TFDV update for OSS build.

PiperOrigin-RevId: 762730883

Author: genehwung

BUILD

@@ -0,0 +1,18 @@
+load("@bazel_gazelle//:def.bzl", "gazelle")
+
+package(
+ default_visibility = [":__subpackages__"],
+)
+
+licenses(["notice"])
+
+exports_files(["LICENSE"])
+
+gazelle(
+ name = "gazelle-update-repos",
+ args = [
+ "-from_file=go.mod",
+ "-to_macro=deps.bzl%go_dependencies",
+ ],
+ command = "update-repos",
+)

RELEASE.md

@@ -20,7 +20,7 @@
 
 ## Bug Fixes and Other Changes
 
-* Relax dependency on Protobuf to include version 5.x
+* Relax dependency on Protobuf to include version 6.5
 
 ## Known Issues
 

WORKSPACE

@@ -2,36 +2,55 @@ workspace(name = "tensorflow_data_validation")
 
 load("@bazel_tools//tools/build_defs/repo:http.bzl", "http_archive")
 
+http_archive(
+ name = "google_bazel_common",
+ sha256 = "82a49fb27c01ad184db948747733159022f9464fc2e62da996fa700594d9ea42",
+ strip_prefix = "bazel-common-2a6b6406e12208e02b2060df0631fb30919080f3",
+ urls = ["https://github.com/google/bazel-common/archive/2a6b6406e12208e02b2060df0631fb30919080f3.zip"],
+)
+
+################################################################################
+# Generic Bazel Support #
+################################################################################
+
+http_archive(
+ name = "rules_proto",
+ sha256 = "6fb6767d1bef535310547e03247f7518b03487740c11b6c6adb7952033fe1295",
+ strip_prefix = "rules_proto-6.0.2",
+ url = "https://github.com/bazelbuild/rules_proto/releases/download/6.0.2/rules_proto-6.0.2.tar.gz",
+)
+
+load("@rules_proto//proto:repositories.bzl", "rules_proto_dependencies")
+
+rules_proto_dependencies()
+
+load("@rules_proto//proto:setup.bzl", "rules_proto_setup")
+
+rules_proto_setup()
+
+load("@rules_proto//proto:toolchains.bzl", "rules_proto_toolchains")
+
+rules_proto_toolchains()
+
 # Install version 0.9.0 of rules_foreign_cc, as default version causes an
 # invalid escape sequence error to be raised, which can't be avoided with
 # the --incompatible_restrict_string_escapes=false flag (flag was removed in
 # Bazel 5.0).
 RULES_FOREIGN_CC_VERSION = "0.9.0"
+
 http_archive(
  name = "rules_foreign_cc",
+ patch_tool = "patch",
+ patches = ["//third_party:rules_foreign_cc.patch"],
  sha256 = "2a4d07cd64b0719b39a7c12218a3e507672b82a97b98c6a89d38565894cf7c51",
  strip_prefix = "rules_foreign_cc-%s" % RULES_FOREIGN_CC_VERSION,
  url = "https://github.com/bazelbuild/rules_foreign_cc/archive/refs/tags/%s.tar.gz" % RULES_FOREIGN_CC_VERSION,
- patch_tool = "patch",
- patches = ["//third_party:rules_foreign_cc.patch",],
 )
 
 load("@rules_foreign_cc//foreign_cc:repositories.bzl", "rules_foreign_cc_dependencies")
-rules_foreign_cc_dependencies()
 
-# To update package to a new revision, for example ABSL
-# 1. Update the 'COM_GOOGLE_ABSL_COMMIT' var below to include the new git hash.
-# 2. Get the sha256 hash of the archive with a command such as...
-# curl -L https://github.com/abseil/abseil-cpp/archive/<git hash>.tar.gz | sha256sum
-# and update the 'sha256' arg with the result.
-# 3. Request the new archive to be mirrored on mirror.bazel.build for more
-# reliable downloads.
+rules_foreign_cc_dependencies()
 
-# TODO(caveness): Clean up dependencies that were needed by TF but are no longer
-# required.
-# Needed by tf_py_wrap_cc rule from Tensorflow.
-# When upgrading tensorflow version, also check tensorflow/WORKSPACE for the
-# version of this -- keep in sync.
 http_archive(
  name = "bazel_skylib",
  sha256 = "97e70364e9249702246c0e9444bccdc4b847bed1eb03c5a3ece4f83dfe6abc44",
@@ -41,77 +60,177 @@ http_archive(
  ],
 )
 
-# TensorFlow depends on "io_bazel_rules_closure" so we need this here.
-# Needs to be kept in sync with the same target in TensorFlow's WORKSPACE file.
-# LINT.IfChange(io_bazel_rules_clousure)
+_PROTOBUF_COMMIT = "4.25.6" # 4.25.6
+
 http_archive(
- name = "io_bazel_rules_closure",
- sha256 = "5b00383d08dd71f28503736db0500b6fb4dda47489ff5fc6bed42557c07c6ba9",
- strip_prefix = "rules_closure-308b05b2419edb5c8ee0471b67a40403df940149",
+ name = "com_google_protobuf",
+ sha256 = "ff6e9c3db65f985461d200c96c771328b6186ee0b10bc7cb2bbc87cf02ebd864",
+ strip_prefix = "protobuf-%s" % _PROTOBUF_COMMIT,
  urls = [
- "http://mirror.tensorflow.org/github.com/bazelbuild/rules_closure/archive/308b05b2419edb5c8ee0471b67a40403df940149.tar.gz",
- "https://github.com/bazelbuild/rules_closure/archive/308b05b2419edb5c8ee0471b67a40403df940149.tar.gz", # 2020-02-14
+ "https://github.com/protocolbuffers/protobuf/archive/v4.25.6.zip",
  ],
 )
-# LINT.ThenChange(:tf_commit)
 
-# External proto rules.
+# Needed by abseil-py by zetasql.
 http_archive(
- name = "rules_proto",
- sha256 = "66bfdf8782796239d3875d37e7de19b1d94301e8972b3cbd2446b332429b4df1",
- strip_prefix = "rules_proto-4.0.0",
+ name = "six_archive",
+ build_file = "//third_party:six.BUILD",
+ sha256 = "105f8d68616f8248e24bf0e9372ef04d3cc10104f1980f54d57b2ce73a5ad56a",
+ strip_prefix = "six-1.10.0",
  urls = [
- "https://mirror.bazel.build/github.com/bazelbuild/rules_proto/archive/refs/tags/4.0.0.tar.gz",
- "https://github.com/bazelbuild/rules_proto/archive/refs/tags/4.0.0.tar.gz",
+ "http://mirror.bazel.build/pypi.python.org/packages/source/s/six/six-1.10.0.tar.gz",
+ "https://pypi.python.org/packages/source/s/six/six-1.10.0.tar.gz",
  ],
 )
-load("@rules_proto//proto:repositories.bzl", "rules_proto_dependencies", "rules_proto_toolchains")
-rules_proto_dependencies()
-rules_proto_toolchains()
 
-# TODO(b/239095455): Change to using a tfx-bsl workspace macro to load these
-# dependencies.
-# Needed by zetasql.
-_PROTOBUF_COMMIT = "3.21.9"
+load("@com_google_protobuf//:protobuf_deps.bzl", "protobuf_deps")
+
+protobuf_deps()
+
+# Use the last commit on the relevant release branch to update.
+# LINT.IfChange(arrow_archive_version)
+ARROW_COMMIT = "347a88ff9d20e2a4061eec0b455b8ea1aa8335dc" # 6.0.1
+# LINT.ThenChange(third_party/arrow.BUILD:arrow_gen_version)
 
+# `shasum -a 256` can be used to get `sha256` from the downloaded archive on
+# Linux.
 http_archive(
- name = "com_google_protobuf",
- sha256 = "f66073dee0bc159157b0bd7f502d7d1ee0bc76b3c1eac9836927511bdc4b3fc1",
- strip_prefix = "protobuf-%s" % _PROTOBUF_COMMIT,
+ name = "arrow",
+ build_file = "//third_party:arrow.BUILD",
+ patches = ["//third_party:arrow.patch"],
+ sha256 = "55fc466d0043c4cce0756bc18e1e62b3233be74c9afe8dc0d18420b9a5fd9714",
+ strip_prefix = "arrow-%s" % ARROW_COMMIT,
+ urls = ["https://github.com/apache/arrow/archive/%s.zip" % ARROW_COMMIT],
+)
+
+COM_GOOGLE_ABSL_COMMIT = "4447c7562e3bc702ade25105912dce503f0c4010" # lts_2023_08_0
+
+http_archive(
+ name = "com_google_absl",
+ sha256 = "df8b3e0da03567badd9440377810c39a38ab3346fa89df077bb52e68e4d61e74",
+ strip_prefix = "abseil-cpp-%s" % COM_GOOGLE_ABSL_COMMIT,
+ url = "https://github.com/abseil/abseil-cpp/archive/%s.tar.gz" % COM_GOOGLE_ABSL_COMMIT,
+)
+
+# Will be loaded by workspace.bzl from head
+# TFMD_COMMIT = "404805761e614561cceedc429e67c357c62be26d" # 1.17.1
+
+# http_archive(
+# name = "com_tensorflow_metadata",
+# sha256 = "1b72e0e5085812cd9b19e004a381b544542f9545a081f0f738c5ed6b8bb886a2",
+# strip_prefix = "metadata-%s" % TFMD_COMMIT,
+# urls = ["https://github.com/tensorflow/metadata/archive/%s.zip" % TFMD_COMMIT],
+# )
+
+# TODO(b/177694034): Follow the new format for tensorflow import after TF 2.5.
+#here
+TENSORFLOW_COMMIT = "3c92ac03cab816044f7b18a86eb86aa01a294d95" # 2.17.1
+
+http_archive(
+ name = "org_tensorflow_no_deps",
+ patches = [
+ "//third_party:tensorflow_expose_example_proto.patch",
+ ],
+ sha256 = "317dd95c4830a408b14f3e802698eb68d70d81c7c7cfcd3d28b0ba023fe84a68",
+ strip_prefix = "tensorflow-%s" % TENSORFLOW_COMMIT,
  urls = [
- "https://github.com/protocolbuffers/protobuf/archive/v3.21.9.zip"
+ "https://github.com/tensorflow/tensorflow/archive/%s.tar.gz" % TENSORFLOW_COMMIT,
  ],
 )
 
-load("@com_google_protobuf//:protobuf_deps.bzl", "protobuf_deps")
-protobuf_deps()
+PYBIND11_COMMIT = "8a099e44b3d5f85b20f05828d919d2332a8de841" # 2.11.1
 
-# Needed by abseil-py by zetasql.
 http_archive(
- name = "six_archive",
+ name = "pybind11",
+ build_file = "//third_party:pybind11.BUILD",
+ sha256 = "8f4b7f28d214e36301435c055076c36186388dc9617117802cba8a059347cb00",
+ strip_prefix = "pybind11-%s" % PYBIND11_COMMIT,
+ urls = ["https://github.com/pybind/pybind11/archive/%s.zip" % PYBIND11_COMMIT],
+)
+
+load("//third_party:python_configure.bzl", "local_python_configure")
+
+local_python_configure(name = "local_config_python")
+
+http_archive(
+ name = "com_google_farmhash",
+ build_file = "//third_party:farmhash.BUILD",
+ sha256 = "6560547c63e4af82b0f202cb710ceabb3f21347a4b996db565a411da5b17aba0", # SHARED_FARMHASH_SHA
+ strip_prefix = "farmhash-816a4ae622e964763ca0862d9dbd19324a1eaf45",
  urls = [
- "http://mirror.bazel.build/pypi.python.org/packages/source/s/six/six-1.10.0.tar.gz",
- "https://pypi.python.org/packages/source/s/six/six-1.10.0.tar.gz",
+ "https://github.com/google/farmhash/archive/816a4ae622e964763ca0862d9dbd19324a1eaf45.tar.gz",
  ],
- sha256 = "105f8d68616f8248e24bf0e9372ef04d3cc10104f1980f54d57b2ce73a5ad56a",
- strip_prefix = "six-1.10.0",
- build_file = "//third_party:six.BUILD"
 )
 
-COM_GOOGLE_ABSL_COMMIT = "92fdbfb301f8b301b28ab5c99e7361e775c2fb8a" # 2022-08-25 Abseil Logging library first release
+################################################################################
+# Google APIs protos #
+################################################################################
+http_archive(
+ name = "com_google_googleapis",
+ patch_args = ["-p1"],
+ patches = ["//third_party:googleapis.patch"],
+ sha256 = "28e7fe3a640dd1f47622a4c263c40d5509c008cc20f97bd366076d5546cccb64",
+ strip_prefix = "googleapis-4ce00b00904a7ce1df8c157e54fcbf96fda0dc49",
+ url = "https://github.com/googleapis/googleapis/archive/4ce00b00904a7ce1df8c157e54fcbf96fda0dc49.tar.gz",
+)
+
+load("@com_google_googleapis//:repository_rules.bzl", "switched_rules_by_language")
+
+switched_rules_by_language(
+ name = "com_google_googleapis_imports",
+ cc = True,
+ go = True,
+)
+
+###############################################################################
+# Gazelle Support #
+###############################################################################
+
+_rules_go_version = "v0.48.1"
+
 http_archive(
- name = "com_google_absl",
- url = "https://github.com/abseil/abseil-cpp/archive/%s.tar.gz" % COM_GOOGLE_ABSL_COMMIT,
- sha256 = "71d38c5f44997a5ccbc338f904c8682b40c25cad60b9cbaf27087a917228d5fa",
- strip_prefix = "abseil-cpp-%s" % COM_GOOGLE_ABSL_COMMIT
+ name = "io_bazel_rules_go",
+ sha256 = "b2038e2de2cace18f032249cb4bb0048abf583a36369fa98f687af1b3f880b26",
+ urls = [
+ "https://mirror.bazel.build/github.com/bazelbuild/rules_go/releases/download/{0}/rules_go-{0}.zip".format(_rules_go_version),
+ "https://github.com/bazelbuild/rules_go/releases/download/{0}/rules_go-{0}.zip.format(_rules_go_version)",
+ ],
 )
 
-ZETASQL_COMMIT = "ac37cf5c0d80b5605176fc0f29e87b12f00be693" # 08/10/2022
+load("@io_bazel_rules_go//go:deps.bzl", "go_register_toolchains", "go_rules_dependencies")
+
+go_rules_dependencies()
+
+go_register_toolchains(version = "1.21.11")
+
+_bazel_gazelle_version = "0.36.0"
+
+http_archive(
+ name = "bazel_gazelle",
+ sha256 = "75df288c4b31c81eb50f51e2e14f4763cb7548daae126817247064637fd9ea62",
+ urls = [
+ "https://mirror.bazel.build/github.com/bazelbuild/bazel-gazelle/releases/download/v{0}/bazel-gazelle-v{0}.tar.gz".format(_bazel_gazelle_version),
+ "https://github.com/bazelbuild/bazel-gazelle/releases/download/v{0}/bazel-gazelle-v{0}.tar.gz".format(_bazel_gazelle_version),
+ ],
+)
+
+load("@bazel_gazelle//:deps.bzl", "gazelle_dependencies") #, "go_repository")
+
+gazelle_dependencies()
+
+################################################################################
+# ZetaSQL #
+################################################################################
+
+ZETASQL_COMMIT = "a516c6b26d183efc4f56293256bba92e243b7a61" # 11/01/2024
+
 http_archive(
  name = "com_google_zetasql",
- urls = ["https://github.com/google/zetasql/archive/%s.zip" % ZETASQL_COMMIT],
+ patch_args = ["-p1"],
+ patches = ["//third_party:zetasql.patch"],
+ sha256 = "1afc2210d4aad371eff0a6bfdd8417ba99e02183a35dff167af2fa6097643f26",
  strip_prefix = "zetasql-%s" % ZETASQL_COMMIT,
- sha256 = "651a768cd51627f58aa6de7039aba9ddab22f4b0450521169800555269447840",
+ urls = ["https://github.com/google/zetasql/archive/%s.tar.gz" % ZETASQL_COMMIT],
 )
 
 load("@com_google_zetasql//bazel:zetasql_deps_step_1.bzl", "zetasql_deps_step_1")
@@ -123,45 +242,39 @@ load("@com_google_zetasql//bazel:zetasql_deps_step_2.bzl", "zetasql_deps_step_2"
 zetasql_deps_step_2(
  analyzer_deps = True,
  evaluator_deps = True,
- tools_deps = False,
  java_deps = False,
  testing_deps = False,
+ tools_deps = False,
 )
 
-# This is part of what zetasql_deps_step_3() does.
-load("@com_google_googleapis//:repository_rules.bzl", "switched_rules_by_language")
-switched_rules_by_language(
- name = "com_google_googleapis_imports",
- cc = True,
-)
+# No need to run zetasql_deps_step_3 and zetasql_deps_step_4 since all necessary dependencies are
+# already installed.
+
+# load("@com_google_zetasql//bazel:zetasql_deps_step_3.bzl", "zetasql_deps_step_3")
+
+# zetasql_deps_step_3()
+
+# load("@com_google_zetasql//bazel:zetasql_deps_step_4.bzl", "zetasql_deps_step_4")
+
+# zetasql_deps_step_4()
 
 _PLATFORMS_VERSION = "0.0.6"
+
 http_archive(
  name = "platforms",
+ sha256 = "5308fc1d8865406a49427ba24a9ab53087f17f5266a7aabbfc28823f3916e1ca",
  urls = [
  "https://mirror.bazel.build/github.com/bazelbuild/platforms/releases/download/%s/platforms-%s.tar.gz" % (_PLATFORMS_VERSION, _PLATFORMS_VERSION),
  "https://github.com/bazelbuild/platforms/releases/download/%s/platforms-%s.tar.gz" % (_PLATFORMS_VERSION, _PLATFORMS_VERSION),
  ],
- sha256 = "5308fc1d8865406a49427ba24a9ab53087f17f5266a7aabbfc28823f3916e1ca",
 )
 
+# Specify the minimum required bazel version.
+load("@bazel_skylib//lib:versions.bzl", "versions")
+
+versions.check("6.5.0")
+
 # Please add all new TensorFlow Data Validation dependencies in workspace.bzl.
 load("//tensorflow_data_validation:workspace.bzl", "tf_data_validation_workspace")
 
 tf_data_validation_workspace()
-
-load("@com_github_tfx_bsl//third_party:python_configure.bzl", "local_python_configure")
-local_python_configure(name = "local_config_python")
-
-PYBIND11_COMMIT = "8a099e44b3d5f85b20f05828d919d2332a8de841" # 2.11.1
-http_archive(
- name = "pybind11",
- build_file = "@com_github_tfx_bsl//third_party:pybind11.BUILD",
- strip_prefix = "pybind11-%s" % PYBIND11_COMMIT,
- urls = ["https://github.com/pybind/pybind11/archive/%s.zip" % PYBIND11_COMMIT],
- sha256 = "8f4b7f28d214e36301435c055076c36186388dc9617117802cba8a059347cb00",
-)
-
-# Specify the minimum required bazel version.
-load("@bazel_skylib//lib:versions.bzl", "versions")
-versions.check("6.1.0")

tensorflow_data_validation/BUILD

@@ -31,11 +31,11 @@ sh_binary(
  srcs = ["move_generated_files.sh"],
  data = select({
  "//conditions:default": [
- "//tensorflow_data_validation/anomalies/proto:custom_validation_config_pb2.py",
- "//tensorflow_data_validation/anomalies/proto:validation_config_pb2.py",
- "//tensorflow_data_validation/anomalies/proto:validation_metadata_pb2.py",
+ "//tensorflow_data_validation/anomalies/proto:custom_validation_config_proto_py_pb2",
+ "//tensorflow_data_validation/anomalies/proto:validation_config_proto_py_pb2",
+ "//tensorflow_data_validation/anomalies/proto:validation_metadata_proto_py_pb2",
  "//tensorflow_data_validation/pywrap:tensorflow_data_validation_extension.so",
- "//tensorflow_data_validation/skew/protos:feature_skew_results_pb2.py",
+ "//tensorflow_data_validation/skew/protos:feature_skew_results_proto_py_pb2",
  ],
  }),
 )