def test_parse_from_sequence_example_missing_feature_list(self):
with tf.Graph().as_default():
empty_proto = text_format.Parse(
"""
feature_lists {
feature_list {
key: "utility2"
value {
feature { float_list { value: 0.0 } }
}
}
}
""", tf.train.SequenceExample())
features = data_lib.parse_from_sequence_example(
tf.convert_to_tensor(value=[empty_proto.SerializeToString()]),
list_size=2,
context_feature_spec=None,
example_feature_spec={
"utility": EXAMPLE_FEATURE_SPEC["utility"]
})
features_0 = data_lib.parse_from_sequence_example(
tf.convert_to_tensor(value=[empty_proto.SerializeToString()]),
context_feature_spec=None,
example_feature_spec={
"utility": EXAMPLE_FEATURE_SPEC["utility"],
"utility2": EXAMPLE_FEATURE_SPEC["utility"]
})
with tf.compat.v1.Session() as sess:
sess.run(tf.compat.v1.local_variables_initializer())
feature_map, feature_0_map = sess.run([features, features_0])
self.assertAllEqual([1, 2, 1], feature_map["utility"].shape)
self.assertAllEqual([1, 1, 1], feature_0_map["utility"].shape)
def test_parse_from_sequence_example_missing_frame_exception(self):
with tf.Graph().as_default():
missing_frame_proto = text_format.Parse(
"""
feature_lists {
feature_list {
key: "utility"
value {
feature { float_list { value: 0.0 } }
feature { }
}
}
}
""", tf.train.SequenceExample())
features = data_lib.parse_from_sequence_example(
tf.convert_to_tensor(
value=[missing_frame_proto.SerializeToString()]),
list_size=2,
context_feature_spec=None,
example_feature_spec={
"utility": EXAMPLE_FEATURE_SPEC["utility"]
})
with tf.compat.v1.Session() as sess:
sess.run(tf.compat.v1.local_variables_initializer())
with self.assertRaisesRegexp(
tf.errors.InvalidArgumentError,
r"Unexpected number of elements in feature utility"):
sess.run(features)
def test_parse_from_sequence_example_with_small_list_size(self):
with tf.Graph().as_default():
features = data_lib.parse_from_sequence_example(
tf.convert_to_tensor(value=[
SEQ_EXAMPLE_PROTO_1.SerializeToString(),
]),
list_size=1,
context_feature_spec=CONTEXT_FEATURE_SPEC,
example_feature_spec=EXAMPLE_FEATURE_SPEC)
with tf.compat.v1.Session() as sess:
sess.run(tf.compat.v1.local_variables_initializer())
feature_map = sess.run(features)
self.assertEqual(sorted(feature_map),
["query_length", "unigrams", "utility"])
self.assertAllEqual(feature_map["unigrams"].dense_shape,
[1, 1, 3])
self.assertAllEqual(feature_map["unigrams"].indices,
[[0, 0, 0]])
self.assertAllEqual(feature_map["unigrams"].values,
[b"tensorflow"])
self.assertAllEqual(feature_map["query_length"], [[3]])
self.assertAllEqual(feature_map["utility"], [[[0.]]])
# Check static shapes for dense tensors.
self.assertAllEqual([1, 1], feature_map["query_length"].shape)
self.assertAllEqual([1, 1, 1], feature_map["utility"].shape)
def test_parse_from_sequence_example(self):
features = data_lib.parse_from_sequence_example(
tf.convert_to_tensor(value=[
SEQ_EXAMPLE_PROTO_1.SerializeToString(),
SEQ_EXAMPLE_PROTO_2.SerializeToString(),
]),
context_feature_spec=CONTEXT_FEATURE_SPEC,
example_feature_spec=EXAMPLE_FEATURE_SPEC)
with tf.compat.v1.Session() as sess:
sess.run(tf.compat.v1.local_variables_initializer())
feature_map = sess.run(features)
self.assertEqual(
sorted(feature_map), ["query_length", "unigrams", "utility"])
self.assertAllEqual(feature_map["unigrams"].dense_shape, [2, 2, 3])
self.assertAllEqual(
feature_map["unigrams"].indices,
[[0, 0, 0], [0, 1, 0], [0, 1, 1], [0, 1, 2], [1, 0, 0]])
self.assertAllEqual(feature_map["unigrams"].values,
[b"tensorflow", b"learning", b"to", b"rank", b"gbdt"])
self.assertAllEqual(feature_map["query_length"], [[3], [2]])
self.assertAllEqual(feature_map["utility"], [[[0.], [1.]], [[0.], [-1.]]])
# Check static shapes for dense tensors.
self.assertAllEqual([2, 1],
features["query_length"].get_shape().as_list())
self.assertAllEqual([2, 2, 1], features["utility"].get_shape().as_list())
def test_parse_from_sequence_example(self):
features = data_lib.parse_from_sequence_example(
ops.convert_to_tensor([
SEQ_EXAMPLE_PROTO_1.SerializeToString(),
SEQ_EXAMPLE_PROTO_2.SerializeToString(),
]),
list_size=2,
context_feature_spec=CONTEXT_FEATURE_SPEC,
example_feature_spec=EXAMPLE_FEATURE_SPEC)
with session.Session() as sess:
sess.run(variables.local_variables_initializer())
queue_runner.start_queue_runners()
feature_map = sess.run(features)
self.assertEqual(sorted(feature_map),
["query_length", "unigrams", "utility"])
self.assertAllEqual(feature_map["unigrams"].dense_shape, [2, 2, 3])
self.assertAllEqual(
feature_map["unigrams"].indices,
[[0, 0, 0], [0, 1, 0], [0, 1, 1], [0, 1, 2], [1, 0, 0]])
self.assertAllEqual(
feature_map["unigrams"].values,
[b"tensorflow", b"learning", b"to", b"rank", b"gbdt"])
self.assertAllEqual(feature_map["query_length"], [[3], [2]])
self.assertAllEqual(feature_map["utility"],
[[[0.], [1.]], [[0.], [0.]]])
def test_parse_from_sequence_example_with_sizes(self):
with tf.Graph().as_default():
features = data_lib.parse_from_sequence_example(
tf.convert_to_tensor(value=[
SEQ_EXAMPLE_PROTO_1.SerializeToString(),
SEQ_EXAMPLE_PROTO_2.SerializeToString(),
]),
context_feature_spec=CONTEXT_FEATURE_SPEC,
example_feature_spec=EXAMPLE_FEATURE_SPEC,
size_feature_name=_SIZE)
with tf.compat.v1.Session() as sess:
sess.run(tf.compat.v1.local_variables_initializer())
features = sess.run(features)
self.assertAllEqual(features[_SIZE], [2, 1])