RasaPi Running via Docker

Rasa Open Source
1

Hello community and Docker aficionados!

@stephens did a fantastic start on trying to get a Docker image of the RasaPi system running on the Raspberry Pi. Here is his write-up:

He needs help getting it running - seems like there is some issues with opencv-python.

To replicate the errors, run the following as a script:

sudo echo "https://apt.dockerproject.org/repo/ raspbian-RELEASE main"
sudo apt-get install -y docker.io
sudo apt install python3-pip
sudo pip3 install docker-compose

Here is his Dockerfile:

# time sudo docker build -t rasapi:balena .
#FROM balenalib/raspberry-pi-debian:buster
# time sudo docker build -t rasapi:buster .
FROM debian:buster

RUN apt-get update -qq && \
apt-get install -y --no-install-recommends \
  libbz2-dev \
  libssl-dev \
  xz-utils \
  wget \
  curl \
  build-essential \
  cmake \
  unzip \
  pkg-config \
  zlib1g-dev \
  apt-utils \
  sudo

# Get and install Python3.6
RUN wget --no-check-certificate https://www.python.org/ftp/python/3.6.8/Python-3.6.8.tar.xz && \
  tar -xvf Python-3.6.8.tar.xz && \
  cd Python-3.6.8 && \
  ./configure && \
  make -j4 && \
  make install

# Update Python3.6 packages
RUN python3.6 -m pip install --upgrade pip setuptools --user

# Install additional dependencies
# Enables access to Tensorflow whl
# Dependency for the h5py python package
RUN apt-get install python3-pip libhdf5-dev -y

RUN python3.6 --version

# Install Tensorflow
RUN python3.6 -m pip install tensorflow==1.14.0 -v --user --extra-index-url https://www.piwheels.org/simple

RUN  apt-get install -y --no-install-recommends \
  libjpeg-dev libpng-dev libtiff-dev \
  libavcodec-dev libavformat-dev libswscale-dev libv4l-dev \
  libxvidcore-dev libx264-dev \
  libgtk-3-dev \
  libcanberra-gtk* \
  libatlas-base-dev gfortran \
  python3-dev

WORKDIR /root

# Download OpenCV and clarify naming scheme
RUN wget --no-check-certificate -O opencv.zip https://github.com/opencv/opencv/archive/4.0.0.zip && \
  wget --no-check-certificate -O opencv_contrib.zip https://github.com/opencv/opencv_contrib/archive/4.0.0.zip && \
  ls -l

RUN unzip opencv.zip && \
  unzip opencv_contrib.zip && \
  mv opencv-4.0.0 opencv && \
  mv opencv_contrib-4.0.0 opencv_contrib

RUN ls -l

RUN cd opencv && \
  ls -l && \
  mkdir build && \
  cd build && \
  cmake -D CMAKE_BUILD_TYPE=RELEASE \
    -D CMAKE_INSTALL_PREFIX=/usr/local \
    -D OPENCV_EXTRA_MODULES_PATH=/root/opencv_contrib/modules \
    -D ENABLE_NEON=ON \
    -D ENABLE_VFPV3=ON \
    -D BUILD_TESTS=OFF \
    -D OPENCV_ENABLE_NONFREE=ON \
    -D INSTALL_PYTHON_EXAMPLES=OFF \
    -D BUILD_EXAMPLES=OFF .. && \
  make -j4 && \
  sudo make install && \
  sudo ldconfig

RUN apt-get install -y --no-install-recommends \
  dphys-swapfile \
  git

#RUN sudo /bin/sh -c "/sbin/dphys-swapfile setup"
#RUN cat /etc/dphys-swapfile

# Increasing swap size to make OpenCV
#RUN SWAPSIZE=2048 && \
#  sed -i "s/^CONF_SWAPSIZE.*/CONF_SWAPSIZE=${SWAPSIZE}/" /etc/dphys-swapfile && \
#  sudo /etc/init.d/dphys-swapfile stop && \
#  sudo /etc/init.d/dphys-swapfile start

# Link cv2 to Python3.6
RUN cd /usr/local/lib/python3.6/site-packages/ && \
  sudo ln -s /usr/local/python/cv2/python-3.6/cv2.cpython-36m-arm-linux-gnueabihf.so cv2.so

RUN cd ~ && \
  git clone https://github.com/explosion/spaCy && \
  git clone https://github.com/tensorflow/tensor2tensor && \
  git clone https://github.com/google/dopamine.git && \
  wget --no-check-certificate https://github.com/RasaHQ/rasa/archive/1.4.0.zip && unzip 1.4.0.zip

RUN sudo echo "[global]" > /etc/pip.conf && \
  sudo echo "extra-index-url=https://www.piwheels.org/simple" >> /etc/pip.conf && \
  sudo cat /etc/pip.conf

WORKDIR /root/dopamine

# Installing dopamine-rl
RUN cd ~/dopamine && \
  sed -i '/opencv-python/d' setup.py

RUN  echo grep opencf && \
  find . -exec grep -l "opencv" {} \; && \
  pwd && \
  ls -l
RUN  pwd && \
  cat setup.py && \
  python3.6 -m pip install . --user

# Installing tensor2tensor
RUN cd ~/tensor2tensor && \
  sed -i '/opencv-python/d' setup.py && \
  sed -i '/dopamine-rl/d' setup.py && \
  python3.6 -m pip install . --user --force-reinstall

# Installing other RASA dependencies
RUN sudo apt install libpq-dev/buster -y  && \
  python3.6 -m pip install psycopg2 --user

# Installing RASA
RUN cd ~/rasa-1.4.0 && \
  sed -i '/tensor2tensor/d' setup.py && \
  sed -i '/tensor2tensor/d' requirements.txt && \
  python3.6 -m pip install -r requirements.txt --user --force-reinstall && \
  python3.6 -m pip install . --user --force-reinstall

Finally, after saving this Dockerfile, run:

sudo docker build .

Can anyone help with running this Dockerfile and replicating on the Pi’s? Looping in @Tobias_Wochinger and @JulianGerhard to see if you know of anyone else who can help with this.

Thanks and anyone who helps here, let us know what you find out! I am going to be trying this at some point to see what I can uncover.

TheCedarPrince

1 Like
2

Like I mentioned in Slack I’ll work on this some over the weekend when I get home and have access to my pi’s.

Thanks again both of ya for this work, it’s gonna be a cool project IMO.

2 Likes
3

Hey @btotharye,

Awesome. Just a heads-up, the entire method has only been tested on the RPi4 with 4GB of RAM. Other Pi models might not be strong enough to handle this.

TheCedarPrince

4

ok cool I’ll test them just so we can confirm I have a mixture of different rpi’s at the house anyways.

1 Like
5

Hi @btotharye,

just to complete what @TheCedarPrince has written so far: Feel free to get back to me if there are any questions. I have tested everything in advance on a Raspi3 aswell. However what @TheCedarPrince wrote somewhere is true: don’t expect it to be a quick installation :smiley:

If you want to see the RasPi Prince wrote about in action you can ask Karen for the video.

Kind regards and have a nice weekend both of you Julian

1 Like