Adding pre-commit hooks

.github/workflows/formatting.yml

@@ -0,0 +1,13 @@
+name: pre-commit checks
+on:
+  workflow_dispatch:
+    branches: [ main ]
+  pull_request:
+    branches: [ main ]
+jobs:
+  pre-commit:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v2
+    - uses: actions/setup-python@v2
+    - uses: pre-commit/[email protected]

.github/workflows/main.yml

@@ -9,7 +9,7 @@ on:
     branches: [ main ]
 
 
-  
+
 # A workflow run is made up of one or more jobs that can run sequentially or in parallel
 jobs:
   # This workflow contains a single job called "build"
@@ -21,19 +21,19 @@ jobs:
     steps:
     # Checks-out your repository under $GITHUB_WORKSPACE, so your job can access it
     - uses: actions/checkout@v2
-        
+
     - name: Doxygen Action
       uses: mattnotmitt/[email protected]
       with:
         # Path to Doxyfile
         doxyfile-path: "./docs/Doxyfile" # default is ./Doxyfile
         # Working directory
         working-directory: "." # default is .
-    
+
     - name: Deploy
       uses: peaceiris/actions-gh-pages@v3
       with:
         github_token: ${{ secrets.GITHUB_TOKEN }}
-        # Default Doxyfile build documentation to html directory. 
+        # Default Doxyfile build documentation to html directory.
         # Change the directory if changes in Doxyfile
-        publish_dir: ./html
+        publish_dir: ./html

.gitignore

@@ -1,3 +1,3 @@
 .vscode
 build*
-install*
+install*

.pre-commit-config.yaml

@@ -0,0 +1,25 @@
+repos:
+-   repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v4.0.1
+    hooks:
+    -   id: check-yaml
+    -   id: end-of-file-fixer
+    -   id: trailing-whitespace
+        exclude: test-suite/oracle
+-   repo: https://github.com/psf/black
+    rev: 22.3.0
+    hooks:
+    -   id: black
+-   repo: https://github.com/pocc/pre-commit-hooks
+    rev: v1.1.1
+    hooks:
+        - id: clang-format
+          args: [-i]
+# -   repo: https://github.com/dfm/black_nbconvert
+#     rev: v0.4.0
+#     hooks:
+#         - id: black_nbconvert
+- repo: https://github.com/maxwinterstein/shfmt-py
+  rev: v3.4.3.1
+  hooks:
+    - id: shfmt

CMakeLists.txt

@@ -14,7 +14,7 @@ message("LLVM_MLBRIDGE: ${LLVM_MLBRIDGE}")
 
 if(NOT LLVM_MLBRIDGE)
   find_package(LLVM 10.0.0 REQUIRED CONFIG)
-  include_directories(${LLVM_INCLUDE_DIRS}) 
+  include_directories(${LLVM_INCLUDE_DIRS})
   link_directories(${LLVM_LIBRARY_DIR})
   add_compile_definitions(C_LIBRARY)
 endif()
@@ -25,9 +25,9 @@ if (DEFINED LLVM_HAVE_TF_AOT OR LLVM_HAVE_TFLITE)
   set(LLVM_INLINER_MODEL_PATH_DEFAULT "models/inliner-Oz")
 
   set(LLVM_INLINER_MODEL_CURRENT_URL "<UNSPECIFIED>" CACHE STRING "URL to download the LLVM inliner model")
-  
+
   if (DEFINED LLVM_HAVE_TF_AOT)
-    
+
     tf_find_and_compile(
       ${LLVM_INLINER_MODEL_PATH}
       ${LLVM_INLINER_MODEL_CURRENT_URL}
@@ -41,23 +41,23 @@ if (DEFINED LLVM_HAVE_TF_AOT OR LLVM_HAVE_TFLITE)
       )
     endif()
   endif()
-  
+
 add_subdirectory(MLModelRunner)
 add_subdirectory(SerDes)
 
 if(LLVM_MLBRIDGE)
   include(AddLLVM)
   include(HandleLLVMOptions)
   include(LLVMDistributionSupport)
-  
+
   add_llvm_library(LLVMMLBridge
   tools.cpp
   ${GeneratedMLSources}
-  
+
   ADDITIONAL_HEADER_DIRS
   ${CMAKE_CURRENT_SOURCE_DIR}/include
   ${TENSORFLOW_AOT_PATH}/include
-  
+
   LINK_LIBS
   ModelRunnerLib
   $<TARGET_OBJECTS:SerDesLib>
@@ -68,7 +68,7 @@ endif(LLVM_MLBRIDGE)
 
 llvm_map_components_to_libnames(llvm_libs support core irreader analysis TransformUtils)
 
-add_custom_target(protobuf_version ALL 
+add_custom_target(protobuf_version ALL
 COMMAND ${CMAKE_COMMAND} -E echo "protoc path = $<TARGET_FILE:protobuf::protoc> Using Protobuf ${Protobuf_VERSION}")
 
 add_library(MLCompilerBridgeC STATIC $<TARGET_OBJECTS:ModelRunnerCWrapper>)

CompilerInterface/BaseCompilerInterface.py

@@ -7,7 +7,7 @@
 class BaseCompilerInterface(ABC):
     ## Initializes correct SerDes object
     def __init__(self, data_format=None):
-        self.serdes_obj = SerDes(data_format) 
+        self.serdes_obj = SerDes(data_format)
 
     ## Places data for next request into a buffer after serialization.
     # @param Unserialized data for next query to compiler
@@ -18,5 +18,3 @@ def populate_buffer(self, data):
     ## Sends query to compiler and returns deserialized result.
     def evaluate(self):
         pass
-
-

CompilerInterface/GrpcCompilerInterface.py

@@ -14,22 +14,32 @@ class GrpcCompilerInterface(BaseCompilerInterface):
     ## Initializes GrpcCompilerInterface object.
     # @param mode Can be 'client' or 'server'.
     # @param stub_class gRPC stub class used in 'client' mode
-    # @param hostip 
-    # @param hostport 
+    # @param hostip
+    # @param hostport
     # @param add_server_method used in 'server' mode
     # @param grpc_service_obj used in 'server' mode
-    def __init__(self, mode, stub_class=None, hostip='127.0.0.1', hostport=50051, add_server_method=None,grpc_service_obj=None ):
-        super().__init__('protobuf')
+    def __init__(
+        self,
+        mode,
+        stub_class=None,
+        hostip="127.0.0.1",
+        hostport=50051,
+        add_server_method=None,
+        grpc_service_obj=None,
+    ):
+        super().__init__("protobuf")
         self.mode = mode
-        self.host=hostip
+        self.host = hostip
         self.server_port = hostport
 
-        if self.mode == 'client':
-            self.channel = grpc.insecure_channel('{}:{}'.format(self.host,self.server_port))
+        if self.mode == "client":
+            self.channel = grpc.insecure_channel(
+                "{}:{}".format(self.host, self.server_port)
+            )
             print("Setting stub", stub_class)
             self.stub = stub_class(self.channel)
-        
-        elif self.mode == 'server':
+
+        elif self.mode == "server":
             self.grpc_service_obj = grpc_service_obj
             self.add_server_method = add_server_method
             self.start_server()
@@ -38,23 +48,24 @@ def __del__(self):
         pass
 
     ## Sends query to compiler and returns deserialized result.
-    def evaluate(self, mode = None):
+    def evaluate(self, mode=None):
         out = self.serdes_obj.getOutputBuffer()
         return self.stub.queryCompiler(out)
 
-
     ## Starts gRPC server
     def start_server(self):
-        if self.mode == 'server':
-            server=grpc.server(futures.ThreadPoolExecutor(max_workers=20),options = [
-                        ('grpc.max_send_message_length', 200*1024*1024), #50MB
-                                ('grpc.max_receive_message_length', 200*1024*1024) #50MB
-                                    ])
+        if self.mode == "server":
+            server = grpc.server(
+                futures.ThreadPoolExecutor(max_workers=20),
+                options=[
+                    ("grpc.max_send_message_length", 200 * 1024 * 1024),  # 50MB
+                    ("grpc.max_receive_message_length", 200 * 1024 * 1024),  # 50MB
+                ],
+            )
 
-            self.add_server_method(self.grpc_service_obj,server)
-            server.add_insecure_port('{}:{}'.format(self.host,self.server_port))
+            self.add_server_method(self.grpc_service_obj, server)
+            server.add_insecure_port("{}:{}".format(self.host, self.server_port))
 
             server.start()
-            print("Server Running")        
+            print("Server Running")
             server.wait_for_termination()
-

CompilerInterface/PipeCompilerInterface.py

@@ -6,7 +6,7 @@
 ## This class implements methods for communication with compiler using pipes.
 class PipeCompilerInterface(BaseCompilerInterface):
     ## Initializes PipeCompilerInterface object.
-    # @param data_format Data format for serialization 
+    # @param data_format Data format for serialization
     # @param pipe_name Name for pipe file
     def __init__(self, data_format=None, pipe_name=None):
         super().__init__(data_format)
@@ -23,29 +23,28 @@ def __del__(self):
         self.remove_pipes()
 
     ## Sends query to compiler and returns deserialized result.
-    def evaluate(self, mode = None):
+    def evaluate(self, mode=None):
         out = self.serdes_obj.getOutputBuffer()
         if out is not None:
             self.tc.write(out)
             self.tc.flush()
 
-        if mode == 'exit':
+        if mode == "exit":
             return None
 
         result = self.serdes_obj.deserializeData(self.fc)
 
         return result
 
-
-    ## Creates pipe files for communication. 
+    ## Creates pipe files for communication.
     def init_pipes(self):
         self.to_compiler = self.pipe_name + ".in"
         self.from_compiler = self.pipe_name + ".out"
         if os.path.exists(self.to_compiler):
             os.remove(self.to_compiler)
         if os.path.exists(self.from_compiler):
             os.remove(self.from_compiler)
-            
+
         os.mkfifo(self.to_compiler, 0o666)
         os.mkfifo(self.from_compiler, 0o666)
 

CompilerInterface/SerDes.py

@@ -13,7 +13,7 @@ def __init__(self, data_format):
         self.read_stream_iter = None
 
         self.serMap = {
-            "json": self.serializeJson,             
+            "json": self.serializeJson,
             "bytes": self.serializeBytes,
             "protobuf": self.serializeProtobuf,
         }
@@ -39,11 +39,13 @@ def deserializeJson(self, datastream):
     ##  Deserializes and returns bitstream data
     def deserializeBytes(self, datastream):
         if self.read_stream_iter is None:
-            self.read_stream_iter = log_reader.read_stream2(datastream)        # try to make it indep
+            self.read_stream_iter = log_reader.read_stream2(
+                datastream
+            )  # try to make it indep
         hdr = datastream.read(8)
         context, observation_id, features, score = next(self.read_stream_iter)
         return features
-    
+
     # Not implemented
     def deserializeProtobuf(self, datastream):
         raise NotImplementedError
@@ -52,7 +54,7 @@ def deserializeProtobuf(self, datastream):
     def serializeData(self, data):
         self.serMap[self.data_format](data)
 
-    ## Serializes data to JSON 
+    ## Serializes data to JSON
     def serializeJson(self, data):
         msg = json.dumps({"out": data}).encode("utf-8")
         hdr = len(msg).to_bytes(8, "little")
@@ -72,7 +74,7 @@ def serializeProtobuf(self, data):
         self.buffer = data
 
     ## Returns value in buffer and empties it
-    # @return Data from output buffer 
+    # @return Data from output buffer
     def getOutputBuffer(self):
         out = self.buffer
         self.buffer = None

MLModelRunner/CMakeLists.txt

@@ -30,7 +30,7 @@ if (NOT TENSORFLOW_AOT_PATH STREQUAL "")
   install(TARGETS tf_xla_runtime EXPORT LLVMExports
     ARCHIVE DESTINATION lib${LLVM_LIBDIR_SUFFIX} COMPONENT tf_xla_runtime)
   set_property(GLOBAL APPEND PROPERTY LLVM_EXPORTS tf_xla_runtime)
-  
+
   # Once we add more modules, we should handle this more automatically.
   if (DEFINED LLVM_OVERRIDE_MODEL_HEADER_INLINERSIZEMODEL)
     set(LLVM_INLINER_MODEL_PATH "none")

MLModelRunner/ONNXModelRunner/CMakeLists.txt

@@ -50,5 +50,5 @@ add_library(ONNXModelRunnerLib OBJECT onnx.cpp
     )
 endif(LLVM_MLBRIDGE)
 
-target_link_libraries(ONNXModelRunnerLib PUBLIC onnxruntime)  
+target_link_libraries(ONNXModelRunnerLib PUBLIC onnxruntime)
 target_include_directories(ONNXModelRunnerLib PUBLIC "${ONNXRUNTIME_ROOTDIR}/include" "${ONNXRUNTIME_ROOTDIR}/include/onnxruntime/core/session" ${TENSORFLOW_AOT_PATH}/include)

MLModelRunner/ONNXModelRunner/README.md

@@ -12,14 +12,14 @@ Integration of a trained model happnes in two steps.
 
 ## Directory structure
 
-The direcory and its subdirctories contain source code for OONX model integrations as well as base classes of vaious components used by and RL algorithm. 
+The direcory and its subdirctories contain source code for OONX model integrations as well as base classes of vaious components used by and RL algorithm.
 
     .
     ├── onnx.cpp                # Defination of ONNXModel class
     ├── agent.cpp               # RL Agent class which will compute action by quring model
-    ├── Include               
+    ├── Include
     │   ├── agent.h             # Declare RL agent class
-    │   ├── driver.h            # A interface class to query RL baaed model 
+    │   ├── driver.h            # A interface class to query RL baaed model
     │   ├── environment.h       # Declare a base Environment class which can be extended further
     │   ├── onnx.h              # Declaration of ONNXModel class
     │   └── utils.h             # Utility defination used my Agent class
@@ -89,11 +89,10 @@ struct Hello : public FunctionPass, Environment {
 
 bool runOnFunction(Function &F) override {
  ...
- // Creates instance of DriverService with InferenceEngine as parent class 
+ // Creates instance of DriverService with InferenceEngine as parent class
  InferenceEngine* inference_driver = new DriverService(Environment* env);
  inference_driver->getPredictions(PassData passData, OptInfo &predictions);
  ...
 }
 
 ```
-

MLModelRunner/ONNXModelRunner/agent.cpp

@@ -37,4 +37,4 @@ unsigned Agent::computeAction(Observation &input) {
   LLVM_DEBUG(llvm::errs() << "\nmax value and index are " << *max << " "
                           << argmaxVal << "\n");
   return argmaxVal;
-}
+}

MLModelRunner/ONNXModelRunner/onnx.cpp

@@ -15,7 +15,6 @@
 #include <assert.h>
 #include <iostream>
 #include <numeric>
-#include <iostream>
 
 ONNXModel::ONNXModel(const char *model_path) : model_path(model_path) {
   env = new Ort::Env(ORT_LOGGING_LEVEL_WARNING, "test");

MLModelRunner/Utils/CMakeLists.txt

@@ -5,4 +5,4 @@ if(LLVM_MLBRIDGE)
   target_include_directories(ModelRunnerUtils PUBLIC ${CMAKE_CURRENT_SOURCE_DIR}/../../include)
 endif(LLVM_MLBRIDGE)
 
-add_library(ModelRunnerCUtils OBJECT MLConfig.cpp)
+add_library(ModelRunnerCUtils OBJECT MLConfig.cpp)