Remove zetasql (#272)

* adds linting configuration and workflow

* update error codes

* reformat linting ignore rules

* add patch to to ignore hooks

* run linter

* fix formatting error

* Remove ZetaSQL from workspace

* add import for re2.BUILD remove tfx_bsl sql_util dependencies

* fix test call

* don't test 3.9 right now

* fix tests

* revert tfx-bsl repo and branch to official

---------

Co-authored-by: smokestacklightnin <[email protected]>
Co-authored-by: pdmurray <[email protected]>

Author: 3 people

.bazelrc

@@ -1,7 +1,11 @@
 # Needed to work with ZetaSQL dependency.
+# Zetasql is removed.
+# This is a candidate for removal
 build --cxxopt="-std=c++17"
 
 # Needed to avoid zetasql proto error.
+# Zetasql is removed.
+# This is a candidate for removal
 build --protocopt=--experimental_allow_proto3_optional
 
 # icu@: In create_linking_context: in call to create_linking_context(),

.github/workflows/test.yml

@@ -29,7 +29,9 @@ jobs:
     - name: Install built wheel
       shell: bash
       run: |
-        pip install dist/*.whl['test']
+        PYTHON_VERSION_TAG="cp$(echo ${{ matrix.python-version }} | sed 's/\.//')"
+        WHEEL_FILE=$(ls dist/*${PYTHON_VERSION_TAG}*.whl)
+        pip install "${WHEEL_FILE}[test]"
 
     - name: Run Test
       run: |

WORKSPACE

@@ -71,18 +71,6 @@ http_archive(
     ],
 )
 
-# Needed by abseil-py by zetasql.
-http_archive(
-    name = "six_archive",
-    build_file = "//third_party:six.BUILD",
-    sha256 = "105f8d68616f8248e24bf0e9372ef04d3cc10104f1980f54d57b2ce73a5ad56a",
-    strip_prefix = "six-1.10.0",
-    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",
-    ],
-)
-
 load("@com_google_protobuf//:protobuf_deps.bzl", "protobuf_deps")
 
 protobuf_deps()
@@ -112,6 +100,16 @@ http_archive(
     url = "https://github.com/abseil/abseil-cpp/archive/%s.tar.gz" % COM_GOOGLE_ABSL_COMMIT,
 )
 
+
+# re2 required for google tests
+http_archive(
+    name = "com_googlesource_code_re2",
+    #    build_file = "//third_party:re2.BUILD",
+    sha256 = "b90430b2a9240df4459108b3e291be80ae92c68a47bc06ef2dc419c5724de061",
+    strip_prefix = "re2-a276a8c738735a0fe45a6ee590fe2df69bcf4502",
+    urls = ["https://github.com/google/re2/archive/a276a8c738735a0fe45a6ee590fe2df69bcf4502.tar.gz"],
+)
+
 # Will be loaded by workspace.bzl from head
 # TFMD_COMMIT = "404805761e614561cceedc429e67c357c62be26d"  # 1.17.1
 
@@ -218,46 +216,6 @@ load("@bazel_gazelle//:deps.bzl", "gazelle_dependencies")  #, "go_repository")
 
 gazelle_dependencies()
 
-################################################################################
-# ZetaSQL                                                                      #
-################################################################################
-
-ZETASQL_COMMIT = "a516c6b26d183efc4f56293256bba92e243b7a61"  # 11/01/2024
-
-http_archive(
-    name = "com_google_zetasql",
-    patch_args = ["-p1"],
-    patches = ["//third_party:zetasql.patch"],
-    sha256 = "1afc2210d4aad371eff0a6bfdd8417ba99e02183a35dff167af2fa6097643f26",
-    strip_prefix = "zetasql-%s" % ZETASQL_COMMIT,
-    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")
-
-zetasql_deps_step_1()
-
-load("@com_google_zetasql//bazel:zetasql_deps_step_2.bzl", "zetasql_deps_step_2")
-
-zetasql_deps_step_2(
-    analyzer_deps = True,
-    evaluator_deps = True,
-    java_deps = False,
-    testing_deps = False,
-    tools_deps = False,
-)
-
-# 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(

tensorflow_data_validation/BUILD

@@ -31,7 +31,6 @@ sh_binary(
     srcs = ["move_generated_files.sh"],
     data = select({
         "//conditions:default": [
-            "//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/anomalies/BUILD

@@ -425,38 +425,6 @@ cc_test(
     ],
 )
 
-cc_library(
-    name = "custom_validation",
-    srcs = ["custom_validation.cc"],
-    hdrs = ["custom_validation.h"],
-    deps = [
-        ":path",
-        ":schema",
-        ":status_util",
-        "//tensorflow_data_validation/anomalies/proto:custom_validation_config_proto_cc_pb2",
-        "@com_github_tensorflow_metadata//tensorflow_metadata/proto/v0:metadata_v0_proto_cc_pb2",
-        "@com_github_tfx_bsl//tfx_bsl/cc/statistics:sql_util",
-        "@com_google_absl//absl/base:log_severity",
-        "@com_google_absl//absl/container:flat_hash_map",
-        "@com_google_absl//absl/log",
-        "@com_google_absl//absl/status",
-        "@com_google_absl//absl/strings",
-        "@com_google_absl//absl/types:optional",
-    ],
-)
-
-cc_test(
-    name = "custom_validation_test",
-    srcs = ["custom_validation_test.cc"],
-    deps = [
-        ":custom_validation",
-        ":test_util",
-        "@com_github_tensorflow_metadata//tensorflow_metadata/proto/v0:metadata_v0_proto_cc_pb2",
-        "@com_google_absl//absl/types:optional",
-        "@com_google_googletest//:gtest_main",
-    ],
-)
-
 cc_library(
     name = "telemetry",
     srcs = ["telemetry.cc"],