Remove repeated prepare_images in processor tests (huggingface#33163)

* Remove repeated prepare_images

* Address comments - update docstring; explanatory comment

tests/models/align/test_processor_align.py

@@ -18,7 +18,6 @@
 import tempfile
 import unittest
 
-import numpy as np
 import pytest
 
 from transformers import BertTokenizer, BertTokenizerFast
@@ -30,8 +29,6 @@
 
 
 if is_vision_available():
-    from PIL import Image
-
     from transformers import AlignProcessor, EfficientNetImageProcessor
 
 
@@ -86,15 +83,6 @@ def get_image_processor(self, **kwargs):
     def tearDown(self):
         shutil.rmtree(self.tmpdirname)
 
-    def prepare_image_inputs(self):
-        """This function prepares a list of PIL images, or a list of numpy arrays if one specifies numpify=True,
-        or a list of PyTorch tensors if one specifies torchify=True.
-        """
-
-        image_inputs = [np.random.randint(255, size=(3, 30, 400), dtype=np.uint8)]
-        image_inputs = [Image.fromarray(np.moveaxis(x, 0, -1)) for x in image_inputs]
-        return image_inputs
-
     def test_save_load_pretrained_default(self):
         tokenizer_slow = self.get_tokenizer()
         tokenizer_fast = self.get_rust_tokenizer()

tests/models/blip/test_processor_blip.py

@@ -15,21 +15,22 @@
 import tempfile
 import unittest
 
-import numpy as np
 import pytest
 
 from transformers.testing_utils import require_vision
 from transformers.utils import is_vision_available
 
+from ...test_processing_common import ProcessorTesterMixin
 
-if is_vision_available():
-    from PIL import Image
 
+if is_vision_available():
     from transformers import AutoProcessor, BertTokenizer, BlipImageProcessor, BlipProcessor, PreTrainedTokenizerFast
 
 
 @require_vision
-class BlipProcessorTest(unittest.TestCase):
+class BlipProcessorTest(ProcessorTesterMixin, unittest.TestCase):
+    processor_class = BlipProcessor
+
     def setUp(self):
         self.tmpdirname = tempfile.mkdtemp()
 
@@ -49,17 +50,6 @@ def get_image_processor(self, **kwargs):
     def tearDown(self):
         shutil.rmtree(self.tmpdirname)
 
-    def prepare_image_inputs(self):
-        """This function prepares a list of PIL images, or a list of numpy arrays if one specifies numpify=True,
-        or a list of PyTorch tensors if one specifies torchify=True.
-        """
-
-        image_inputs = [np.random.randint(255, size=(3, 30, 400), dtype=np.uint8)]
-
-        image_inputs = [Image.fromarray(np.moveaxis(x, 0, -1)) for x in image_inputs]
-
-        return image_inputs
-
     def test_save_load_pretrained_additional_features(self):
         processor = BlipProcessor(tokenizer=self.get_tokenizer(), image_processor=self.get_image_processor())
         processor.save_pretrained(self.tmpdirname)

tests/models/blip_2/test_processor_blip_2.py

@@ -15,21 +15,22 @@
 import tempfile
 import unittest
 
-import numpy as np
 import pytest
 
 from transformers.testing_utils import require_vision
 from transformers.utils import is_vision_available
 
+from ...test_processing_common import ProcessorTesterMixin
 
-if is_vision_available():
-    from PIL import Image
 
+if is_vision_available():
     from transformers import AutoProcessor, Blip2Processor, BlipImageProcessor, GPT2Tokenizer, PreTrainedTokenizerFast
 
 
 @require_vision
-class Blip2ProcessorTest(unittest.TestCase):
+class Blip2ProcessorTest(ProcessorTesterMixin, unittest.TestCase):
+    processor_class = Blip2Processor
+
     def setUp(self):
         self.tmpdirname = tempfile.mkdtemp()
 
@@ -49,17 +50,6 @@ def get_image_processor(self, **kwargs):
     def tearDown(self):
         shutil.rmtree(self.tmpdirname)
 
-    def prepare_image_inputs(self):
-        """This function prepares a list of PIL images, or a list of numpy arrays if one specifies numpify=True,
-        or a list of PyTorch tensors if one specifies torchify=True.
-        """
-
-        image_inputs = [np.random.randint(255, size=(3, 30, 400), dtype=np.uint8)]
-
-        image_inputs = [Image.fromarray(np.moveaxis(x, 0, -1)) for x in image_inputs]
-
-        return image_inputs
-
     def test_save_load_pretrained_additional_features(self):
         processor = Blip2Processor(tokenizer=self.get_tokenizer(), image_processor=self.get_image_processor())
         processor.save_pretrained(self.tmpdirname)

tests/models/chinese_clip/test_processor_chinese_clip.py

@@ -18,23 +18,24 @@
 import tempfile
 import unittest
 
-import numpy as np
 import pytest
 
 from transformers import BertTokenizer, BertTokenizerFast
 from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES
 from transformers.testing_utils import require_vision
 from transformers.utils import FEATURE_EXTRACTOR_NAME, is_vision_available
 
+from ...test_processing_common import ProcessorTesterMixin
 
-if is_vision_available():
-    from PIL import Image
 
+if is_vision_available():
     from transformers import ChineseCLIPImageProcessor, ChineseCLIPProcessor
 
 
 @require_vision
-class ChineseCLIPProcessorTest(unittest.TestCase):
+class ChineseCLIPProcessorTest(ProcessorTesterMixin, unittest.TestCase):
+    processor_class = ChineseCLIPProcessor
+
     def setUp(self):
         self.tmpdirname = tempfile.mkdtemp()
 
@@ -76,6 +77,11 @@ def setUp(self):
         with open(self.image_processor_file, "w", encoding="utf-8") as fp:
             json.dump(image_processor_map, fp)
 
+        tokenizer = self.get_tokenizer()
+        image_processor = self.get_image_processor()
+        processor = ChineseCLIPProcessor(tokenizer=tokenizer, image_processor=image_processor)
+        processor.save_pretrained(self.tmpdirname)
+
     def get_tokenizer(self, **kwargs):
         return BertTokenizer.from_pretrained(self.tmpdirname, **kwargs)
 
@@ -88,17 +94,6 @@ def get_image_processor(self, **kwargs):
     def tearDown(self):
         shutil.rmtree(self.tmpdirname)
 
-    def prepare_image_inputs(self):
-        """This function prepares a list of PIL images, or a list of numpy arrays if one specifies numpify=True,
-        or a list of PyTorch tensors if one specifies torchify=True.
-        """
-
-        image_inputs = [np.random.randint(255, size=(3, 30, 400), dtype=np.uint8)]
-
-        image_inputs = [Image.fromarray(np.moveaxis(x, 0, -1)) for x in image_inputs]
-
-        return image_inputs
-
     def test_save_load_pretrained_default(self):
         tokenizer_slow = self.get_tokenizer()
         tokenizer_fast = self.get_rust_tokenizer()

tests/models/clip/test_processor_clip.py

@@ -18,7 +18,6 @@
 import tempfile
 import unittest
 
-import numpy as np
 import pytest
 
 from transformers import CLIPTokenizer, CLIPTokenizerFast
@@ -30,8 +29,6 @@
 
 
 if is_vision_available():
-    from PIL import Image
-
     from transformers import CLIPImageProcessor, CLIPProcessor
 
 
@@ -79,17 +76,6 @@ def get_image_processor(self, **kwargs):
     def tearDown(self):
         shutil.rmtree(self.tmpdirname)
 
-    def prepare_image_inputs(self):
-        """This function prepares a list of PIL images, or a list of numpy arrays if one specifies numpify=True,
-        or a list of PyTorch tensors if one specifies torchify=True.
-        """
-
-        image_inputs = [np.random.randint(255, size=(3, 30, 400), dtype=np.uint8)]
-
-        image_inputs = [Image.fromarray(np.moveaxis(x, 0, -1)) for x in image_inputs]
-
-        return image_inputs
-
     def test_save_load_pretrained_default(self):
         tokenizer_slow = self.get_tokenizer()
         tokenizer_fast = self.get_rust_tokenizer()

tests/models/clipseg/test_processor_clipseg.py

@@ -18,23 +18,24 @@
 import tempfile
 import unittest
 
-import numpy as np
 import pytest
 
 from transformers import CLIPTokenizer, CLIPTokenizerFast
 from transformers.models.clip.tokenization_clip import VOCAB_FILES_NAMES
 from transformers.testing_utils import require_vision
 from transformers.utils import IMAGE_PROCESSOR_NAME, is_vision_available
 
+from ...test_processing_common import ProcessorTesterMixin
 
-if is_vision_available():
-    from PIL import Image
 
+if is_vision_available():
     from transformers import CLIPSegProcessor, ViTImageProcessor
 
 
 @require_vision
-class CLIPSegProcessorTest(unittest.TestCase):
+class CLIPSegProcessorTest(ProcessorTesterMixin, unittest.TestCase):
+    processor_class = CLIPSegProcessor
+
     def setUp(self):
         self.tmpdirname = tempfile.mkdtemp()
 
@@ -75,16 +76,6 @@ def get_image_processor(self, **kwargs):
     def tearDown(self):
         shutil.rmtree(self.tmpdirname)
 
-    def prepare_image_inputs(self):
-        """This function prepares a list of PIL images, or a list of numpy arrays if one specifies numpify=True,
-        or a list of PyTorch tensors if one specifies torchify=True."""
-
-        image_inputs = [np.random.randint(255, size=(3, 30, 400), dtype=np.uint8)]
-
-        image_inputs = [Image.fromarray(np.moveaxis(x, 0, -1)) for x in image_inputs]
-
-        return image_inputs
-
     def test_save_load_pretrained_default(self):
         tokenizer_slow = self.get_tokenizer()
         tokenizer_fast = self.get_rust_tokenizer()

tests/models/flava/test_processor_flava.py

@@ -19,18 +19,17 @@
 import tempfile
 import unittest
 
-import numpy as np
 import pytest
 
 from transformers import BertTokenizer, BertTokenizerFast
 from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES
 from transformers.testing_utils import require_vision
 from transformers.utils import IMAGE_PROCESSOR_NAME, is_vision_available
 
+from ...test_processing_common import ProcessorTesterMixin
 
-if is_vision_available():
-    from PIL import Image
 
+if is_vision_available():
     from transformers import FlavaImageProcessor, FlavaProcessor
     from transformers.models.flava.image_processing_flava import (
         FLAVA_CODEBOOK_MEAN,
@@ -41,7 +40,9 @@
 
 
 @require_vision
-class FlavaProcessorTest(unittest.TestCase):
+class FlavaProcessorTest(ProcessorTesterMixin, unittest.TestCase):
+    processor_class = FlavaProcessor
+
     def setUp(self):
         self.tmpdirname = tempfile.mkdtemp()
 
@@ -91,17 +92,6 @@ def get_image_processor(self, **kwargs):
     def tearDown(self):
         shutil.rmtree(self.tmpdirname)
 
-    def prepare_image_inputs(self):
-        """This function prepares a list of PIL images, or a list of numpy arrays if one specifies numpify=True,
-        or a list of PyTorch tensors if one specifies torchify=True.
-        """
-
-        image_inputs = [np.random.randint(255, size=(3, 30, 400), dtype=np.uint8)]
-
-        image_inputs = [Image.fromarray(np.moveaxis(x, 0, -1)) for x in image_inputs]
-
-        return image_inputs
-
     def test_save_load_pretrained_default(self):
         tokenizer_slow = self.get_tokenizer()
         tokenizer_fast = self.get_rust_tokenizer()

tests/models/git/test_processor_git.py

@@ -15,21 +15,22 @@
 import tempfile
 import unittest
 
-import numpy as np
 import pytest
 
 from transformers.testing_utils import require_vision
 from transformers.utils import is_vision_available
 
+from ...test_processing_common import ProcessorTesterMixin
 
-if is_vision_available():
-    from PIL import Image
 
+if is_vision_available():
     from transformers import AutoProcessor, BertTokenizer, CLIPImageProcessor, GitProcessor, PreTrainedTokenizerFast
 
 
 @require_vision
-class GitProcessorTest(unittest.TestCase):
+class GitProcessorTest(ProcessorTesterMixin, unittest.TestCase):
+    processor_class = GitProcessor
+
     def setUp(self):
         self.tmpdirname = tempfile.mkdtemp()
 
@@ -51,17 +52,6 @@ def get_image_processor(self, **kwargs):
     def tearDown(self):
         shutil.rmtree(self.tmpdirname)
 
-    def prepare_image_inputs(self):
-        """This function prepares a list of PIL images, or a list of numpy arrays if one specifies numpify=True,
-        or a list of PyTorch tensors if one specifies torchify=True.
-        """
-
-        image_inputs = [np.random.randint(255, size=(3, 30, 400), dtype=np.uint8)]
-
-        image_inputs = [Image.fromarray(np.moveaxis(x, 0, -1)) for x in image_inputs]
-
-        return image_inputs
-
     def test_save_load_pretrained_additional_features(self):
         processor = GitProcessor(tokenizer=self.get_tokenizer(), image_processor=self.get_image_processor())
         processor.save_pretrained(self.tmpdirname)

tests/models/grounding_dino/test_processor_grounding_dino.py

@@ -18,7 +18,6 @@
 import tempfile
 import unittest
 
-import numpy as np
 import pytest
 
 from transformers import BertTokenizer, BertTokenizerFast, GroundingDinoProcessor
@@ -35,8 +34,6 @@
     from transformers.models.grounding_dino.modeling_grounding_dino import GroundingDinoObjectDetectionOutput
 
 if is_vision_available():
-    from PIL import Image
-
     from transformers import GroundingDinoImageProcessor
 
 
@@ -96,18 +93,6 @@ def get_image_processor(self, **kwargs):
     def tearDown(self):
         shutil.rmtree(self.tmpdirname)
 
-    # Copied from tests.models.clip.test_processor_clip.CLIPProcessorTest.prepare_image_inputs
-    def prepare_image_inputs(self):
-        """This function prepares a list of PIL images, or a list of numpy arrays if one specifies numpify=True,
-        or a list of PyTorch tensors if one specifies torchify=True.
-        """
-
-        image_inputs = [np.random.randint(255, size=(3, 30, 400), dtype=np.uint8)]
-
-        image_inputs = [Image.fromarray(np.moveaxis(x, 0, -1)) for x in image_inputs]
-
-        return image_inputs
-
     def get_fake_grounding_dino_output(self):
         torch.manual_seed(42)
         return GroundingDinoObjectDetectionOutput(