Adding support for state-of-the-art Yolov4 with OPENCV 4.4-pre .

multiple improvment for error handeling and logging

Author: hadikoub

.DS_Store

@@ -0,0 +1,21 @@
+{
+    // Use IntelliSense to learn about possible attributes.
+    // Hover to view descriptions of existing attributes.
+    // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "name": "Python: Remote Attach",
+            "type": "python",
+            "request": "attach",
+            "port": 4000,
+            "host": "0.0.0.0",
+            "pathMappings": [
+                {
+                    "localRoot": "${workspaceFolder}",
+                    "remoteRoot": "${workspaceFolder}`"
+                }
+            ]
+        }
+    ]
+}

.vscode/launch.json

@@ -1,10 +1,10 @@
-# YOLO v3 CPU Inference API for Windows and Linux
+# YOLO v4-v3 CPU Inference API for Windows and Linux
 
-This is a repository for an object detection inference API using the Yolov3 Opencv.
+This is a repository for an object detection inference API using the Yolov4 and Yolo v3  Opencv.
 
 The inference REST API works on CPU and doesn't require any GPU usage. It's supported on both Windows and Linux Operating systems.
 
-Models trained using our training Yolov3 repository can be deployed in this API. Several object detection models can be loaded and used at the same time.
+Models trained using our training Yolov4 or Yolov3  repository can be deployed in this API. Several object detection models can be loaded and used at the same time.
 
 ![predict image](./docs/4.gif)
 
@@ -44,13 +44,12 @@ To [install Docker on Windows](https://docs.docker.com/docker-for-windows/instal
 In order to build the project run the following command from the project's root directory:
 
 ```sh
-sudo docker build -t yolov3_inference_api_cpu -f ./docker/dockerfile .
+sudo docker build -t yolov4_inference_api_cpu -f ./docker/dockerfile .
 ```
 ### Behind a proxy
 
 ```sh
-sudo docker build --build-arg http_proxy='' --build-arg https_proxy='' -t yolov3_inference_api_cpu -f ./docker/dockerfile .
-
+sudo docker build --build-arg http_proxy='' --build-arg https_proxy='' -t yolov4_inference_api_cpu -f ./docker/dockerfile .
 ```
 
 ## Run The Docker Container
@@ -60,12 +59,12 @@ To run the API, go the to the API's directory and run the following:
 #### Using Linux based docker:
 
 ```sh
-sudo docker run -itv $(pwd)/models:/models -p <docker_host_port>:7770 yolov3_inference_api_cpu
+sudo docker run -itv $(pwd)/models:/models -p <docker_host_port>:7770 yolov4_inference_api_cpu
 ```
 #### Using Windows based docker:
 
 ```sh
-docker run -itv ${PWD}/models:/models -p <docker_host_port>:7770 yolov3_inference_api_cpu
+docker run -itv ${PWD}/models:/models -p <docker_host_port>:7770 yolov4_inference_api_cpu
 ```
 
 The <docker_host_port> can be any unique port of your choice.
@@ -150,7 +149,7 @@ Inside each subfolder there should be a:
 
   ```json
     {
-      "inference_engine_name": "yolov3_opencv_cpu_detection",
+      "inference_engine_name": "yolov4_opencv_cpu_detection",
       "confidence": 60,
       "nms_threshold": 0.6,
       "image": {
@@ -171,6 +170,9 @@ Inside each subfolder there should be a:
     }
   ```
   P.S
+  
+  - You can choose **"inference_engine_name"**: between **yolov4_opencv_cpu_detection** and **yolov3_opencv_cpu_detection** depending on the model you have.
+  
   - You can change confidence and nms_threshold values while running the API
   - The API will return bounding boxes with a confidence higher than the "confidence" value. A high "confidence" can show you only accurate predictions
 

README.md

@@ -3,11 +3,50 @@ FROM python:3.6
 LABEL maintainer="antoine.charbel@inmind.ai"
 
 COPY docker/requirements.txt .
+ENV HOME /root
 
 COPY src/main /main
 
+RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get -y --no-install-recommends install \
+	locales \
+        wget \
+        software-properties-common \
+        ca-certificates \
+        build-essential \
+        cmake \
+        git \
+        libopencv-dev \
+        python3-dev \
+        python3-pip \
+        libgtk2.0-dev\
+        pkg-config\
+    && apt-get -y autoremove \
+    && apt-get clean \
+    && pip3 install setuptools wheel \
+    && rm -rf /var/lib/apt/lists/{apt,dpkg,cache,log} /tmp/* /var/tmp/*
+
 RUN pip install -r requirements.txt
 
-WORKDIR /main
 
+WORKDIR ${HOME}
+RUN git clone http://github.com/opencv/opencv.git && cd opencv \
+    && git checkout master                   \
+    && mkdir build && cd build              \
+    && cmake -D CMAKE_BUILD_TYPE=RELEASE    \
+        -D CMAKE_INSTALL_PREFIX=/usr/local  \
+        -D WITH_CUDA=OFF                     \
+        -D WITH_OPENCL=OFF                  \
+        -D ENABLE_FAST_MATH=1               \
+        -D CUDA_FAST_MATH=1                 \
+        -D WITH_CUBLAS=1                    \
+        -D BUILD_DOCS=OFF                   \
+        -D BUILD_PERF_TESTS=OFF             \
+        -D BUILD_TESTS=OFF                  \
+        ..                                  \
+    && make -j `nproc`                      \
+    && make install                         \
+    && cd ${HOME} && rm -rf ./opencv/
+
+
+WORKDIR /main
 CMD ["uvicorn", "start:app", "--host", "0.0.0.0", "--port", "7770"]

docker/dockerfile

@@ -130,6 +130,8 @@ def get_labels(self, model_name):
         :param model_name: Model name
         :return: List of model labels
         """
+        if model_name not in self.models_hash_dict and model_name not in self.models_hash_dict.values():
+            raise ModelNotFound
         if not self.model_loaded(model_name):
             self.load_model(model_name)
         return self.models_dict[model_name].labels
@@ -140,6 +142,8 @@ def get_labels_custom(self, model_name):
         :param model_name: Model name
         :return: A list of mode's labels with their hashed values
         """
+        if model_name not in self.models_hash_dict and model_name not in self.models_hash_dict.values():
+            raise ModelNotFound 
         if re.match(r'[0-9a-fA-F]{8}\-[0-9a-fA-F]{4}\-[0-9a-fA-F]{4}\-[0-9a-fA-F]{4}\-[0-9a-fA-F]{12}', model_name,
                     flags=0):
             for key, value in self.models_hash_dict.items():
@@ -162,6 +166,8 @@ def get_config(self, model_name):
         :param model_name: Model name
         :return: List of model's configuration
         """
+        if model_name not in self.models_hash_dict and model_name not in self.models_hash_dict.values():
+            raise ModelNotFound 
         if not self.model_loaded(model_name):
             self.load_model(model_name)
         return self.models_dict[model_name].configuration

src/.DS_Store

@@ -26,7 +26,8 @@ def get_engine(path_to_model):
 		try:
 			# import one of the available inference engine class (in this project there's only one), and return a
 			# model instance
-			return getattr(__import__(inference_engine_name), 'InferenceEngine')(path_to_model)
+			if inference_engine_name=='yolov3_opencv_cpu_detection' or inference_engine_name=='yolov4_opencv_cpu_detection':
+				return getattr(__import__("yolov3_opencv_cpu_detection"), 'InferenceEngine')(path_to_model)
 		except ApplicationError as e:
 			raise e
 		except Exception as e:

src/main/deep_learning_service.py

@@ -40,10 +40,13 @@ def load_custom():
 	:return: All the available models with their respective hashed values
 	"""
 	try:
+		error_logging.info('request successful;')
 		return dl_service.load_all_models()
 	except ApplicationError as e:
+		error_logging.warning(str(e))
 		return ApiResponse(success=False, error=e)
-	except Exception:
+	except Exception as e:
+		error_logging.error(str(e))
 		return ApiResponse(success=False, error='unexpected server error')
 
 
@@ -76,7 +79,15 @@ def get_labels_custom(model: str = Form(...)):
 	:param model: Model name or model hash
 	:return: A list of the model's labels with their hashed values
 	"""
-	return dl_service.get_labels_custom(model)
+	try :
+		error_logging.info('request successful;')
+		return dl_service.get_labels_custom(model)
+	except ModelNotFound as e :		
+		error_logging.warning(model+' '+str(e))
+		return ApiResponse(success=False, error=e)
+	except Exception as e:
+		error_logging.error(model + ' ' + str(e))
+		return ApiResponse(success=False, error='unexpected server error')
 
 
 @app.get('/models/{model_name}/load')
@@ -89,9 +100,14 @@ async def load(model_name: str, force: bool = False):
 	"""
 	try:
 		dl_service.load_model(model_name, force)
+		error_logging.info('request successful;')
 		return ApiResponse(success=True)
 	except ApplicationError as e:
+		error_logging.warning(str(e))
 		return ApiResponse(success=False, error=e)
+	except Exception as e:
+		error_logging.error(str(e))
+		return ApiResponse(success=False, error='unexpected server error')
 
 
 @app.get('/models')
@@ -101,7 +117,12 @@ async def list_models(user_agent: str = Header(None)):
 	:param user_agent:
 	:return: APIResponse
 	"""
-	return ApiResponse(data={'models': dl_service.list_models()})
+	try :
+		error_logging.info('request successful;')
+		return ApiResponse(data={'models': dl_service.list_models()})
+	except Exception as e:
+		error_logging.error(str(e))
+		return ApiResponse(success=False, error='unexpected server error')
 
 
 @app.post('/models/{model_name}/predict')
@@ -140,7 +161,6 @@ async def run_model_batch(model_name: str, input_data: List[UploadFile] = File(.
 		error_logging.warning(model_name + ';' + str(e))
 		return ApiResponse(success=False, error=e)
 	except Exception as e:
-		print(e)
 		error_logging.error(model_name + ' ' + str(e))
 		return ApiResponse(success=False, error='unexpected server error')
 
@@ -172,9 +192,16 @@ async def list_model_labels(model_name: str):
 	:param model_name: Model name
 	:return: List of model's labels
 	"""
-	labels = dl_service.get_labels(model_name)
-	return ApiResponse(data=labels)
-
+	try :
+		labels = dl_service.get_labels(model_name)
+		error_logging.info('request successful;' + str(labels))
+		return ApiResponse(data=labels)
+	except ModelNotFound as e :
+		error_logging.warning(model_name + ';' + str(e))
+		return ApiResponse (success=False, error=e)
+	except Exception as e:
+		error_logging.error(model_name + ' ' + str(e))
+		return ApiResponse(success=False, error='unexpected server error')
 
 @app.get('/models/{model_name}/config')
 async def list_model_config(model_name: str):
@@ -183,5 +210,13 @@ async def list_model_config(model_name: str):
 	:param model_name: Model name
 	:return: List of model's configuration
 	"""
-	config = dl_service.get_config(model_name)
-	return ApiResponse(data=config)
+	try:
+		config = dl_service.get_config(model_name)
+		error_logging.info('request successful;' + str(config))
+		return ApiResponse(data=config)
+	except ModelNotFound as e :
+		error_logging.warning(model_name + ';' + str(e))
+		return ApiResponse(success=False, error=e)
+	except Exception as e:
+		error_logging.error(model_name + ' ' + str(e))
+		return ApiResponse(success=False, error='unexpected server error')