Merge branch 'check-images' into 'master'

Add check_images command

See merge request !203

arkindex/images/management/commands/check_images.py

+from django.core.management.base import CommandError
+from ponos.management.base import PonosCommand
+from arkindex.project.argparse import CorpusArgument, ElementArgument
+from arkindex.images.models import Image, ImageStatus
+import logging
+
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s [%(levelname)s] %(message)s',
+)
+logger = logging.getLogger()
+
+
+class Command(PonosCommand):
+    help = 'Check images availability'
+
+    def add_arguments(self, parser):
+        super().add_arguments(parser)
+        parser.add_argument(
+            '--force',
+            action='store_true',
+            default=False,
+        )
+        parser.add_argument(
+            '--corpus',
+            help='A corpus to run checks on',
+            type=CorpusArgument(),
+        )
+        parser.add_argument(
+            '--element',
+            help='One or more elements to run checks on',
+            type=ElementArgument(),
+            nargs='+',
+        )
+
+    def validate_args(self, corpus=None, element=None, force=False, **options):
+        if corpus and element:
+            raise CommandError('--corpus and --element cannot be used together')
+        if corpus:
+            images = Image.objects.filter(zones__elements__corpus=corpus).distinct()
+        elif element:
+            images = Image.objects.filter(zones__elements__in=element).distinct()
+        else:
+            images = Image.objects.all()
+
+        if not force:
+            images = images.exclude(status=ImageStatus.Checked)
+
+        return {'images': images}
+
+    def run(self, images=[]):
+        for image in images:
+            logger.info('Checking image {} at {}'.format(str(image.id), image.url))
+            image.perform_check(save=True)

arkindex/images/tests.py → arkindex/images/tests/test_bulk_transcriptions.py

 from arkindex.project.tests import FixtureTestCase
 from arkindex.project.polygon import Polygon
 from arkindex.documents.models import Transcription, TranscriptionType, DataSource
-from arkindex.images.models import Image, ImageServer, ImageStatus
 from arkindex.images.importer import build_transcriptions, save_transcriptions
 
 
@@ -11,7 +10,7 @@ class TestBulkTranscriptions(FixtureTestCase):
     @classmethod
     def setUpTestData(cls):
         super().setUpTestData()
-        cls.img = Image.objects.get(path='img6', server=cls.imgsrv)
+        cls.img = cls.imgsrv.images.get(path='img6')
         pagezone = cls.img.zones.first()
         cls.page = pagezone.elements.first()
         cls.src = DataSource.objects.get(slug='test')
@@ -133,25 +132,3 @@ class TestBulkTranscriptions(FixtureTestCase):
         # should not recreate
         trpolygons = build_transcriptions(self.img, self.page, self.src, items)
         self.assertEqual(len(trpolygons), 0)
-
-
-class TestImageServer(FixtureTestCase):
-    '''
-    Test Image server creation
-    '''
-    def test_check(self):
-        server = ImageServer.objects.create(name='dummy', url='http://test/iiif')
-        image = server.find_image('https://test/iiif/dog.jpg', width=800, height=600)
-        self.assertEqual(image.status, ImageStatus.Error)  # connection error, no exception
-
-
-class TestImage(FixtureTestCase):
-
-    def test_path_not_encoded(self):
-        """
-        Assert that an image's path is not URL encoded when created
-        This can yield to images with URLs such as http://.../test%34test to be encoded as http://.../test%2534test
-        """
-        img = Image.objects.create(server=self.imgsrv, path='!#%:+*')
-        img.refresh_from_db()
-        self.assertEqual(img.path, '!#%:+*')

arkindex/images/tests/test_check_images.py

+from arkindex.project.tests import FixtureTestCase
+from arkindex.images.models import ImageStatus
+from django.core.management import call_command
+from django.core.management.base import CommandError
+from unittest.mock import patch
+
+
+class TestCheckImages(FixtureTestCase):
+
+    @classmethod
+    def setUpClass(cls):
+        super().setUpClass()
+        cls.check_mock = patch('arkindex.images.management.commands.check_images.Image.perform_check').start()
+
+    @classmethod
+    def setUpTestData(cls):
+        super().setUpTestData()
+        cls.p1 = cls.corpus.elements.get(name='Volume 1, page 1r')
+        cls.p2 = cls.corpus.elements.get(name='Volume 1, page 1v')
+
+        # Set two images to Checked
+        cls.imgsrv.images.filter(path='img5').update(status=ImageStatus.Checked)
+        cls.imgsrv.images.filter(path='img6').update(status=ImageStatus.Checked)
+
+        # Create an image linked to zero elements
+        cls.imgsrv.images.create(path='am-outside')
+
+    def tearDown(self):
+        super().tearDown()
+        self.check_mock.reset_mock()
+
+    def test_nothing(self):
+        call_command(
+            'check_images',
+        )
+        self.assertEqual(self.check_mock.call_count, 5)
+
+    def test_corpus(self):
+        call_command(
+            'check_images',
+            corpus=self.corpus,
+        )
+        self.assertEqual(self.check_mock.call_count, 4)
+
+    def test_element(self):
+        call_command(
+            'check_images',
+            element=[self.p1, self.p2],
+        )
+        self.assertEqual(self.check_mock.call_count, 2)
+
+    def test_corpus_xor_element(self):
+        with self.assertRaises(CommandError):
+            call_command(
+                'check_images',
+                corpus=self.corpus,
+                element=[self.p1, ],
+            )
+
+    def test_force(self):
+        call_command(
+            'check_images',
+            force=True,
+        )
+        self.assertEqual(self.check_mock.call_count, 7)

arkindex/images/tests/test_image.py

+from arkindex.project.tests import FixtureTestCase
+from arkindex.images.models import Image
+
+
+class TestImage(FixtureTestCase):
+
+    def test_path_not_encoded(self):
+        """
+        Assert that an image's path is not URL encoded when created
+        This can yield to images with URLs such as http://.../test%34test to be encoded as http://.../test%2534test
+        """
+        img = Image.objects.create(server=self.imgsrv, path='!#%:+*')
+        img.refresh_from_db()
+        self.assertEqual(img.path, '!#%:+*')