Refactor CV-SDK

Refactor examples to use the new Computer Vision solution

This change refactors the examples to use the new Computer Vision
solution instead of the old one (CV SDK). It also includes various
other updates and improvements, such as:

- Update the README files with relevant information
- Update the acap-runtime version to 2.0.0 (container based)
- Add QEMU to the GitHub workflow for building

Author: Corallo

.github/workflows/example-checks.yml

@@ -15,11 +15,17 @@ jobs:
         run: |
           # Find directories that do not start with dot
           acap_list="$(find . -maxdepth 1 -type d -printf '%f\n' |grep -v '^[.]' | sort)"
-          # Find lines starting with '* ['
-          # The grep removes anything but '[*]', then remove entries with space
-          # or capital letter and finally removes the brackets.
-          endline="$(sed -n '/## Docker Hub images/=' README.md)"
-          readme_examples="$(sed -n "1,${endline}!d ; /* \[/p" README.md | grep -oe '\[.*\]' | grep -v '[A-Z ]' | grep -oe '[0-9a-z-]*')"
+
+          # Find example application links in README.md
+          # The regular expression '^\* \[[0-9a-z-]*\]' matches lines starting with '* ['
+          # followed by alphanumeric characters and hyphens, and a closing ']'
+          example_links="$(grep -o '^\* \[[0-9a-z-]*\]' README.md)"
+
+          # Remove '* [' prefix and trailing ']' from example links
+          # The sed command uses two substitutions:
+          # 1. 's/^\* \[//' removes the '* [' prefix from the start of the line
+          # 2. 's/\]$//' removes the trailing ']' from the end of the line
+          readme_examples="$(echo "$example_links" | sed 's/^\* \[//;s/\]$//')"
 
           # Control that all examples have an entry in README
           common=$(printf "${acap_list}\n${readme_examples}" | sort | uniq -d)

.github/workflows/hello-world-python.yml

@@ -25,4 +25,5 @@ jobs:
           imagetag: ${{ env.EXREPO }}_${{ env.EXNAME }}:1.0
         run: |
           cd $EXNAME
+          docker run --rm --privileged multiarch/qemu-user-static --credential yes --persistent yes
           docker build --no-cache --tag $imagetag .

.github/workflows/object-detector-python.yml

@@ -31,5 +31,6 @@ jobs:
           modeltag: ${{ env.EXREPO }}_${{ env.EXNAME }}-${{ matrix.chip }}-model:1.0
         run: |
           cd $EXNAME
+          docker run --rm --privileged multiarch/qemu-user-static --credential yes --persistent yes
           docker build --no-cache --build-arg CHIP=${{ matrix.chip }} --tag $imagetag .
           docker build --file Dockerfile.model --tag $modeltag  .

.github/workflows/opencv-qr-decoder-python.yml

@@ -25,4 +25,5 @@ jobs:
           imagetag: ${{ env.EXREPO }}_${{ env.EXNAME }}:1.0
         run: |
           cd $EXNAME
+          docker run --rm --privileged multiarch/qemu-user-static --credential yes --persistent yes
           docker build --no-cache --tag $imagetag .

.github/workflows/parameter-api-python.yml

@@ -25,4 +25,5 @@ jobs:
           imagetag: ${{ env.EXREPO }}_${{ env.EXNAME }}:1.0
         run: |
           cd $EXNAME
+          docker run --rm --privileged multiarch/qemu-user-static --credential yes --persistent yes
           docker build --no-cache --tag $imagetag .

README.md

@@ -32,15 +32,22 @@ developers analytics experience. All examples are using Docker framework and has
 README file in its directory which shows overview, example directory structure and
 step-by-step instructions on how to run applications on the camera.
 
-### Requirements
+## Requirements
 
 ### Supported architectures
 
 The examples support the following architectures:
 
 * aarch64
 
-### Example applications for video analytics
+### Required components
+
+The example applications make use of two special components:
+
+* [Docker ACAP](https://github.com/AxisCommunications/docker-acap) needs to be installed and running on the Axis device. It's required to deploy and run the example application.
+* [ACAP runtime](https://github.com/AxisCommunications/acap-runtime) is used in a few examples to get access to [APIs](https://axiscommunications.github.io/acap-documentation/docs/api/computer-vision-sdk-apis.html). ACAP Runtime is a container image and the example applications will pull it from Docker Hub when starting.
+
+## Example applications
 
 Below is a list of examples available in the repository:
 
@@ -60,12 +67,6 @@ Below is a list of examples available in the repository:
 * [web-server](./web-server/)
   * A C++ example which runs a Monkey web server on the camera.
 
-### Docker Hub images
-
-The examples are based on the [ACAP Computer Vision SDK](https://github.com/AxisCommunications/acap-computer-vision-sdk).
-This SDK is an image which contains APIs and tooling to build computer vision apps for running on camera, with support for Python.
-Additionally, there is the [ACAP Native SDK](https://github.com/AxisCommunications/acap-native-sdk), which is more geared towards building ACAPs that uses AXIS-developed APIs directly, and primarily does so using C/C++.
-
 ## How to work with Github repository
 
 You can help to make this repo a better one using the following commands.

hello-world-python/Dockerfile

@@ -1,17 +1,15 @@
 # syntax=docker/dockerfile:1
 
-ARG REPO=axisecp
-ARG SDK_VERSION=1.15
 ARG UBUNTU_VERSION=22.04
 
-
-FROM ${REPO}/acap-computer-vision-sdk:${SDK_VERSION}-aarch64 AS cv-sdk
-
 FROM arm64v8/ubuntu:${UBUNTU_VERSION}
 
-# Get the Python package from the CV SDK
-COPY --from=cv-sdk /axis/python /
+RUN DEBIAN_FRONTEND=noninteractive \
+    apt-get update && apt-get install -y --no-install-recommends \
+    python3 \
+    && rm -rf /var/lib/apt/lists/*
+
+WORKDIR /opt/app
 
-WORKDIR /app
-COPY app/* /app
+COPY app/* .
 CMD ["python3", "simply_hello.py"]

hello-world-python/README.md

@@ -2,9 +2,9 @@
 
 # A minimal Python application
 
-This example demonstrates how to create a simple Python application using the ACAP Computer Vision SDK and run it on an edge device.
+This example demonstrates how to create a simple containerized Python application and run it on an edge device.
 
-Going from zero to a Python application running on an AXIS device is quite easy. First, the application script is written, as in the hello-world script in [app/simply_hello.py](app/simply_hello.py). Next, the [Dockerfile](Dockerfile) which builds the application image is constructed. This needs to pull in packages from the ACAP Computer Vision SDK, as is done using the `COPY` commands. Finally, the application needs to be built and uploaded, as is specified below.
+Going from zero to a Python application running on an Axis device is quite easy. First, the application script is written, as in the hello-world script in [app/simply_hello.py](app/simply_hello.py). Next, the [Dockerfile](Dockerfile) which builds the application image is constructed. The application needs to be built and uploaded, as is specified below.
 
 ## Example structure
 

minimal-ml-inference/Dockerfile

@@ -1,24 +1,36 @@
 # syntax=docker/dockerfile:1
 
+ARG RUNTIME_VERSION=2.0.0
 ARG REPO=axisecp
-ARG SDK_VERSION=1.15
+ARG IMAGE=acap-runtime
 ARG UBUNTU_VERSION=22.04
 
-# Specify which ACAP Computer Vision SDK to use
-FROM ${REPO}/acap-computer-vision-sdk:${SDK_VERSION}-aarch64 AS cv-sdk
+FROM ${REPO}/${IMAGE}:${RUNTIME_VERSION}-protofiles AS proto-image
+FROM arm64v8/ubuntu:${UBUNTU_VERSION} AS runtime-image
 
-# Define the runtime image
-FROM arm64v8/ubuntu:${UBUNTU_VERSION}
+RUN DEBIAN_FRONTEND=noninteractive \
+    apt-get update && apt-get install -y --no-install-recommends \
+    python3 \
+    python3-pip \
+    ffmpeg \
+    libsm6 \
+    libxext6 \
+    && rm -rf /var/lib/apt/lists/*
 
-# Get packages from the CV SDK
-COPY --from=cv-sdk /axis/python /
-COPY --from=cv-sdk /axis/python-tfserving /
-COPY --from=cv-sdk /axis/python-vdoproto /
-COPY --from=cv-sdk /axis/python-numpy /
-COPY --from=cv-sdk /axis/opencv /
-COPY --from=cv-sdk /axis/openblas /
+# Copy list of grpc dependencies
+COPY --from=proto-image /build/requirements.txt .
+
+RUN pip install --no-cache-dir -r requirements.txt \
+    numpy \
+    opencv-python
+
+WORKDIR /opt/app
+
+# Install prebuilt TensorFlow and VDO proto files
+COPY --from=proto-image /build/tf/proto_utils .
+COPY --from=proto-image /build/vdo/proto_utils .
 
 # Copy the application script to the container
-WORKDIR /app
-COPY simple_inference.py .
+COPY app/* .
+
 CMD ["python3", "simple_inference.py"]

minimal-ml-inference/Dockerfile.model

@@ -4,7 +4,6 @@ ARG UBUNTU_VERSION=22.04
 
 FROM arm64v8/alpine as model-image
 
-
 # Get SSD Mobilenet V2
 ADD https://github.com/google-coral/edgetpu/raw/master/test_data/ssd_mobilenet_v2_coco_quant_postprocess.tflite models/
 ADD https://github.com/google-coral/edgetpu/raw/master/test_data/coco_labels.txt models/