Updates (vendored) bleach to 6.2.0 and removes dependencies on six and html5lib (#7013)

## Motivation for features / changes

Addresses #7010.

The old version of bleach that we were using, needed dependencies `six`
and `html5lib` to be installed, and the old version of "six" that we
were using for CI via bazel is not supported in python >= 3.12, which
causes some issues at least for local builds and CI when using newer
versions of python.

## Technical description of changes

Update the version of bleach specified in the `http_archive` rule used
by bazel and remove rules referring to `six` and `html5lib`.

## Detailed steps to verify changes work correctly (as executed by you)

Created a virtualenv with python 3.12+ and ran `bazel run //tensorboard
-- --logdir=<logdir>`, and verified that it runs correctly. Also CI
should be passing for this to get merged.

Author: arcra

tensorboard/pip_package/BUILD

@@ -86,7 +86,6 @@ genrule(
         "@npm//:node_modules/d3/LICENSE",
         "@com_google_fonts_roboto//:LICENSE",
         "@org_mozilla_bleach//:LICENSE",
-        "@org_html5lib//:LICENSE",
         "@org_pythonhosted_webencodings//:LICENSE",
         "//third_party:bh_tsne.LICENSE",
     ],

tensorboard/pip_package/build_pip_package.sh

@@ -86,15 +86,12 @@ build() (
 
   mkdir -p tensorboard/_vendor
   >tensorboard/_vendor/__init__.py
-  cp -LR "${RUNFILES}/org_html5lib/html5lib" tensorboard/_vendor
   cp -LR "${RUNFILES}/org_mozilla_bleach/bleach" tensorboard/_vendor
   cp -LR "${RUNFILES}/org_pythonhosted_webencodings/webencodings" tensorboard/_vendor
 
   chmod -R u+w,go+r .
 
   find tensorboard -name \*.py -exec $sedi -e '
-      s/^import html5lib$/from tensorboard._vendor import html5lib/
-      s/^from html5lib/from tensorboard._vendor.html5lib/
       s/^import bleach$/from tensorboard._vendor import bleach/
       s/^from bleach/from tensorboard._vendor.bleach/
       s/^import webencodings$/from tensorboard._vendor import webencodings/

tensorboard/pip_package/requirements.txt

@@ -33,9 +33,5 @@ packaging
 # https://github.com/protocolbuffers/protobuf/issues/13485
 protobuf >= 3.19.6, != 4.24.0
 setuptools >= 41.0.0  # Note: provides pkg_resources as well as setuptools
-# A dependency of our vendored packages. This lower bound has not been correctly
-# vetted, but I wanted to be the least restrictive we can, since it's not a new
-# requirement, and likely this will not disrupt existing users of the package.
-six > 1.9
 tensorboard-data-server >= 0.7.0, < 0.8.0
 werkzeug >= 1.0.1

tensorboard/plugins/text/BUILD

@@ -21,7 +21,6 @@ py_library(
         "//tensorboard/backend:http_util",
         "//tensorboard/data:provider",
         "//tensorboard/plugins:base_plugin",
-        "@org_mozilla_bleach",
         "@org_pocoo_werkzeug",
         "@org_pythonhosted_markdown",
     ],

third_party/bleach.BUILD

@@ -15,15 +15,11 @@ py_library(
     srcs = [
         "bleach/__init__.py",
         "bleach/callbacks.py",
-        "bleach/encoding.py",
+        "bleach/html5lib_shim.py",
         "bleach/linkifier.py",
+        "bleach/parse_shim.py",
         "bleach/sanitizer.py",
-        "bleach/utils.py",
-        "bleach/version.py",
-    ],
-    srcs_version = "PY3",
-    deps = [
-        "@org_html5lib",
-        "@org_pythonhosted_six",
-    ],
+        "bleach/six_shim.py",
+    ] + glob(["bleach/_vendor/**/*.py"]),
+    deps = ["@org_pythonhosted_webencodings"],
 )

third_party/html5lib.BUILD

@@ -44,25 +44,14 @@ def tensorboard_python_workspace():
         build_file = str(Label("//third_party:urllib3.BUILD")),
     )
 
-    http_archive(
-        name = "org_html5lib",
-        urls = [
-            "http://mirror.tensorflow.org/github.com/html5lib/html5lib-python/archive/1.1.tar.gz",
-            "https://github.com/html5lib/html5lib-python/archive/1.1.tar.gz",
-        ],
-        sha256 = "66e9e24a53c10c27abb6be8a3cf2cf55824c6ea1cef8570a633cb223ec46e894",
-        strip_prefix = "html5lib-python-1.1",
-        build_file = str(Label("//third_party:html5lib.BUILD")),
-    )
-
     http_archive(
         name = "org_mozilla_bleach",
         urls = [
-            "http://mirror.tensorflow.org/github.com/mozilla/bleach/archive/v2.0.tar.gz",
-            "https://github.com/mozilla/bleach/archive/v2.0.tar.gz",
+            "http://mirror.tensorflow.org/files.pythonhosted.org/packages/76/9a/0e33f5054c54d349ea62c277191c020c2d6ef1d65ab2cb1993f91ec846d1/bleach-6.2.0.tar.gz",
+            "https://files.pythonhosted.org/packages/76/9a/0e33f5054c54d349ea62c277191c020c2d6ef1d65ab2cb1993f91ec846d1/bleach-6.2.0.tar.gz"
         ],
-        strip_prefix = "bleach-2.0",
-        sha256 = "789dcf3e7daf79c4c78518c6ebafd51bbaf111ac4263a97c08cf8d6a27eda820",
+        strip_prefix = "bleach-6.2.0",
+        sha256 = "123e894118b8a599fd80d3ec1a6d4cc7ce4e5882b1317a7e1ba69b56e95f991f",
         build_file = str(Label("//third_party:bleach.BUILD")),
     )
 
@@ -99,13 +88,3 @@ def tensorboard_python_workspace():
         ],
     )
 
-    http_archive(
-        name = "org_pythonhosted_six",
-        urls = [
-            "http://mirror.tensorflow.org/pypi.python.org/packages/source/s/six/six-1.13.0.tar.gz",
-            "https://pypi.python.org/packages/source/s/six/six-1.13.0.tar.gz",
-        ],
-        sha256 = "30f610279e8b2578cab6db20741130331735c781b56053c59c4076da27f06b66",
-        strip_prefix = "six-1.13.0",
-        build_file = str(Label("//third_party:six.BUILD")),
-    )

third_party/python.bzl

@@ -35,12 +35,6 @@ def tensorboard_workspace(name = ""):
     tensorboard_js_workspace()
     tensorboard_rust_workspace()
 
-    # Protobuf's BUILD file depends on //external:six.
-    native.bind(
-        name = "six",
-        actual = "@org_pythonhosted_six",
-    )
-
     # Needed by Protobuf.
     native.bind(
         name = "grpc_python_plugin",