Skip to content

Latest commit

 

History

History
295 lines (227 loc) · 7.04 KB

BUILD.md

File metadata and controls

295 lines (227 loc) · 7.04 KB
Raw

Getting Started

The simplest way to build Aria Data Tools is either to use:

Requirements:

  • C++17 compiler
  • Python3 (>=3.8 recommended)

Using pre-compiled package (pip install)

In a nutshell 
pip install projectaria-tools==0.0.3

python
>>> import projectaria_tools
>>> dir(projectaria_tools)
['__doc__', '__loader__', '__name__', '__package__', '__spec__', 'dataprovider', 'mps_io', 'sensors']
>>> quit()

Using isolated/virtual environment 
# Here’s an e2e example that:
# - creates a virtual environment in python3.9
# - installs `projectaria_tools` pkg and imports it

python3.9 -m venv projectaria_env
source ./projectaria_env/bin/activate
pip install projectaria_tools
python
>>> import projectaria_tools
>>> dir(projectaria_tools)
['__doc__', '__loader__', '__name__', '__package__', '__spec__', 'dataprovider', 'mps_io', 'sensors']
>>> quit()
deactivate
rm -rf ./projectaria_env

Using build from source

We are detailing here instruction to install Aria Data Tools on Debian/Fedora/MacOS systems. The build process is separated in 2 steps:

  1. Installing 3rd party requirements 2.a Cloning, compiling Aria Data Tools [C++] 2.b Cloning, compiling Aria Data Tools [Python]

1. Installing 3rd party requirements

Pick your platform and follow along the provided instructions:

Debian/Ubuntu 
# Install build essentials && python pip essentials
sudo apt-get install -y cmake git build-essential
sudo apt-get install -y libpython3-dev python3-pip

# Install VRS dependencies
sudo apt-get install -y libgtest-dev libgmock-dev \
    libfmt-dev  \
    libturbojpeg-dev libpng-dev \
    liblz4-dev libzstd-dev libxxhash-dev \
    libboost-system-dev libboost-filesystem-dev libboost-thread-dev libboost-chrono-dev libboost-date-time-dev

# Installing cereal library (header only) to get the last version
cd /tmp
git clone https://github.com/USCiLab/cereal.git -b v1.3.2
cd cereal
cmake -DSKIP_PORTABILITY_TEST=1 -DJUST_INSTALL_CEREAL=ON .
sudo make -j2 install
rm -rf /tmp/cereal
Fedora 
# Install build essentials
sudo dnf install -y git cmake gcc gcc-c++ make

# Install python (or use your system config)
sudo dnf install -y python3-devel; pip3 install --upgrade pip; pip3 install pybind11[global] numpy

# Install VRS dependencies
sudo dnf install -y gtest-devel gmock-devel glog-devel \
                 fmt-devel lz4-devel libzstd-devel xxhash-devel \
                 boost-devel libpng-devel libjpeg-turbo-devel turbojpeg-devel;

# Installing cereal library (header only) to get the last version
cd /tmp
git clone https://github.com/USCiLab/cereal.git -b v1.3.2
cd cereal
cmake -DSKIP_PORTABILITY_TEST=1 -DJUST_INSTALL_CEREAL=ON .
sudo make -j2 install
rm -rf /tmp/cereal
MacOS 
# Install build essentials with Homebrew
brew install git cmake

# Install python (or use your system config)
brew install python3

# Install VRS dependencies
brew install boost fmt sophus cereal googletest glog lz4 zstd xxhash libpng jpeg-turbo

2.a Cloning, compiling Aria Data Tools [C++]

Debian/Ubuntu/Fedora/MacOS 
# Clone the project
git clone https://github.com/facebookresearch/Aria_data_tools.git --recursive
mkdir build
cmake -DCMAKE_BUILD_TYPE=RELEASE -S ./Aria_data_tools/src -B build
cd build
make -j

2.b Cloning, compiling Aria Data Tools [Python]

Debian/Ubuntu/Fedora/MacOS 
# Go to source folder & build/install python bindings
cd Aria_data_tools/src
pip3 install --global-option=build_ext --global-option="-j2" .;

# Testing successful install
python
>>> import projectaria.tools
>>> dir(projectaria.tools.datatools)
['__doc__', '__loader__', '__name__', '__package__', '__spec__', 'dataprovider', 'mpsIO', 'sensors']

A note on Visualization samples

Visualization samples using Pangolin are optional and can be de/activated easily at the cmake configure step -DBUILD_WITH_PANGOLIN=ON. Here are the instructions to setup Pangolin and its internal dependencies.

1. Installing Pangolin (minimum dependencies)

Ubuntu/Debian 
sudo apt-get install -y libeigen3-dev libglew-dev libgl1-mesa-dev
Fedora 
sudo dnf install -y glew-devel eigen3
MacOS 
brew install eigen glew

2. Clone, build and install Pangolin

Debian/Ubuntu/Fedora/MacOS 
cd /tmp
git clone https://github.com/stevenlovegrove/Pangolin.git -b v0.8
# Package installed above should be enough,
# FYI you can also install all Pangolin REQUIRED dependencies by using
# "./Pangolin/scripts/install_prerequisites.sh required"
mkdir Pangolin_Build && cd Pangolin_Build
cmake -DCMAKE_BUILD_TYPE=Release -DBUILD_TOOLS=OFF -DBUILD_PANGOLIN_PYTHON=OFF -DBUILD_EXAMPLES=OFF ../Pangolin/
sudo make -j install

3. Compiling Aria Data Tools with Pangolin

Debian/Ubuntu/Fedora/MacOS 
# Cloning the project
git clone https://github.com/facebookresearch/Aria_data_tools.git --recursive
# Building the project
mkdir build
cmake -DCMAKE_BUILD_TYPE=RELEASE -DBUILD_WITH_PANGOLIN=ON  -S ./Aria_data_tools/src -B build
cd build
make -j
# Testing a visualization sample
./visualization/aria_viewer ../Aria_data_tools/data/aria_unit_test_sequence_calib.vrs

Container (Build and run in isolation)

Here are the instruction to compile and run Aria Data Tools in isolation in a container. Podman or Docker can be used.

Debian/Ubuntu/Fedora/MacOS 
git clone https://github.com/facebookresearch/Aria_data_tools.git --recursive

cd Aria_data_tools/src

podman build . -t aria_data_tools

podman run -it aria_data_tools

Working with container

X11 forwarding

  • If you setup allow X11 forwarding you can see GUI elements from your container on your host machine (tested on Linux):
podman run -it -e DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix localhost/aria_data_tools:latest

Mapping volume to use HOST data in your container

  • Adding volume(s) in your container allow you to quickly iterate on code or local data:
podman run -it --volume <SOME_PATH>/aria_data_tools/:/tmp/src_aria --volume '<SOME_PATH>Documents/TestingData':/tmp/data_test aria_data_tools