Merge pull request #253 from RockefellerArchiveCenter/development

Factor out method to get TIFF files

create_derivatives/migrations/0004_alter_bag_origin.py

@@ -0,0 +1,18 @@
+# Generated by Django 4.0.10 on 2024-10-21 20:44
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('create_derivatives', '0003_bag_pdf_path'),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name='bag',
+            name='origin',
+            field=models.CharField(choices=[('aurora', 'Aurora'), ('legacy_digital', 'Legacy Digital Processing'), ('digitization', 'Digitization'), ('av_digitization', 'Audiovisual Digitization')], default='aurora', max_length=20),
+        ),
+    ]

create_derivatives/routines.py

@@ -76,8 +76,11 @@ def get_tiff_file_paths(self, bag_path):
             tiff_files (list of pathlib.Paths): absolute filepaths for TIFF files.
         """
         service_dir = Path(bag_path, "data", "service")
-        if service_dir.is_dir() and any(service_dir.iterdir()):
-            tiff_files_dir = Path(bag_path, "data", "service")
+        master_dir = Path(bag_path, "data", "master")
+        for tiff_dir in service_dir, master_dir:
+            if tiff_dir.is_dir() and any(tiff_dir.iterdir()):
+                tiff_files_dir = tiff_dir
+                break
         else:
             tiff_files_dir = Path(bag_path, "data")
         return matching_files(tiff_files_dir, prepend=True)
@@ -149,24 +152,6 @@ def process_bag(self, bag):
         tiff_files = self.get_tiff_file_paths(bag.bag_path)
         self.convert_to_strips(tiff_files)
 
-    def get_tiff_file_paths(self, bag_path):
-        """Determines the location of TIFF files in the bag.
-
-        Args:
-            bag_path (str): root bag path.
-        Returns:
-            tiff_files (list of pathlib.Paths): absolute filepaths for TIFF files.
-        """
-        service_dir = Path(bag_path, "data", "service")
-        master_dir = Path(bag_path, "data", "master")
-        for tiff_dir in service_dir, master_dir:
-            if tiff_dir.is_dir() and any(tiff_dir.iterdir()):
-                tiff_files_dir = tiff_dir
-                break
-        else:
-            tiff_files_dir = Path(bag_path, "data")
-        return matching_files(tiff_files_dir, prepend=True)
-
     def convert_to_strips(self, tiff_files):
         """Converts tiled TIFFs to stripped TIFFs.