Project import generated by Copybara.

PiperOrigin-RevId: 177753043

README.md

@@ -57,6 +57,7 @@ releasing a new version.
 |tensorflow-transform                                                            |tensorflow    |apache-beam[gcp]|
 |--------------------------------------------------------------------------------|--------------|----------------|
 |[GitHub master](https://github.com/tensorflow/transform/blob/master/RELEASE.md) |nightly (1.x) |2.2.0           |
+|[0.4.0](https://github.com/tensorflow/transform/blob/v0.4.0/RELEASE.md)         |1.4           |2.2.0           |
 |[0.3.1](https://github.com/tensorflow/transform/blob/v0.3.1/RELEASE.md)         |1.3           |2.1.1           |
 |[0.3.0](https://github.com/tensorflow/transform/blob/v0.3.0/RELEASE.md)         |1.3           |2.1.1           |
 |[0.1.10](https://github.com/tensorflow/transform/blob/v0.1.10/RELEASE.md)       |1.0           |2.0.0           |

RELEASE.md

@@ -1,11 +1,12 @@
-# Current version (not yet released; still in development)
+# Release 0.4.0
 
 ## Major Features and Improvements
 * Added a combine_analyzer() that supports user provided combiner, conforming to
   beam.CombinFn(). This allows users to implement custom combiners
   (e.g. median), to complement analyzers (like min, max) that are
   prepackaged in TFT.
-* Quantiles Analyzer (`tft.quantiles`).
+* Quantiles Analyzer (`tft.quantiles`), with a corresponding `tft.bucketize`
+  mapper.
 
 ## Bug Fixes and Other Changes
 * Depends on `apache-beam[gcp]>=2.2,<3`.
@@ -33,9 +34,15 @@
 * Some functions now introduce a new name scope when they did not before so the
   names of tensors may change.  This will only affect you if you directly lookup
   tensors by name in the graph produced by tf.Transform.
-* Various Analyzer Specs (_NumericCombineSpec, _UniquesSpec, _QuantilesSpec) are
-  now private. Analyzers are accessible only via the top-level TFT functions (
-  min, max, sum, size, mean, var, uniques, quantiles).
+* Various Analyzer Specs (\_NumericCombineSpec, \_UniquesSpec, \_QuantilesSpec)
+  are now private. Analyzers are accessible only via the top-level TFT functions
+  (min, max, sum, size, mean, var, uniques, quantiles).
+
+## Upcoming deprecations
+* The `serving_input_fn`s on `tensorflow_transform/saved/input_fn_maker.py` will
+be removed on a future version and should not be used on new code,
+see the `examples` directory for details on how to migrate your code to define
+their own serving functions.
 
 # Release 0.3.1
 

setup.py

@@ -17,7 +17,7 @@
 from setuptools import setup
 
 # Tensorflow transform version.
-__version__ = '0.4.0dev'
+__version__ = '0.4.0'
 
 
 def _make_required_install_packages():

tensorflow_transform/analyzers.py

@@ -30,6 +30,7 @@
 
 import re
 
+import numpy as np
 import tensorflow as tf
 
 
@@ -102,32 +103,102 @@ def output_is_asset(self, output_tensor):
     return self._output_is_asset_map[output_tensor]
 
 
-class _NumericCombineSpec(object):
-  """Operation to combine numeric values."""
+class CombinerSpec(object):
+  """Analyze using combiner function.
+
+  This object mirrors a beam.CombineFn, that will receive a beam PCollection
+  representing the batched input tensors.
+  """
+
+  def create_accumulator(self):
+    """Return a fresh, empty accumulator.
+
+    Returns: An empty accumulator.  This can be an Python value.
+    """
+    raise NotImplementedError
+
+  def add_input(self, accumulator, element):
+    """Return result of folding element into accumulator.
+
+    Args:
+      accumulator: the current accumulator
+      element: the element to add, which will be an ndarray representing the
+         value of the input for a batch.
+
+    Returns: An accumulator that includes the additional element.
+    """
+    raise NotImplementedError
+
+  def merge_accumulators(self, accumulators):
+    """Merges several accumulators to a single accumulator value.
+
+    Args:
+      accumulators: the accumulators to merge
+
+    Returns: The sole merged accumulator.
+    """
+    raise NotImplementedError
+
+  def extract_output(self, accumulator):
+    """Return result of converting accumulator into the output value.
+
+    Args:
+      accumulator: the final accumulator value.  Should be a list of ndarrays.
+
+    Returns: An ndarray representing the result of this combiner.
+    """
+    raise NotImplementedError
 
-  MIN = 'min'
-  MAX = 'max'
-  SUM = 'sum'
 
-  def __init__(self, dtype, combiner_type, reduce_instance_dims):
-    self._dtype = dtype
-    self._combiner_type = combiner_type
+def combine_analyzer(x, output_dtype, output_shape, combiner_spec, name):
+  """Applies the combiner over the whole dataset.
+
+  Args:
+    x: An input `Tensor` or `SparseTensor`.
+    output_dtype: The dtype of the output of the analyzer.
+    output_shape: The shape of the output of the analyzer.
+    combiner_spec: A subclass of CombinerSpec.
+    name: Similar to a TF op name.  Used to define a unique scope for this
+      analyzer, which can be used for debugging info.
+
+  Returns:
+    The combined values, which is a `Tensor` with type output_dtype and shape
+    `output_shape`.  These must be compatible with the combiner_spec.
+  """
+  return Analyzer([x], [(output_dtype, output_shape, False)], combiner_spec,
+                  name).outputs[0]
+
+
+class _NumPyCombinerSpec(CombinerSpec):
+  """Combines the PCollection only on the 0th dimension using nparray."""
+
+  def __init__(self, fn, reduce_instance_dims):
+    self._fn = fn
     self._reduce_instance_dims = reduce_instance_dims
 
-  @property
-  def dtype(self):
-    return self._dtype
+  def create_accumulator(self):
+    return None
 
-  @property
-  def combiner_type(self):
-    return self._combiner_type
+  def add_input(self, accumulator, next_input):
+    if self._reduce_instance_dims:
+      batch = self._fn(next_input)
+    else:
+      batch = self._fn(next_input, axis=0)
+    if accumulator is None:
+      return batch
+    else:
+      return self._fn((accumulator, batch), axis=0)
 
-  @property
-  def reduce_instance_dims(self):
-    return self._reduce_instance_dims
+  def merge_accumulators(self, accumulators):
+    # numpy's sum, min, max, etc functions operate on array-like objects, but
+    # not arbitrary iterables. Convert the provided accumulators into a list
+    return self._fn(list(accumulators), axis=0)
+
+  def extract_output(self, accumulator):
+    return [accumulator]
 
 
-def _numeric_combine(x, combiner_type, reduce_instance_dims=True, name=None):
+def _numeric_combine(x, fn, reduce_instance_dims=True, name=None):
   """Apply an analyzer with _NumericCombineSpec to given input."""
   if not isinstance(x, tf.Tensor):
     raise TypeError('Expected a Tensor, but got %r' % x)
@@ -143,10 +214,9 @@ def _numeric_combine(x, combiner_type, reduce_instance_dims=True, name=None):
     # If reducing over batch dimensions, with unknown shape, the result will
     # also have unknown shape.
     shape = None
-  spec = _NumericCombineSpec(x.dtype, combiner_type, reduce_instance_dims)
-  return Analyzer(
-      [x], [(x.dtype, shape, False)], spec,
-      name if name is not None else combiner_type).outputs[0]
+  return combine_analyzer(
+      x, x.dtype, shape, _NumPyCombinerSpec(fn, reduce_instance_dims),
+      name if name is not None else fn.__name__)
 
 
 def min(x, reduce_instance_dims=True, name=None):  # pylint: disable=redefined-builtin
@@ -162,8 +232,7 @@ def min(x, reduce_instance_dims=True, name=None):  # pylint: disable=redefined-b
   Returns:
     A `Tensor`. Has the same type as `x`.
   """
-  return _numeric_combine(
-      x, _NumericCombineSpec.MIN, reduce_instance_dims, name)
+  return _numeric_combine(x, np.min, reduce_instance_dims, name)
 
 
 def max(x, reduce_instance_dims=True, name=None):  # pylint: disable=redefined-builtin
@@ -179,8 +248,7 @@ def max(x, reduce_instance_dims=True, name=None):  # pylint: disable=redefined-b
   Returns:
     A `Tensor`. Has the same type as `x`.
   """
-  return _numeric_combine(
-      x, _NumericCombineSpec.MAX, reduce_instance_dims, name)
+  return _numeric_combine(x, np.max, reduce_instance_dims, name)
 
 
 def sum(x, reduce_instance_dims=True, name=None):  # pylint: disable=redefined-builtin
@@ -196,8 +264,7 @@ def sum(x, reduce_instance_dims=True, name=None):  # pylint: disable=redefined-b
   Returns:
     A `Tensor`. Has the same type as `x`.
   """
-  return _numeric_combine(
-      x, _NumericCombineSpec.SUM, reduce_instance_dims, name)
+  return _numeric_combine(x, np.sum, reduce_instance_dims, name)
 
 
 def size(x, reduce_instance_dims=True, name=None):
@@ -271,18 +338,13 @@ def var(x, reduce_instance_dims=True, name=None):
 class _UniquesSpec(object):
   """Operation to compute unique values."""
 
-  def __init__(self, dtype, top_k, frequency_threshold,
+  def __init__(self, top_k, frequency_threshold,
                vocab_filename, store_frequency):
-    self._dtype = dtype
     self._top_k = top_k
     self._frequency_threshold = frequency_threshold
     self._vocab_filename = vocab_filename
     self._store_frequency = store_frequency
 
-  @property
-  def dtype(self):
-    return self._dtype
-
   @property
   def top_k(self):
     return self._top_k
@@ -400,8 +462,8 @@ def uniques(x, top_k=None, frequency_threshold=None,
     # Make the file name path safe.
     vocab_filename = sanitized_vocab_filename(vocab_filename, prefix=prefix)
 
-    spec = _UniquesSpec(tf.string, top_k, frequency_threshold,
-                        vocab_filename, store_frequency)
+    spec = _UniquesSpec(top_k, frequency_threshold, vocab_filename,
+                        store_frequency)
     return Analyzer([x], [(tf.string, [], True)], spec, 'uniques').outputs[0]
 
 
@@ -469,50 +531,3 @@ def quantiles(x, num_buckets, epsilon, name=None):
     # Drop the fist and last quantile boundaries, so that we end-up with
     # num_buckets-1 boundaries, and hence num_buckets buckets.
     return quantile_boundaries[0:1, 1:-1]
-
-
-class _CombinerSpec(object):
-  """Analyze using combiner function.
-
-  Args:
-    combiner: Object of a class that implements beam.CombineFn() interface.
-      In addtion, the combiner class must implement a @property method called
-      output_dtype() that returns the tf.DType of the output of the combiner.
-  """
-
-  def __init__(self, combiner):
-    self._combiner = combiner
-
-  @property
-  def combiner(self):
-    return self._combiner
-
-  @property
-  def output_dtype(self):
-    return self._combiner.output_dtype
-
-
-def combine_analyzer(x, combiner, name=None):
-  """Applies the combiner over the whole dataset.
-
-  Args:
-    x: An input `Tensor` or `SparseTensor`.
-    combiner: Object of a class that implements beam.CombineFn() interface.
-      In addtion, the combiner class must implement a @property method called
-      output_dtype() that returns the type of the output of the combiner.
-    name: (Optional) A name for this operation.
-
-  Returns:
-    The combined values as a list, where the each element in the list
-    is of type combiner.output_dtype().
-  """
-
-  # The TF node name will be of the form:
-  # original_scope/{combine_analyzer|name}/{class-name-of-combiner}
-  with tf.name_scope(name, 'combine_analyzer'):
-    spec = _CombinerSpec(combiner)
-    return Analyzer(
-        [x],
-        [(spec.output_dtype, [1, None], False)],
-        spec,
-        type(combiner).__name__).outputs[0]