Project import generated by Copybara.

PiperOrigin-RevId: 236960625

README.md

@@ -10,6 +10,7 @@ The metadata serialization formats include:
 
 * A schema describing tabular data (e.g., tf.Examples).
 * A collection of summary statistics over such datasets.
+* A problem statement quantifying the objectives of a model.
 
 The metadata may be produced by hand or automatically during input data
 analysis, and may be consumed for data validation, exploration, and

RELEASE.md

@@ -1,8 +1,19 @@
+# Version 0.13.0
+
+## Major Features and Improvements
+
+* Added ProblemStatement.
+
+## Bug Fixes and Other Changes
+
+## Breaking changes
+
+## Deprecations
+
 # Version 0.12.0
 
 ## Major Features and Improvements
 
-* Added ProblemStatement
 * Add support for declaring sparse features.
 * Add support for schema diff regions.
 

setup.py

@@ -16,8 +16,13 @@
 from setuptools import find_packages
 from setuptools import setup
 
+with open('tensorflow_metadata/version.py') as fp:
+  globals_dict = {}
+  exec (fp.read(), globals_dict)  # pylint: disable=exec-used
+
 # tf.Metadata version.
-__version__ = '0.12.0'
+__version__ = globals_dict['__version__']
+
 
 # Note: In order for the README to be rendered correctly, make sure to have the
 # following minimum required versions of the respective packages when building
@@ -58,8 +63,7 @@
     install_requires=[
         'googleapis-common-protos',
 
-        # TF now requires protobuf>=3.6.1.
-        'protobuf>=3.6.1,<4',
+        'protobuf>=3.7,<4',
     ],
     python_requires='>=2.7,<4',
     packages=find_packages(),
@@ -68,5 +72,5 @@
     long_description=_LONG_DESCRIPTION,
     long_description_content_type='text/markdown',
     keywords='tensorflow metadata tfx',
-    download_url='https://pypi.org/project/tensorflow-metadata',
+    download_url='https://github.com/tensorflow/metadata/tags',
     requires=[])

tensorflow_metadata/BUILD

@@ -0,0 +1,9 @@
+licenses(["notice"])  # Apache 2.0
+
+sh_binary(
+    name = "build_pip_package",
+    srcs = ["build_pip_package.sh"],
+    data = [
+        "//tensorflow_metadata/proto/v0:metadata_v0_proto_py_pb2",
+    ],
+)

tensorflow_metadata/build_pip_package.sh

@@ -0,0 +1,54 @@
+#!/bin/bash
+# Copyright 2018 Google LLC
+#
+# 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.
+
+# Convenience binary to build TFMD from source.
+
+# Put wrapped c++ files in place
+
+# Usage: build_pip_package.sh [--python_bin_path PYTHON_BIN_PATH]
+
+if [[ -z "$1" ]]; then
+  PYTHON_BIN_PATH=python
+else
+  if [[ "$1" == --python_bin_path ]]; then
+    shift
+    PYTHON_BIN_PATH=$1
+  else
+    printf "Unrecognized argument $1"
+    exit 1
+  fi
+fi
+
+set -u -x
+
+cp -f tensorflow_metadata/proto/v0/schema_pb2.py \
+  ${BUILD_WORKSPACE_DIRECTORY}/tensorflow_metadata/proto/v0
+
+cp -f tensorflow_metadata/proto/v0/path_pb2.py \
+  ${BUILD_WORKSPACE_DIRECTORY}/tensorflow_metadata/proto/v0
+
+cp -f tensorflow_metadata/proto/v0/anomalies_pb2.py \
+  ${BUILD_WORKSPACE_DIRECTORY}/tensorflow_metadata/proto/v0
+
+cp -f tensorflow_metadata/proto/v0/statistics_pb2.py \
+  ${BUILD_WORKSPACE_DIRECTORY}/tensorflow_metadata/proto/v0
+
+# Create the wheel
+cd ${BUILD_WORKSPACE_DIRECTORY}
+
+${PYTHON_BIN_PATH} setup.py bdist_wheel
+
+# Cleanup
+cd -

tensorflow_metadata/proto/v0/BUILD

@@ -24,7 +24,9 @@ cc_proto_library(
     name = "metadata_v0_proto_cc_pb2",
     srcs = [
         "anomalies.proto",
+        "metric.proto",
         "path.proto",
+        "problem_statement.proto",
         "schema.proto",
         "statistics.proto",
     ],
@@ -37,7 +39,9 @@ py_proto_library(
     name = "metadata_v0_proto_py_pb2",
     srcs = [
         "anomalies.proto",
+        "metric.proto",
         "path.proto",
+        "problem_statement.proto",
         "schema.proto",
         "statistics.proto",
     ],

tensorflow_metadata/proto/v0/metric.proto

@@ -0,0 +1,184 @@
+// Copyright 2017 The TensorFlow Authors. All Rights Reserved.
+//
+// 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.
+// =============================================================================
+
+syntax = "proto3";
+
+option cc_enable_arenas = true;
+
+package tensorflow.metadata.v0;
+
+option java_package = "org.tensorflow.metadata.v0";
+option java_multiple_files = true;
+
+
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/binary_accuracy
+message BinaryAccuracy {
+}
+
+
+// categorical_accuracy(...)
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/categorical_accuracy
+message CategoricalAccuracy {
+}
+
+// categorical_crossentropy(...)
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/categorical_crossentropy
+message CategoricalCrossEntropy {
+}
+
+// cosine(...)
+// cosine_proximity(...)
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/cosine_proximity
+message Cosine {
+}
+
+
+// Linear Hinge Loss
+// hinge(...)
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/hinge
+message Hinge {
+}
+
+// kld(...)
+// kullback_leibler_divergence(...)
+// KLD(...)
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/kullback_leibler_divergence
+message KullbackLeiblerDivergence {
+}
+
+// MAE(...)
+// mae(...)
+// mean_absolute_error(...)
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/mean_absolute_error
+message MeanAbsoluteError {
+}
+
+// MAPE(...)
+// mape(...)
+// mean_absolute_percentage_error(...)
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/mean_absolute_percentage_error
+message MeanAbsolutePercentageError {
+}
+
+// MSE(...)
+// mse(...)
+// mean_squared_error(...)
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/mean_squared_error
+message MeanSquaredError {
+}
+
+// msle(...)
+// MSLE(...)
+// mean_squared_logarithmic_error(...)
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/mean_squared_logarithmic_error
+message MeanSquaredLogarithmicError {
+}
+
+// poisson(...)
+message Poisson {
+}
+
+// squared_hinge(...)
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/squared_hinge
+message SquaredHinge {
+}
+
+// top_k_categorical_accuracy(...)
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/top_k_categorical_accuracy
+message TopKCategoricalAccuracy {
+
+}
+
+// sparse_top_k_categorical_accuracy(...)
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/sparse_top_k_categorical_accuracy
+message SparseTopKCategoricalAccuracy {
+}
+
+// Binary cross entropy as a metric is equal to the negative log likelihood
+// (see logistic regression).
+// In addition, when used to solve a binary classification task, binary cross
+// entropy implies that the binary label will maximize binary accuracy.
+// binary_crossentropy(...)
+// https://www.tensorflow.org/api_docs/python/tf/keras/metrics/binary_crossentropy
+message BinaryCrossEntropy {
+}
+
+// AKA the negative log likelihood or log loss.
+// Given a label y\in {0,1} and a predicted probability p in [0,1]:
+// -yln(p)-(1-y)ln(1-p)
+// TODO(martinz): if this is interpreted the same as binary_cross_entropy,
+// we may need to revisit the semantics.
+message LogisticRegression {
+}
+
+
+message AUC {
+}
+
+message PrecisionAtK {
+
+}
+
+message MeanReciprocalRank {
+}
+
+message BlockUtility {
+  repeated double weight = 1;
+}
+
+// A custom metric.
+message CustomMetric {
+  // The name of a metric computed by the model.
+  string name = 1;
+}
+
+// Objective functions are performance metrics that are differentiable, and
+// therefore can be directly optimized.
+// TODO(martinz): take metrics that are appropriate and make them objectives.
+message ObjectiveFunction {
+  oneof objective {
+    BinaryCrossEntropy binary_cross_entropy = 1;
+    LogisticRegression logistic_regression = 2;
+    MeanSquaredError squared_error = 3;
+    CustomMetric custom_metric = 4;
+  }
+}
+
+// Performance metrics measure the quality of a model. They need not be
+// differentiable.
+message PerformanceMetric {
+  oneof performance_metric {
+    AUC auc = 1;
+    BinaryAccuracy binary_accuracy = 2;
+    BinaryCrossEntropy binary_cross_entropy = 3;
+    BlockUtility block_utility = 4;
+    CategoricalAccuracy categorical_accuracy = 5;
+    CategoricalCrossEntropy categorical_cross_entropy = 6;
+    Cosine cosine = 7;
+    Hinge hinge = 8;
+    KullbackLeiblerDivergence kullback_leibler_divergence = 9;
+    LogisticRegression logistic_regression = 10;
+    MeanAbsoluteError mean_absolute_error = 11;
+    MeanAbsolutePercentageError mean_absolute_percentage_error = 12;
+    MeanSquaredError squared_error = 13;
+    MeanSquaredLogarithmicError mean_squared_logarithmic_error = 14;
+    MeanReciprocalRank mean_reciprocal_rank = 15;
+    Poisson poisson = 16;
+    PrecisionAtK precision_at_k = 17;
+    SquaredHinge squared_hinge = 18;
+    SparseTopKCategoricalAccuracy sparse_top_k_categorical_accuracy = 19;
+    TopKCategoricalAccuracy top_k_categorical_accuracy = 20;
+  }
+}