Get Started with TensorFlow Transform

导语：
    本文是 tensorflow 手册翻译系列的第二十九篇。
    本文档介绍tf.Transform的基本概念以及如何使用它们。

import tensorflow as tf
import tensorflow_transform as tft
import tensorflow_transform.beam as tft_beam

def preprocessing_fn(inputs):
  x = inputs['x']
  y = inputs['y']
  s = inputs['s']
  x_centered = x - tft.mean(x)
  y_normalized = tft.scale_to_0_1(y)
  s_integerized = tft.compute_and_apply_vocabulary(s)
  x_centered_times_y_normalized = x_centered * y_normalized
  return {
      'x_centered': x_centered,
      'y_normalized': y_normalized,
      'x_centered_times_y_normalized': x_centered_times_y_normalized,
      's_integerized': s_integerized
  }

raw_data = [
    {'x': 1, 'y': 1, 's': 'hello'},
    {'x': 2, 'y': 2, 's': 'world'},
    {'x': 3, 'y': 3, 's': 'hello'}
]

raw_data_metadata = ...
transformed_dataset, transform_fn = (
    (raw_data, raw_data_metadata) | tft_beam.AnalyzeAndTransformDataset(
        preprocessing_fn))
transformed_data, transformed_metadata = transformed_dataset

[{u's_integerized': 0,
  u'x_centered': -1.0,
  u'x_centered_times_y_normalized': -0.0,
  u'y_normalized': 0.0},
 {u's_integerized': 1,
  u'x_centered': 0.0,
  u'x_centered_times_y_normalized': 0.0,
  u'y_normalized': 0.5},
 {u's_integerized': 0,
  u'x_centered': 1.0,
  u'x_centered_times_y_normalized': 1.0,
  u'y_normalized': 1.0}]

transformed_data, transform_fn = (
    my_data | tft_beam.AnalyzeAndTransformDataset(preprocessing_fn))

transform_fn = my_data | tft_beam.AnalyzeDataset(preprocessing_fn)
transformed_data = (my_data, transform_fn) | tft_beam.TransformDataset()

from tensorflow_transform.tf_metadata import dataset_metadata
from tensorflow_transform.tf_metadata import dataset_schema

raw_data_metadata = dataset_metadata.DatasetMetadata(
    dataset_schema.from_feature_spec({
        's': tf.FixedLenFeature([], tf.string),
        'y': tf.FixedLenFeature([], tf.float32),
        'x': tf.FixedLenFeature([], tf.float32),
    }))

39, State-gov, 77516, Bachelors, 13, Never-married, Adm-clerical, Not-in-family, White, Male, 2174, 0, 40, United-States, <=50K
50, Self-emp-not-inc, 83311, Bachelors, 13, Married-civ-spouse, Exec-managerial, Husband, White, Male, 0, 0, 13, United-States, <=50K

converter = tft.coders.CsvCoder(ordered_columns, raw_data_schema)

raw_data = (
    p
    | 'ReadTrainData' >> textio.ReadFromText(train_data_file)
    | ...
    | 'DecodeTrainData' >> beam.Map(converter.decode))

def preprocessing_fn(inputs):
  """Preprocess input columns into transformed columns."""
  # Since we are modifying some features and leaving others unchanged, we
  # start by setting `outputs` to a copy of `inputs.
  outputs = inputs.copy()

  # Scale numeric columns to have range [0, 1].
  for key in NUMERIC_FEATURE_KEYS:
    outputs[key] = tft.scale_to_0_1(outputs[key])

  for key in OPTIONAL_NUMERIC_FEATURE_KEYS:
    # This is a SparseTensor because it is optional. Here we fill in a default
    # value when it is missing.
    dense = tf.sparse_to_dense(outputs[key].indices,
                               [outputs[key].dense_shape[0], 1],
                               outputs[key].values, default_value=0.)
    # Reshaping from a batch of vectors of size 1 to a batch to scalars.
    dense = tf.squeeze(dense, axis=1)
    outputs[key] = tft.scale_to_0_1(dense)

  # For all categorical columns except the label column, we generate a
  # vocabulary but do not modify the feature.  This vocabulary is instead
  # used in the trainer, by means of a feature column, to convert the feature
  # from a string to an integer id.
  for key in CATEGORICAL_FEATURE_KEYS:
    tft.vocabulary(inputs[key], vocab_filename=key)

  # For the label column we provide the mapping from string to index.
  initializer = tf.lookup.KeyValueTensorInitializer(
      keys=['>50K', '<=50K'],
      values=tf.cast(tf.range(2), tf.int64),
      key_dtype=tf.string,
      value_dtype=tf.int64)
  table = tf.lookup.StaticHashTable(initializer, default_value=-1)

  outputs[LABEL_KEY] = table.lookup(outputs[LABEL_KEY])

  return outputs

transformed_data | "WriteTrainData" >> tfrecordio.WriteToTFRecord(
    transformed_eval_data_base,
    coder=tft.coders.ExampleProtoCoder(transformed_metadata))

_ = (
    transform_fn
    | 'WriteTransformFn' >> tft_beam.WriteTransformFn(working_dir))
transformed_metadata | 'WriteMetadata' >> tft_beam.WriteMetadata(
    transformed_metadata_file, pipeline=p)

  wget https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.data
  wget https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.test

real_valued_columns = [feature_column.real_valued_column(key)
                       for key in NUMERIC_COLUMNS]

one_hot_columns = [
    feature_column.sparse_column_with_integerized_feature(
        key, bucket_size=bucket_size)
    for key, bucket_size in zip(CATEGORICAL_COLUMNS, BUCKET_SIZES)]

estimator = learn.LinearClassifier(real_valued_columns + one_hot_columns)

def _make_training_input_fn(tf_transform_output, transformed_examples,
                            batch_size):
  ...
  def input_fn():
    """Input function for training and eval."""
    dataset = tf.contrib.data.make_batched_features_dataset(
        ..., tf_transform_output.transformed_feature_spec(), ...)

    transformed_features = dataset.make_one_shot_iterator().get_next()
    ...

  return input_fn