Migrated the Federated Language project to a repository dedicated to …

…this project.

To: https://github.com/google-parfait/federated-language/
From: https://github.com/google-parfait/tensorflow-federated/

TensorFlow Federated (TFF) is being split into multiple projects to enable the federated language, runtimes, and libraries to evolve independently from each other and from TensorFlow. While TFF already supports both TensorFlow and JAX, multiple projects will make it easier and explicit how to extend support to other frameworks (e.g., PyTorch). Additionally, this change will increase the velocity of using TFF to develop new ideas both internally and externally.

PiperOrigin-RevId: 702437793

.bazelrc

@@ -27,11 +27,8 @@ build --compilation_mode=opt
 # build --define=use_fast_cpp_protos=true
 build --define=allow_oversize_protos=true
 
-# pybind_abseil does not include a __init__.py, which breaks dependencies when
-# bazel produces an empty __init__.py and hides the symbols from the `.so`
-# python extension. We must ask bazel not to generate these and rely on
-# explicit __init__.py files.
-# build --incompatible_default_to_explicit_init_py
+# Do not automatically create `__init__.py` in the runfiles of Python targets.
+build --incompatible_default_to_explicit_init_py
 
 # Haswell processor and later optimizations. This covers most processors deployed
 # today, includin Colab CPU runtimes.

WORKSPACE

@@ -87,6 +87,22 @@ http_archive(
 #     },
 # )
 
+http_archive(
+    name = "federated_language",
+    patches = [
+        "//third_party/federated_language:proto_library_loads.patch",
+        "//third_party/federated_language:python_deps.patch",
+        # Must come after `python_deps.patch`, this patches the output of `python_deps.patch`.
+        "//third_party/federated_language:structure_visibility.patch",
+    ],
+    repo_mapping = {
+        "@protobuf": "@com_google_protobuf",
+    },
+    sha256 = "de4bbfd93ee10c3797463d6a96911963d8969fb5ed0b264ae4f3f3088013fed4",
+    strip_prefix = "federated-language-5405fd4b2965e2f7c6c240b386f0540e4114818e",
+    url = "https://github.com/google-parfait/federated-language/archive/5405fd4b2965e2f7c6c240b386f0540e4114818e.tar.gz",
+)
+
 # The version of TensorFlow should match the version in
 # https://github.com/google-parfait/tensorflow-federated/blob/main/requirements.txt.
 http_archive(

docs/deployment.md

@@ -14,7 +14,7 @@ There are two principal modes of deployment for TFF computations:
 *   **Native backends**. We're going to refer to a backend as *native* if it is
     capable of interpreting the syntactic structure of TFF computations as
     defined in
-    [`computation.proto`](https://github.com/google-parfait/tensorflow-federated/blob/main/tensorflow_federated/proto/v0/computation.proto).
+    [`computation.proto`](https://github.com/google-parfait/federated-language/blob/main/tensorflow_federated/proto/computation.proto).
     A native backend does not necessarily have to support all language
     constructs or intrinsics. Native backends must implement one of the standard
     TFF *executor* interfaces, such as

docs/design/compilation.md

@@ -33,13 +33,13 @@ support [Computations](#computation) backed by other external runtimes.
 ### `Computation`
 
 A
-[pb.Computation](https://github.com/google-parfait/tensorflow-federated/blob/main/tensorflow_federated/proto/v0/computation.proto)
+[pb.Computation](https://github.com/google-parfait/federated-language/blob/main/tensorflow_federated/proto/computation.proto)
 is the Proto or serialized representation of the [AST](#ast).
 
 #### TensorFlow Computation
 
 A
-[pb.Computation](https://github.com/google-parfait/tensorflow-federated/blob/main/tensorflow_federated/proto/v0/computation.proto)
+[pb.Computation](https://github.com/google-parfait/federated-language/blob/main/tensorflow_federated/proto/computation.proto)
 that represents a [Computations](#computation) that will be delegated to the
 [TensorFlow](execution.md#tensorflow) runtime.
 

docs/federated_core.md

@@ -80,7 +80,7 @@ blocks such as `tff.federated_sum`, `tff.federated_reduce`, or
 
 TFF uses an internal language to represent federated computations, the syntax of
 which is defined by the serializable representation in
-[computation.proto](https://github.com/google-parfait/tensorflow-federated/blob/main/tensorflow_federated/proto/v0/computation.proto).
+[computation.proto](https://github.com/google-parfait/federated-language/blob/main/tensorflow_federated/proto/computation.proto).
 Users of FC API generally won't need to interact with this language directly,
 though. Rather, we provide a Python API (the `tff` namespace) that wraps arounds
 it as a way to define computations.

examples/custom_data_backend/BUILD

@@ -6,12 +6,11 @@ cc_library(
     hdrs = ["data_backend_example.h"],
     deps = [
         "//tensorflow_federated/cc/core/impl/executors:data_backend",
-        "//tensorflow_federated/cc/core/impl/executors:status_macros",
         "//tensorflow_federated/cc/core/impl/executors:tensorflow_utils",
-        "//tensorflow_federated/proto/v0:array_cc_proto",
-        "//tensorflow_federated/proto/v0:computation_cc_proto",
         "//tensorflow_federated/proto/v0:executor_cc_proto",
         "@com_google_absl//absl/status",
+        "@federated_language//federated_language/proto:array_cc_proto",
+        "@federated_language//federated_language/proto:computation_cc_proto",
         "@org_tensorflow//tensorflow/core:framework",
         "@org_tensorflow//tensorflow/core:protos_all_cc",
     ],

examples/custom_data_backend/data_backend_example.cc

@@ -19,17 +19,17 @@ limitations under the License
 #include <utility>
 
 #include "google/protobuf/any.pb.h"
+#include "federated_language/proto/array.pb.h"
 #include "tensorflow/core/framework/tensor.h"
 #include "tensorflow/core/framework/tensor.pb.h"
 #include "tensorflow_federated/cc/core/impl/executors/tensorflow_utils.h"
-#include "tensorflow_federated/proto/v0/array.pb.h"
 
 namespace tensorflow_federated_examples {
 
 namespace {
 
-using ::tensorflow_federated::v0::Data;
-using ::tensorflow_federated::v0::Type;
+using ::federated_language::Data;
+using ::federated_language::Type;
 using ::tensorflow_federated::v0::Value;
 
 // Constant URIs and values resolved by `DataBackendExample`.

examples/custom_data_backend/data_backend_example.h

@@ -17,19 +17,18 @@ limitations under the License
 #define THIRD_PARTY_TENSORFLOW_FEDERATED_EXAMPLES_CUSTOM_DATA_BACKEND_DATA_BACKEND_EXAMPLE_H_
 
 #include "absl/status/status.h"
+#include "federated_language/proto/computation.pb.h"
 #include "tensorflow_federated/cc/core/impl/executors/data_backend.h"
-#include "tensorflow_federated/proto/v0/computation.pb.h"
 #include "tensorflow_federated/proto/v0/executor.pb.h"
 
 namespace tensorflow_federated_examples {
 
 // An example implementation of `DataBackend` used to show Python interop.
 class DataBackendExample : public tensorflow_federated::DataBackend {
  public:
-  absl::Status ResolveToValue(
-      const tensorflow_federated::v0::Data& data_reference,
-      const tensorflow_federated::v0::Type& data_type,
-      tensorflow_federated::v0::Value& value_out) final;
+  absl::Status ResolveToValue(const federated_language::Data& data_reference,
+                              const federated_language::Type& data_type,
+                              tensorflow_federated::v0::Value& value_out) final;
 };
 
 }  // namespace tensorflow_federated_examples

tensorflow_federated/BUILD

@@ -20,19 +20,13 @@ py_library(
         "//tensorflow_federated/python/core/environments/jax",
         "//tensorflow_federated/python/core/environments/tensorflow",
         "//tensorflow_federated/python/core/framework",
-        "//tensorflow_federated/python/core/impl/computation:computation_base",
-        "//tensorflow_federated/python/core/impl/federated_context:federated_computation",
-        "//tensorflow_federated/python/core/impl/federated_context:intrinsics",
-        "//tensorflow_federated/python/core/impl/federated_context:value_impl",
         "//tensorflow_federated/python/core/impl/types",
-        "//tensorflow_federated/python/core/impl/types:computation_types",
-        "//tensorflow_federated/python/core/impl/types:placements",
-        "//tensorflow_federated/python/core/impl/types:typed_object",
         "//tensorflow_federated/python/core/templates",
         "//tensorflow_federated/python/core/test",
         "//tensorflow_federated/python/learning",
         "//tensorflow_federated/python/program",
         "//tensorflow_federated/python/simulation",
+        "@federated_language//federated_language",
     ],
 )
 

tensorflow_federated/__init__.py

@@ -29,38 +29,38 @@
 from tensorflow_federated.python.core.environments import jax
 from tensorflow_federated.python.core.environments import tensorflow
 from tensorflow_federated.python.core.impl import types
-from tensorflow_federated.python.core.impl.computation.computation_base import Computation
-from tensorflow_federated.python.core.impl.federated_context.federated_computation import federated_computation
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_aggregate
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_broadcast
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_eval
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_map
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_max
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_mean
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_min
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_secure_select
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_secure_sum
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_secure_sum_bitwidth
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_select
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_sum
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_value
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import federated_zip
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import sequence_map
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import sequence_reduce
-from tensorflow_federated.python.core.impl.federated_context.intrinsics import sequence_sum
-from tensorflow_federated.python.core.impl.federated_context.value_impl import to_value
-from tensorflow_federated.python.core.impl.federated_context.value_impl import Value
-from tensorflow_federated.python.core.impl.types.computation_types import FederatedType
-from tensorflow_federated.python.core.impl.types.computation_types import FunctionType
-from tensorflow_federated.python.core.impl.types.computation_types import SequenceType
-from tensorflow_federated.python.core.impl.types.computation_types import StructType
-from tensorflow_federated.python.core.impl.types.computation_types import StructWithPythonType
-from tensorflow_federated.python.core.impl.types.computation_types import TensorType
-from tensorflow_federated.python.core.impl.types.computation_types import to_type
-from tensorflow_federated.python.core.impl.types.computation_types import Type
-from tensorflow_federated.python.core.impl.types.placements import CLIENTS
-from tensorflow_federated.python.core.impl.types.placements import SERVER
-from tensorflow_federated.python.core.impl.types.typed_object import TypedObject
+from federated_language import Computation
+from federated_language import federated_computation
+from federated_language import federated_aggregate
+from federated_language import federated_broadcast
+from federated_language import federated_eval
+from federated_language import federated_map
+from federated_language import federated_max
+from federated_language import federated_mean
+from federated_language import federated_min
+from federated_language import federated_secure_select
+from federated_language import federated_secure_sum
+from federated_language import federated_secure_sum_bitwidth
+from federated_language import federated_select
+from federated_language import federated_sum
+from federated_language import federated_value
+from federated_language import federated_zip
+from federated_language import sequence_map
+from federated_language import sequence_reduce
+from federated_language import sequence_sum
+from federated_language import to_value
+from federated_language import Value
+from federated_language import FederatedType
+from federated_language import FunctionType
+from federated_language import SequenceType
+from federated_language import StructType
+from federated_language import StructWithPythonType
+from federated_language import TensorType
+from federated_language import to_type
+from federated_language import Type
+from federated_language import CLIENTS
+from federated_language import SERVER
+from federated_language import TypedObject
 from tensorflow_federated.version import __version__
 # pylint: enable=g-importing-member