Make S3 versions the defaults.

PiperOrigin-RevId: 290973467

Author: pierrot0

README.md

@@ -145,7 +145,7 @@ input pipelines with `tf.data` but use whatever you'd like for your model
 components.
 
 ```python
-train_ds = tfds.load("mnist", split=tfds.Split.TRAIN)
+train_ds = tfds.load("mnist", split="train")
 train_ds = train_ds.shuffle(1024).batch(128).repeat(5).prefetch(10)
 for example in tfds.as_numpy(train_ds):
   numpy_images, numpy_labels = example["image"], example["label"]

docs/_index.ipynb

@@ -47,7 +47,7 @@
         "print(tfds.list_builders())\n",
         "\n",
         "# Construct a tf.data.Dataset\n",
-        "dataset = tfds.load(name=\"mnist\", split=tfds.Split.TRAIN)\n",
+        "dataset = tfds.load(name=\"mnist\", split=\"train\")\n",
         "\n",
         "# Build your input pipeline\n",
         "dataset = dataset.shuffle(1024).batch(32).prefetch(tf.data.experimental.AUTOTUNE)\n",

docs/_index.yaml

@@ -33,7 +33,7 @@ landing_page:
         print(tfds.list_builders())
 
         # Construct a tf.data.Dataset
-        dataset = tfds.load(name="mnist", split=tfds.Split.TRAIN)
+        dataset = tfds.load(name="mnist", split="train")
 
         # Build your input pipeline
         dataset = dataset.shuffle(1024).batch(32).prefetch(tf.data.experimental.AUTOTUNE)

docs/api_docs/python/tfds/core/SplitGenerator.md

@@ -33,7 +33,6 @@ source</a>
 ``` python
 __init__(
     name,
-    num_shards=1,
     gen_kwargs=None
 )
 ```

oss_scripts/oss_pip_install.sh

@@ -32,7 +32,7 @@ pip install -q -U numpy
 # data load
 pip install -e .
 python -c "import tensorflow_datasets as tfds"
-python -c "import tensorflow_datasets as tfds; tfds.load('mnist', split=tfds.Split.TRAIN)"
+python -c "import tensorflow_datasets as tfds; tfds.load('mnist:3.*.*', split='train')"
 
 # Then install the test dependencies
 pip install -e .[tests]

tensorflow_datasets/audio/groove.py

@@ -77,15 +77,12 @@ def __init__(self, split_bars=None, include_audio=True, audio_rate=16000,
     else:
       name_parts.append("midionly")
 
+    v1 = tfds.core.Version(
+        "1.0.0", experiments={tfds.core.Experiment.S3: False})
+    v2 = tfds.core.Version(
+        "2.0.0", "New split API (https://tensorflow.org/datasets/splits)")
     super(GrooveConfig, self).__init__(
-        name="-".join(name_parts),
-        version=tfds.core.Version(
-            "1.0.0", experiments={tfds.core.Experiment.S3: False}),
-        supported_versions=[
-            tfds.core.Version(
-                "2.0.0",
-                "New split API (https://tensorflow.org/datasets/splits)"),
-        ],
+        name="-".join(name_parts), version=v2, supported_versions=[v1],
         **kwargs)
     self.split_bars = split_bars
     self.include_audio = include_audio

tensorflow_datasets/audio/librispeech.py

@@ -149,21 +149,16 @@ def _make_builder_configs():
           vocab_size=2**15),
   ]
   configs = []
+  v001 = tfds.core.Version(
+      "0.0.1", experiments={tfds.core.Experiment.S3: False})
+  v1 = tfds.core.Version(
+      "1.0.0", "New split API (https://tensorflow.org/datasets/splits)")
   for text_encoder_config in text_encoder_configs:
     for data in _DATA_OPTIONS:
       config = LibrispeechConfig(
-          version=tfds.core.Version(
-              "0.0.1", experiments={tfds.core.Experiment.S3: False}),
-          supported_versions=[
-              tfds.core.Version(
-                  "1.0.0",
-                  "New split API (https://tensorflow.org/datasets/splits)"),
-          ],
+          version=v1, supported_versions=[v001],
           text_encoder_config=text_encoder_config,
           data=data)
-      # Version history:
-      # 1.0.0: S3 (new shuffling, sharding and slicing mechanism).
-      # 0.0.1: Initial version.
       configs.append(config)
   return configs
 

tensorflow_datasets/core/dataset_builder_test.py

@@ -163,12 +163,12 @@ def test_load_from_gcs(self):
               data_dir=tmp_dir,
               with_info=True)
       self.assertSetEqual(
-          set(["dataset_info.json", "image.image.json",
-               "mnist-test.counts.txt-00000-of-00001",
+          set(["dataset_info.json",
+               "image.image.json",
                "mnist-test.tfrecord-00000-of-00001",
-               "mnist-train.counts.txt-00000-of-00001"] +
-              ["mnist-train.tfrecord-0000%d-of-00010" % i for i in range(10)]),
-          set(tf.io.gfile.listdir(os.path.join(tmp_dir, "mnist/1.0.0"))))
+               "mnist-train.tfrecord-00000-of-00001",
+              ]),
+          set(tf.io.gfile.listdir(os.path.join(tmp_dir, "mnist/3.0.0"))))
 
       self.assertEqual(set(info.splits.keys()), set(["train", "test"]))
 
@@ -350,7 +350,7 @@ def setUp(self):
     def load_mnist_dataset_info(self):
       mnist_info_path = os.path.join(
           utils.tfds_dir(),
-          "testing/test_data/dataset_info/mnist/1.0.0",
+          "testing/test_data/dataset_info/mnist/3.0.0",
       )
       mnist_info_path = os.path.normpath(mnist_info_path)
       self.read_from_directory(mnist_info_path)

tensorflow_datasets/core/dataset_info_test.py

@@ -35,9 +35,9 @@
 
 _TFDS_DIR = py_utils.tfds_dir()
 _INFO_DIR = os.path.join(_TFDS_DIR, "testing", "test_data", "dataset_info",
-                         "mnist", "1.0.0")
+                         "mnist", "3.0.0")
 _INFO_DIR_UNLABELED = os.path.join(_TFDS_DIR, "testing", "test_data",
-                                   "dataset_info", "mnist_unlabeled", "1.0.0")
+                                   "dataset_info", "mnist_unlabeled", "3.0.0")
 _NON_EXISTENT_DIR = os.path.join(_TFDS_DIR, "non_existent_dir")
 
 
@@ -305,7 +305,7 @@ def test_updates_on_bucket_info(self):
 
 INFO_STR = """tfds.core.DatasetInfo(
     name='mnist',
-    version=1.0.0,
+    version=3.0.0,
     description='The MNIST database of handwritten digits.',
     homepage='https://storage.googleapis.com/cvdf-datasets/mnist/',
     features=FeaturesDict({

tensorflow_datasets/image/caltech.py

@@ -49,11 +49,11 @@
 class Caltech101(tfds.core.GeneratorBasedBuilder):
   """Caltech-101."""
 
-  VERSION = tfds.core.Version("1.1.0",
-                              experiments={tfds.core.Experiment.S3: False})
+  VERSION = tfds.core.Version(
+      "3.0.0", "New split API (https://tensorflow.org/datasets/splits)")
   SUPPORTED_VERSIONS = [
       tfds.core.Version(
-          "3.0.0", "New split API (https://tensorflow.org/datasets/splits)"),
+          "1.1.0", experiments={tfds.core.Experiment.S3: False}),
   ]
 
   def _info(self):