Merge pull request #40 from intel/v0.4.0-alpha-preview

v0.4.0-alpha-preview

.gitignore

@@ -3,4 +3,4 @@ __pycache__/
 *.generated
 docker/Dockerfile.env
 docker/final.env
-
+models

.pylintrc

@@ -263,7 +263,9 @@ ignore-on-opaque-inference=yes
 # List of class names for which member attributes should not be checked (useful
 # for classes with dynamically set attributes). This supports the use of
 # qualified names.
-ignored-classes=optparse.Values,thread._local,_thread._local
+ignored-classes=optparse.Values,thread._local,_thread._local,
+                extension__pb2.MediaStreamMessage,
+                MediaStreamMessage	
 
 # List of module names for which member attributes should not be checked
 # (useful for modules/projects where namespaces are manipulated during runtime

Attribution.txt

@@ -49,7 +49,7 @@ https://packages.debian.org/sid/python3-setuptools
 six (MIT) https://pypi.org/project/six/
 https://github.com/benjaminp/six/blob/master/LICENSE
 
-Python 3 (Custom) https://packages.debian.org/sid/python3
+Python 3 (PSF) https://packages.debian.org/sid/python3
 https://docs.python.org/3/license.html
 
 pip3 (MIT) https://packages.debian.org/sid/python3-pip
@@ -90,3 +90,42 @@ https://gstreamer.freedesktop.org/src/gstreamer-vaapi/
 
 Numpy (BSD 3-clause) https://github.com/numpy/numpy
 https://github.com/numpy/numpy/blob/master/LICENSE.txt
+
+jsonschema (MIT) https://github.com/Julian/jsonschema/
+https://github.com/Julian/jsonschema/blob/main/COPYING
+
+strict-rfc3339 (GPLv3) https://pypi.org/project/strict-rfc3339/
+https://pypi.org/project/strict-rfc3339/
+
+pyyaml (MIT) https://pypi.org/project/strict-rfc3339/
+https://pypi.org/project/strict-rfc3339/
+
+protobuf (BSD 3-clause) https://github.com/protocolbuffers/protobuf
+https://github.com/protocolbuffers/protobuf/blob/master/LICENSE
+
+pillow (PIL) https://pillow.readthedocs.io/en/latest/handbook/index.html
+https://github.com/python-pillow/Pillow/blob/master/LICENSE
+
+gunicorn (MIT) http://docs.gunicorn.org/
+https://pypi.org/project/gunicorn/
+
+grpcio (Apache 2.0) https://grpc.io/docs/guides/
+https://pypi.org/project/grpcio/
+
+rfc3987 (GPLv3) https://pypi.org/project/rfc3987/
+https://pypi.org/project/rfc3987/
+
+Live Video Analytics (MIT) https://azure.microsoft.com/en-us/services/media-services/live-video-analytics/
+https://github.com/Azure/live-video-analytics/blob/master/LICENSE
+
+RunIt (three-clause BSD alike license) http://smarden.org/runit
+http://smarden.org/runit/faq.html#license
+
+nginx (BSD 2-clause) http://nginx.org/
+https://nginx.org/LICENSE
+
+libnginx-mod-nchan (MIT) https://nchan.io/
+https://github.com/slact/nchan/blob/master/LICENCE
+
+python-psutil (BSD 3-clause) https://psutil.readthedocs.io/en/latest/
+https://github.com/giampaolo/psutil/blob/master/LICENSE

README.md

@@ -283,7 +283,7 @@ Pipeline Instance Id
 Start a new shell and execute the following command to issue an HTTP POST request, start a pipeline and analyze a sample [video](https://github.com/intel-iot-devkit/sample-videos/blob/master/preview/head-pose-face-detection-male.gif).
 
 ```bash
-curl localhost:8080/pipelines/object_detection/1 -X POST -H \
+curl localhost:8080/pipelines/emotion_recognition/1 -X POST -H \
 'Content-Type: application/json' -d \
 '{ 
   "source": {
@@ -467,9 +467,9 @@ After pretty-printing:
 </details>
 
 # Further Reading
-| **Documentation** | **Reference Guides** |
-| ------------    | ------------------ | 
-| **-** [Defining Media Analytics Pipelines](docs/defining_pipelines.md) <br/> **-** [Building Video Analytics Serving](docs/building_video_analytics_serving.md) <br/> **-** [Running Video Analytics Serving](docs/running_video_analytics_serving.md) | **-** [Video Analytics Serving Architecture Diagram](docs/images/video_analytics_service_architecture.png) <br/> **-** [Microservice Endpoints](docs/restful_microservice_interfaces.md) <br/> **-** [Build Script Reference](docs/build_script_reference.md) <br/> **-** [Run Script Reference](docs/run_script_reference.md) |
+| **Documentation** | **Reference Guides** | **Tutorials** |
+| ------------    | ------------------ | ----------- |
+| **-** [Defining Media Analytics Pipelines](docs/defining_pipelines.md) <br/> **-** [Building Video Analytics Serving](docs/building_video_analytics_serving.md) <br/> **-** [Running Video Analytics Serving](docs/running_video_analytics_serving.md) <br/> **-** [Customizing Pipeline Requests](docs/customizing_pipeline_requests.md) | **-** [Video Analytics Serving Architecture Diagram](docs/images/video_analytics_service_architecture.png) <br/> **-** [Microservice Endpoints](docs/restful_microservice_interfaces.md) <br/> **-** [Build Script Reference](docs/build_script_reference.md) <br/> **-** [Run Script Reference](docs/run_script_reference.md) | <br/> **-** Object Detecion Tutorials <br/> &nbsp;&nbsp;&nbsp;&nbsp; **-** [Changing Object Detection Models](docs/changing_object_detection_models.md) |
 
 ## Related Links
 

docker/Dockerfile

@@ -22,6 +22,11 @@ RUN echo "IMPORTANT: This script will build third party components licensed unde
 # Switching to root user to create user and install packages
 USER root
 
+# Dependencies for OpenVINO
+RUN if [ -f /opt/intel/openvino/install_dependencies/install_openvino_dependencies.sh ]; then \
+       /opt/intel/openvino/install_dependencies/install_openvino_dependencies.sh ;\
+    fi    
+
 # Creating user vaserving and adding it to groups "video" and "users" to use GPU and VPU
 ARG USER=vaserving
 RUN useradd -ms /bin/bash -G video,audio,users ${USER} -d /home/video-analytics-serving && \
@@ -30,9 +35,9 @@ RUN useradd -ms /bin/bash -G video,audio,users ${USER} -d /home/video-analytics-
 # Dependencies installed via apt-get
 RUN DEBIAN_FRONTEND=noninteractive apt-get update && \
     apt-get install -y -q --no-install-recommends \
+    python3 \
     python3-setuptools \
-    python3-pip \
-    python3 && \
+    python3-pip && \
     rm -rf /var/lib/apt/lists/*
 
 # Dependencies installed via pip
@@ -42,6 +47,7 @@ RUN rm -f /requirements.txt
 
 # Video Analytics Serving Python Modules
 COPY ./vaserving /home/video-analytics-serving/vaserving
+COPY --chown=vaserving ./tools /home/video-analytics-serving/tools
 
 # Media Analytics Framework set via environment variable
 ENV FRAMEWORK=${FRAMEWORK}
@@ -104,4 +110,6 @@ FROM ${FINAL_STAGE} as deploy
 
 ARG USER=vaserving
 
+ENV PYTHONPATH=$PYTHONPATH:/home/video-analytics-serving
+
 USER ${USER}

docker/build.sh

@@ -5,27 +5,35 @@
 # SPDX-License-Identifier: BSD-3-Clause
 #
 
+
+DOCKERFILE_DIR=$(dirname "$(readlink -f "$0")")
+SOURCE_DIR=$(dirname $DOCKERFILE_DIR)
+
 BASE_IMAGE=
 BASE_BUILD_CONTEXT=
 BASE_BUILD_DOCKERFILE=
 BASE_BUILD_TAG=
 USER_BASE_BUILD_ARGS=
-MODELS=models
+MODELS=$SOURCE_DIR/models_list/models.list.yml
+MODELS_PATH=models
 PIPELINES=
 FRAMEWORK=
 TAG=
 RUN_PREFIX=
 CREATE_SERVICE=TRUE
 ENVIRONMENT_FILES=()
+DOCKER_RUN_ENVIRONMENT=$(env | cut -f1 -d= | grep -E '_(proxy)$' | sed 's/^/-e / ' | tr '\n' ' ')
 TARGET="deploy"
 
-DOCKERFILE_DIR=$(dirname "$(readlink -f "$0")")
-SOURCE_DIR=$(dirname $DOCKERFILE_DIR)
+
 BUILD_ARGS=$(env | cut -f1 -d= | grep -E '_(proxy|REPO|VER)$' | sed 's/^/--build-arg / ' | tr '\n' ' ')
 BASE_BUILD_ARGS=$(env | cut -f1 -d= | grep -E '_(proxy|REPO|VER)$' | sed 's/^/--build-arg / ' | tr '\n' ' ')
 BUILD_OPTIONS="--network=host "
 BASE_BUILD_OPTIONS="--network=host "
 
+OPEN_MODEL_ZOO_VERSION=
+FORCE_MODEL_DOWNLOAD=
+
 DEFAULT_GSTREAMER_BASE_BUILD_CONTEXT="https://github.com/opencv/gst-video-analytics.git#0d95b489f03d30bd94ffa25a2472eaab7b774821"
 DEFAULT_GSTREAMER_BASE_BUILD_DOCKERFILE="docker/Dockerfile"
 DEFAULT_GSTREAMER_BASE_BUILD_TAG="video-analytics-serving-gstreamer-base"
@@ -93,12 +101,23 @@ get_options() {
             ;;
         --models)
             if [ "$2" ]; then
-                MODELS=$2
+                MODELS=$(realpath $2)
                 shift
             else
                 error 'ERROR: "--models" requires an argument.'
             fi
             ;;
+        --force-model-download)
+            FORCE_MODEL_DOWNLOAD="--force"
+            ;;
+        --open-model-zoo-version)
+            if [ "$2" ]; then
+                OPEN_MODEL_ZOO_VERSION=$2
+                shift
+            else
+                error 'ERROR: "--openvino-version" requires an argument.'
+            fi
+            ;;
         --pipelines)
             if [ "$2" ]; then
                 PIPELINES=$2
@@ -155,7 +174,7 @@ get_options() {
                 error 'ERROR: "--create-service" requires an argument.'
             fi
             ;;
-	--environment-file)
+	    --environment-file)
             if [ "$2" ]; then
                 ENVIRONMENT_FILES+=($2)
                 shift
@@ -175,10 +194,10 @@ get_options() {
             shift
             break
             ;;
-         -?*)
+        -?*)
 	    error 'ERROR: Unknown option: ' $1
             ;;
-	 ?*)
+        ?*)
 	    error 'ERROR: Unknown option: ' $1
             ;;
         *)
@@ -200,6 +219,33 @@ get_options() {
         show_help
     fi
 
+    if [ -f "$MODELS" ]; then
+        if [ -z "$OPEN_MODEL_ZOO_VERSION" ] && [ $FRAMEWORK = 'ffmpeg' ]; then
+            OPEN_MODEL_ZOO_VERSION=2020.3
+        elif [ -z "$OPEN_MODEL_ZOO_VERSION" ]; then
+            OPEN_MODEL_ZOO_VERSION=2020.4
+        fi
+        YML_DIR=$(dirname "${MODELS}")
+        YML_FILE_NAME=$(basename "${MODELS}")
+        VOLUME_MOUNT+="-v $SOURCE_DIR:/home/video-analytics-serving -v $YML_DIR:/models_yml"
+        $RUN_PREFIX docker run -t --rm $DOCKER_RUN_ENVIRONMENT --entrypoint /bin/bash $VOLUME_MOUNT openvino/ubuntu18_data_dev:$OPEN_MODEL_ZOO_VERSION "-i" "-c" "pip3 install jsonschema==3.2.0; python3 /home/video-analytics-serving/tools/model_downloader --model-list /models_yml/$YML_FILE_NAME --output-dir /home/video-analytics-serving/ $FORCE_MODEL_DOWNLOAD"
+
+        #TODO: remove below if condition once aclnet model downloads with model downloader
+        if [ $MODELS = $(realpath $SOURCE_DIR/models_list/models.list.yml) ] && [ -d $SOURCE_DIR/models_list/audio_detection ] ; then
+            if [ ! -d "$SOURCE_DIR/models/audio_detection" ]; then
+                mkdir $SOURCE_DIR/models/audio_detection
+            fi
+            cp -R $SOURCE_DIR/models_list/audio_detection/. $SOURCE_DIR/models/audio_detection
+        fi    
+    elif [ -d "$MODELS" ]; then
+        if [ ! -d "$SOURCE_DIR/models" ]; then
+            mkdir $SOURCE_DIR/models
+        fi
+        cp -R $MODELS/. $SOURCE_DIR/models
+    else
+        error 'ERROR: "'$MODELS'" does not exist.'
+    fi
+
     if [ -z "$PIPELINES" ]; then
         PIPELINES=pipelines/$FRAMEWORK
     elif [ "${PIPELINES^^}" == "NONE" ]; then
@@ -261,6 +307,7 @@ show_image_options() {
     echo "   Build Options: '${BUILD_OPTIONS}'"
     echo "   Build Arguments: '${BUILD_ARGS}'"
     echo "   Models: '${MODELS}'"
+    echo "   Openvino Image: 'openvino/ubuntu18_data_dev:${OPEN_MODEL_ZOO_VERSION}'"
     echo "   Pipelines: '${PIPELINES}'"
     echo "   Framework: '${FRAMEWORK}'"
     echo "   Target: '${TARGET}'"
@@ -273,7 +320,9 @@ show_help() {
     echo "usage: build.sh"
     echo "  [--base base image]"
     echo "  [--framework ffmpeg || gstreamer]"
-    echo "  [--models path to model directory relative to $SOURCE_DIR or NONE]"
+    echo "  [--models path to models directory or model list file or NONE]"
+    echo "  [--open-model-zoo-version specify the version of openvino image to be used for downloading models from Open Model Zoo]"
+    echo "  [--force-model-download force the download of models from Open Model Zoo]"
     echo "  [--pipelines path to pipelines directory relative to $SOURCE_DIR or NONE]"
     echo "  [--build-option additional docker build option that run in the context of docker build. ex. --no-cache]"
     echo "  [--base-build-option additional docker build option for docker build of base image]"
@@ -309,15 +358,15 @@ if [ "$BASE" == "BUILD" ]; then
     BASE_IMAGE=$BASE_BUILD_TAG
 else
     #Ensure image is latest from Docker Hub
-    docker pull $BASE_IMAGE
+    $RUN_PREFIX docker pull $BASE_IMAGE
 fi
 
 # BUILD IMAGE
 
 BUILD_ARGS+=" --build-arg BASE=$BASE_IMAGE "
 BUILD_ARGS+=" --build-arg FRAMEWORK=$FRAMEWORK "
 if [ -n "$MODELS" ]; then
-    BUILD_ARGS+="--build-arg MODELS_PATH=$MODELS "
+    BUILD_ARGS+="--build-arg MODELS_PATH=$MODELS_PATH "
     BUILD_ARGS+="--build-arg MODELS_COMMAND=copy_models "
 else
     BUILD_ARGS+="--build-arg MODELS_COMMAND=do_not_copy_models "
@@ -340,7 +389,7 @@ cp -f $DOCKERFILE_DIR/Dockerfile $DOCKERFILE_DIR/Dockerfile.env
 ENVIRONMENT_FILE_LIST=
 
 if [[ "$BASE_IMAGE" == "openvino/"* ]]; then
-    $RUN_PREFIX docker run -t --rm --entrypoint /bin/bash -e HOSTNAME=BASE $BASE_IMAGE "-i" "-c" "env" > $DOCKERFILE_DIR/openvino_base_environment.txt
+    $RUN_PREFIX docker run -t --rm $DOCKER_RUN_ENVIRONMENT --entrypoint /bin/bash -e HOSTNAME=BASE $BASE_IMAGE "-i" "-c" "env" > $DOCKERFILE_DIR/openvino_base_environment.txt
     ENVIRONMENT_FILE_LIST+="$DOCKERFILE_DIR/openvino_base_environment.txt "
 fi
 

docker/run.sh

@@ -192,6 +192,7 @@ show_options() {
     echo "   Entrypoint: '${ENTRYPOINT}'"
     echo "   EntrypointArgs: '${ENTRYPOINT_ARGS}"
     echo "   User: '${USER}'"
+    echo "   Devices: '${DEVICES}'"
     echo ""
 }
 
@@ -208,6 +209,7 @@ show_help() {
   echo "  [--network name network to pass to docker run]"
   echo "  [--user name of user to pass to docker run]"
   echo "  [--name container name to pass to docker run]"
+  echo "  [--device device to pass to docker run]"
   echo "  [--dev run in developer mode]"  
   exit 0
 }

docs/build_script_reference.md

@@ -8,7 +8,7 @@ $ docker/build.sh --help
 usage: build.sh
   [--base base image]
   [--framework ffmpeg || gstreamer]
-  [--models path to models directory relative to /path/to/video-analytics-serving or NONE]
+  [--models path to models directory or model list file or NONE]
   [--pipelines path to pipelines directory relative to /path/to/Projects/video-analytics-serving or NONE]
   [--base-build-context docker context for building base image]
   [--base-build-dockerfile docker file used to build base image]
@@ -38,8 +38,8 @@ This option is used in conjunction with `--base-build-context` to specify the do
 ## Pipeline Directory (--pipelines)
 Relative path to VA Serving pipelines directory from root of video-analytics-serving project. If not specified, [sample pipelines](../pipelines/gstreamer) are included in the image. If set to `NONE` no pipelines are included and the user must ensure pipelines are made available at runtime by volume mounting.
 
-## Model Directory (--models)
-Models used by pipelines are expected to be in this directory. It is a relative path from root of video-analytics-serving project. If not specified, models for sample pipelines are included in the image. If set to `NONE` no models are included and the user must ensure models are made available at runtime by volume mounting.
+## Model Directory/File List (--models)
+This option can be used to specify path to models directory or a model list file. When its a directory, models used by pipelines are expected to be in this directory. When its a file, the models listed in the file are downloaded and converted to IR format if needed by the [model download tool](../tools/model_downloader/README.md) during build time. If nothing is specified, default models listed in the file `models_list/models.list.yml` are downloaded, converted to IR format if needed and included in the image. If set to `NONE` no models are included and the user must ensure models are made available at runtime by volume mounting.
 
 ## Build Arg (--build-arg)
 Specify a docker build argument when building the VA Serving image.