac206f643e8101bb1efdea2e6cbdfaca1e3ae50cfe3fa34b466b7518337d675"},
 ]
 
 [package.dependencies]

worker.Dockerfile

@@ -1,5 +1,5 @@
 # From python:3.10
-FROM bellingcat/auto-archiver
+FROM bellingcat/auto-archiver:v0.13.4
 
 # set work directory
 WORKDIR /aa-api