New DatasetExtractor using a DatasetWorker

Depends on workers/base-worker!411 (merged)

Closes #3 (closed), #6 (closed)

tests/test_worker.py

@@ -19,11 +19,13 @@ def test_process_split(tmp_path, downloaded_images):
     worker = DatasetExtractor()
     # Parse some arguments
     worker.args = Namespace(database=None)
+    worker.data_folder = tmp_path
     worker.configure_cache()
     worker.cached_images = dict()
 
     # Where to save the downloaded images
-    worker.image_folder = tmp_path
+    worker.images_folder = tmp_path / "images"
+    worker.images_folder.mkdir(parents=True)
 
     first_page_id = UUID("e26e6803-18da-4768-be30-a0a68132107c")
     second_page_id = UUID("c673bd94-96b1-4a2e-8662-a4d806940b5f")
@@ -80,11 +82,6 @@ def test_process_split(tmp_path, downloaded_images):
             == f"https://europe-gamma.iiif.teklia.com/iiif/2/public%2Fiam%2F{page_name}.png"
         )
 
-    assert sorted(tmp_path.rglob("*")) == [
-        tmp_path / f"{first_image_id}.jpg",
-        tmp_path / f"{second_image_id}.jpg",
-    ]
-
     # Should have created 17 transcriptions
     assert CachedTranscription.select().count() == 17
     # Check transcription of first line on first page
@@ -127,3 +124,11 @@ def test_process_split(tmp_path, downloaded_images):
     assert tr_entity.length == 23
     assert tr_entity.confidence == 1.0
     assert tr_entity.worker_run_id is None
+
+    # Full structure of the archive
+    assert sorted(tmp_path.rglob("*")) == [
+        tmp_path / "db.sqlite",
+        tmp_path / "images",
+        tmp_path / "images" / f"{first_image_id}.jpg",
+        tmp_path / "images" / f"{second_image_id}.jpg",
+    ]