Remove TensorFlow Serving (TFS) gRPC/REST frontend

.bazelrc

@@ -56,7 +56,6 @@ build --define=MEDIAPIPE_DISABLE_GPU=1
 build --define=MEDIAPIPE_DISABLE=0
 build --define=CLOUD_DISABLE=0
 build --define=PYTHON_DISABLE=0
-build --define=USE_DROGON=1  # 0=legacy net_http, available only on linux
 
 # Sets the default Apple platform to macOS.
 build --apple_platform_type=macos

BUILD.bazel

@@ -33,15 +33,12 @@ create_config_settings()
 cc_library(
   name = "ovms_dependencies",
   deps = [
-        "@tensorflow_serving//tensorflow_serving/apis:prediction_service_cc_proto",
-        "@tensorflow_serving//tensorflow_serving/apis:model_service_cc_proto",
         "@minitrace//:trace",
         "@com_github_grpc_grpc//:grpc++",
         "@com_github_tencent_rapidjson//:rapidjson",
         "@org_tensorflow//tensorflow/core:framework",
         "@com_github_gabime_spdlog//:spdlog",
         "@com_github_jarro2783_cxxopts//:cxxopts",
-        "@tensorflow_serving//tensorflow_serving/util:json_tensor",
         "//third_party:openvino",
         "//third_party:opencv",
         "//third_party:curl",

WORKSPACE

@@ -158,21 +158,6 @@ cc_library(
 )
 
 
-# Used for gRPC API protos only
-# Tensorflow serving
-git_repository(
-    name = "tensorflow_serving",
-    remote = "https://github.com/tensorflow/serving.git",
-    tag = "2.18.0",
-    patch_args = ["-p1"],
-    patches = ["net_http.patch", "listen.patch", "partial_2.18.patch"]
-    #                             ^^^^^^^^^^^^
-    #                       make bind address configurable
-    #          ^^^^^^^^^^^^
-    #        allow all http methods                ^^^^^^^^^
-    #                                        implements partial responses
-)
-
 ########################################################### Mediapipe
 http_archive(
     name = "com_google_protobuf",
@@ -434,8 +419,6 @@ http_archive(
     repo_mapping = {"@curl" : "@curl"}
 )
 
-load("@tensorflow_serving//tensorflow_serving:workspace.bzl", "tf_serving_workspace")
-
 # Initialize TensorFlow's external dependencies.
 load("@org_tensorflow//tensorflow:workspace3.bzl", "workspace")
 workspace()
@@ -473,7 +456,6 @@ workspace()
 load("@org_tensorflow//tensorflow:workspace0.bzl", "workspace")
 workspace()
 
-tf_serving_workspace() #moved past TF
 # required after update to mp 0.10.18
 load(
     "@org_tensorflow//third_party/gpus/cuda/hermetic:cuda_configure.bzl",

ci/build_test_OnCommit.groovy

@@ -343,7 +343,7 @@ pipeline {
                       def test_doc_files_str = test_doc_files_linux.split('\n').join(' or ')
                       sh "make create-venv && rm -f tests/functional && ln -s ${pwd}/../tests/functional tests/functional"
                       def cmd_venv_activate = ". .venv/bin/activate"
-                      def cmd_export = "export TT_OVMS_C_REPO_PATH=../ && export TT_RUN_REGRESSION_TESTS=True && export TT_REGRESSION_WEEKLY_TESTS=True && export TT_TARGET_DEVICE=CPU,GPU,NPU && export TT_ENABLE_UAT_TESTS=True && export TT_ENABLE_SMOKE_TESTS=False && export TT_OVMS_C_REPO_PATH=${ovms_c_repo_path} && export TT_LOGGING_LEVEL_OVMS=DEBUG && export TT_WAIT_FOR_MESSAGES_TIMEOUT=1500"
+                      def cmd_export = "export TT_OVMS_C_REPO_PATH=../ && export TT_RUN_REGRESSION_TESTS=True && export TT_REGRESSION_WEEKLY_TESTS=True && export TT_TARGET_DEVICE=CPU,GPU,NPU && export TT_ENABLE_UAT_TESTS=True && export TT_ENABLE_SMOKE_TESTS=False && export TT_OVMS_C_REPO_PATH=${ovms_c_repo_path} && export TT_LOGGING_LEVEL_OVMS=DEBUG && export TT_WAIT_FOR_MESSAGES_TIMEOUT=1500 && export CORE_BRANCH=${env.CHANGE_BRANCH ?: 'main'}"
                       def cmd_pytest = "pytest tests/non_functional/documentation -k '${test_doc_files_str}' -n 0 --dist loadgroup"
                       def cmd = ""
                       if ( image_build_needed == "true" ) {
@@ -408,7 +408,7 @@ pipeline {
                       def ovms_c_repo_path = bat(returnStdout: true, script: 'cd .. && cd').trim().split('\n').last().trim()
                       def cmd_link_ovms = "(if exist ${current_path}\\tests\\functional rmdir ${current_path}\\tests\\functional) && mklink /D ${current_path}\\tests\\functional ${ovms_c_repo_path}\\tests\\functional"
                       def cmd_requirements = "(if not exist .venv virtualenv .venv --python=python3.12) && call .venv\\Scripts\\activate.bat && pip install -r requirements.txt"
-                      def cmd_export = "set \"TT_OVMS_C_REPO_PATH=../\" && set \"TT_LOGGING_LEVEL_OVMS=DEBUG\" && set \"TT_RUN_REGRESSION_TESTS=True\" && set \"TT_REGRESSION_WEEKLY_TESTS=True\" && set \"TT_TARGET_DEVICE=CPU,GPU,NPU\" && set \"TT_BASE_OS=windows\" && set \"TT_OVMS_TYPE=BINARY\" && set \"TT_ENABLE_UAT_TESTS=True\" && set \"TT_ENABLE_SMOKE_TESTS=False\" && set \"TT_DISABLE_DMESG_LOG_MONITOR=True\" && set \"TT_OVMS_C_REPO_PATH=${ovms_c_repo_path}\" && set \"TT_WAIT_FOR_MESSAGES_TIMEOUT=1500\" && set \"PYTHONUTF8=1\" && set \"PYTHONIOENCODING=utf-8\""
+                      def cmd_export = "set \"TT_OVMS_C_REPO_PATH=../\" && set \"TT_LOGGING_LEVEL_OVMS=DEBUG\" && set \"TT_RUN_REGRESSION_TESTS=True\" && set \"TT_REGRESSION_WEEKLY_TESTS=True\" && set \"TT_TARGET_DEVICE=CPU,GPU,NPU\" && set \"TT_BASE_OS=windows\" && set \"TT_OVMS_TYPE=BINARY\" && set \"TT_ENABLE_UAT_TESTS=True\" && set \"TT_ENABLE_SMOKE_TESTS=False\" && set \"TT_DISABLE_DMESG_LOG_MONITOR=True\" && set \"TT_OVMS_C_REPO_PATH=${ovms_c_repo_path}\" && set \"TT_WAIT_FOR_MESSAGES_TIMEOUT=1500\" && set \"PYTHONUTF8=1\" && set \"PYTHONIOENCODING=utf-8\" && set \"CORE_BRANCH=${env.CHANGE_BRANCH ?: 'main'}\""
                       def cmd_pytest = "pytest tests/non_functional/documentation -k \"${test_doc_files_str}\" -n 0 --dist loadgroup --basetemp=\"C:\\tmp\\pytest-${BRANCH_NAME}-${BUILD_NUMBER}\""
                       def cmd = ""
                       if ( win_image_build_needed == "true" ) {

client/python/README.md

@@ -2,14 +2,5 @@
 
 OpenVINO Model Server exposes network interface that client applications can interact with.
 
-There's also [tensorflow-serving-api](https://pypi.org/project/tensorflow-serving-api/) package that can be used to send requests to OpenVINO Model Server. 
-
-> **Note**: `tensorflow-serving-api` comes with `tensorflow` dependency which makes the package heavy.
-
-There are sample scripts for `tensorflow-serving-api`:
-- [tensorflow-serving-api samples](tensorflow-serving-api/samples)
-
-Additionally, starting with the 2022.2 release, OpenVINO Model Server also supports [KServe API](https://github.com/kserve/kserve/blob/master/docs/predict-api/v2/required_api.md) calls. You can try it out with the:
- - [KServe samples](kserve-api/samples)
-
-> **Note**: 2025.3 is the last release that includes ovmsclient and it is not supported anymore. If you still need it, use [release 2025.3](https://github.com/openvinotoolkit/model_server/tree/releases/2025/3/client/python/ovmsclient)
+OpenVINO Model Server supports [KServe API](https://github.com/kserve/kserve/blob/master/docs/predict-api/v2/required_api.md) calls. You can try it out with the:
+ - [KServe samples](kserve-api/samples)