Merge pull request #238 from captain-pool:issue_203

PiperOrigin-RevId: 264108693

Author: copybara-github

tensorflow_datasets/image/__init__.py

@@ -70,6 +70,7 @@
 from tensorflow_datasets.image.smallnorb import Smallnorb
 from tensorflow_datasets.image.so2sat import So2sat
 from tensorflow_datasets.image.stanford_dogs import StanfordDogs
+from tensorflow_datasets.image.stanford_online_products import StanfordOnlineProducts
 from tensorflow_datasets.image.sun import Sun397
 from tensorflow_datasets.image.svhn import SvhnCropped
 from tensorflow_datasets.image.uc_merced import UcMerced

tensorflow_datasets/image/stanford_online_products.py

@@ -0,0 +1,88 @@
+# coding=utf-8
+# Copyright 2019 The TensorFlow Datasets Authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Stanford Online Products Dataset."""
+import csv
+import os
+import tensorflow as tf
+import tensorflow_datasets.public_api as tfds
+
+_DOWNLOAD_LINK = "ftp://cs.stanford.edu/cs/cvgl/Stanford_Online_Products.zip"
+_SPLITS = {tfds.Split.TRAIN: "Ebay_train", tfds.Split.TEST: "Ebay_test"}
+
+_SUPER_CLASSES = [
+    "bicycle", "cabinet", "chair", "coffee_maker", "fan", "kettle", "lamp",
+    "mug", "sofa", "stapler", "table", "toaster"
+]
+_CITATION = """\
+@inproceedings{song2016deep,
+ author    = {Song, Hyun Oh and Xiang, Yu and Jegelka, Stefanie and Savarese, Silvio},
+ title     = {Deep Metric Learning via Lifted Structured Feature Embedding},
+ booktitle = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
+ year      = {2016}
+}
+"""
+
+
+class StanfordOnlineProducts(tfds.core.GeneratorBasedBuilder):
+  """Stanford Online Products Dataset."""
+
+  VERSION = tfds.core.Version("1.0.0")
+
+  def _info(self):
+    return tfds.core.DatasetInfo(
+        description=("Stanford Online Products Dataset"),
+        builder=self,
+        citation=_CITATION,
+        urls=["http://cvgl.stanford.edu/projects/lifted_struct/"],
+        features=tfds.features.FeaturesDict({
+            "class_id":
+                tfds.features.ClassLabel(num_classes=22634),
+            "super_class_id/num":
+                tfds.features.ClassLabel(num_classes=len(_SUPER_CLASSES)),
+            "super_class_id":
+                tfds.features.ClassLabel(names=_SUPER_CLASSES),
+            "image":
+                tfds.features.Image()
+        }))
+
+  def _split_generators(self, dl_manager):
+    dl_path = dl_manager.download_and_extract(_DOWNLOAD_LINK)
+    folder_path = os.path.join(dl_path, "Stanford_Online_Products")
+    return [  # pylint:disable=g-complex-comprehension
+        tfds.core.SplitGenerator(
+            name=k,
+            gen_kwargs={"file_path": os.path.join(folder_path, "%s.txt" % v)})
+        for k, v in _SPLITS.items()
+    ]
+
+  def _generate_examples(self, file_path):
+    """Images of Product from the Data Directory.
+
+    Args:
+      file_path: str, path to the Ebay_(train/test/info).txt file. Having
+        Columns ['class_id', 'super_class_id', 'path']
+    Yields:
+      Dataset examples.
+    """
+    with tf.io.gfile.GFile(file_path, "r") as file_:
+      dataset = csv.DictReader(file_, delimiter=" ")
+      for i, row in enumerate(dataset):
+        yield i, {
+            "class_id": int(row["class_id"]) - 1,
+            "super_class_id/num": int(row["super_class_id"]) - 1,
+            "super_class_id": _SUPER_CLASSES[int(row["super_class_id"]) - 1],
+            "image": os.path.join(os.path.dirname(file_path), row["path"])
+        }

tensorflow_datasets/image/stanford_online_products_test.py

@@ -0,0 +1,26 @@
+# coding=utf-8
+# Copyright 2019 The TensorFlow Datasets Authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from tensorflow_datasets.image import stanford_online_products
+import tensorflow_datasets.testing as tfds_test
+
+
+class StanfordOnlineProductsTest(tfds_test.DatasetBuilderTestCase):
+  DATASET_CLASS = stanford_online_products.StanfordOnlineProducts
+  SPLITS = {"train": 3, "test": 3}
+
+
+if __name__ == "__main__":
+  tfds_test.test_main()

tensorflow_datasets/testing/test_data/fake_examples/stanford_online_products/Stanford_Online_Products/Ebay_test.txt

@@ -0,0 +1,4 @@
+ class_id super_class_id path
+115308 21635 11 table_final/381052627539_0.JPG
+80418 14815 4 coffee_maker_final/371344103996_3.JPG
+82898 15277 5 fan_final/321823124835_0.JPG

tensorflow_datasets/testing/test_data/fake_examples/stanford_online_products/Stanford_Online_Products/Ebay_train.txt

@@ -0,0 +1,4 @@
+ class_id super_class_id path
+2607 365 1 bicycle_final/181415767003_7.JPG
+16154 2727 4 coffee_maker_final/111737519223_7.JPG
+25971 4540 6 kettle_final/151781232926_10.JPG