Skip to content

BUILD_FROM_SOURCE.md

Latest commit

 

History

History
148 lines (107 loc) · 6.71 KB

BUILD_FROM_SOURCE.md

File metadata and controls

148 lines (107 loc) · 6.71 KB

PIP install

git clone --branch stable https://github.com/facebookresearch/habitat-sim.git
cd habitat-sim
pip install . --no-build-isolation

  • You can also allow pip to compile a specific version of Habitat. First clone the repo, then pip install . --no-build-isolation in the current git root directory to start the compilation process. To quickly compile the latest main, run pip install git+https://github.com/facebookresearch/habitat-sim.

  • Since pip builds out of tree by default, this process will copy quite a lot of data to your TMPDIR. You can change this location by modifying the TMPDIR env variable. For active development, use an editable install: pip install -e . --no-build-isolation.

  • By default, we build with GUI viewer support and Bullet physics enabled. Override any option via environment variables:

    HABITAT_BUILD_GUI_VIEWERS=OFF pip install . --no-build-isolation   # headless build (no GUI)
HABITAT_WITH_BULLET=OFF pip install . --no-build-isolation         # disable Bullet physics
HABITAT_WITH_CUDA=ON pip install . --no-build-isolation            # enable CUDA

Build from Source

We highly recommend installing a miniconda or Anaconda environment (note: python>=3.9 is required). Once you have Anaconda installed, here are the instructions.

  1. Clone this github repository.

    # Checkout the latest stable release
git clone --branch stable https://github.com/facebookresearch/habitat-sim.git
cd habitat-sim

    List of stable releases is available here. Main branch contains 'bleeding edge' code and under active development.

  2. Install Dependencies

    Common

    # We require python>=3.12 and cmake>=3.22
conda create -n habitat python=3.12 cmake=3.27
conda activate habitat
pip install -r requirements.txt

    Linux (Tested with Ubuntu 18.04 with gcc 7.4.0)

    sudo apt-get update || true
# These are fairly ubiquitous packages and your system likely has them already,
# but if not, let's get the essentials for EGL support:
sudo apt-get install -y --no-install-recommends \
     libjpeg-dev libglm-dev libgl1-mesa-glx libegl1-mesa-dev mesa-utils xorg-dev freeglut3-dev

    See the github actions and workflow configs for a full list of dependencies that our CI installs on a clean Ubuntu VM. If you run into build errors later, this is a good place to check if all dependencies are installed.

  3. Build Habitat-Sim

    Default build (includes GUI viewers and Bullet physics)

    # Assuming we're still within habitat conda environment
pip install . --no-build-isolation

    For headless systems (i.e. without an attached display, e.g. in a cluster)

    HABITAT_BUILD_GUI_VIEWERS=OFF pip install . --no-build-isolation

    For systems with CUDA (to build CUDA features)

    HABITAT_WITH_CUDA=ON pip install . --no-build-isolation

    With audio sensor via rlr-audio-propagation: To use Audio sensors (Linux only), enable the audio flag via:

    HABITAT_WITH_AUDIO=ON pip install . --no-build-isolation

    Note1: Build options stack via environment variables, e.g. to build in headless mode with CUDA:

    HABITAT_BUILD_GUI_VIEWERS=OFF HABITAT_WITH_CUDA=ON pip install . --no-build-isolation

    Note2: some Linux distributions might require an additional --user flag to deal with permission issues.

    Note3: for active development in Habitat, use an editable install (pip install -e . --no-build-isolation) or ./build.sh which wraps the editable install with convenience flags.

    Note4: Audio sensor is only available on Linux.

    Build Configuration Reference

    Environment Variable Default Description
    HABITAT_BUILD_GUI_VIEWERS ON Build GUI viewer applications (set OFF for headless)
    HABITAT_WITH_BULLET ON Enable Bullet physics simulation
    HABITAT_WITH_CUDA OFF Enable CUDA support
    HABITAT_WITH_AUDIO OFF Enable audio sensor (Linux only)
    HABITAT_LTO OFF Enable link-time optimization
    HABITAT_BUILD_TESTS OFF Build C++ tests

    You can also pass CMake arguments directly via pip's --config-settings:

    pip install . --no-build-isolation --config-settings=cmake.define.OPTION=VALUE

    C++ Viewer and Replayer Applications

    Since HABITAT_BUILD_GUI_VIEWERS is ON by default, the C++ viewer and replayer applications are compiled alongside the Python bindings in a standard build.

    • After pip install .: the executables are installed on PATH as viewer and replayer.
    • After pip install -e . (editable/development builds): convenience symlinks are created at build/viewer and build/replayer in the repo root.
    • Standalone CMake (no Python): you can build the C++ apps directly: 
      cmake -B build -S src \
  -DBUILD_GUI_VIEWERS=ON \
  -DBUILD_PYTHON_BINDINGS=OFF \
  -DCMAKE_BUILD_TYPE=RelWithDebInfo
cmake --build build -j$(nproc)
# Binaries at: build/utils/viewer/viewer and build/utils/replayer/replayer

Common build issues

  • If your machine has a custom installation location for the nvidia OpenGL and EGL drivers, you may need to manually provide the EGL_LIBRARY path to cmake as follows. Add -DEGL_LIBRARY=/usr/lib/x86_64-linux-gnu/nvidia-opengl/libEGL.so via --config-settings=cmake.define.EGL_LIBRARY=/usr/lib/x86_64-linux-gnu/nvidia-opengl/libEGL.so. When running any executable adjust the environment as follows: LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu/nvidia-opengl:${LD_LIBRARY_PATH} examples/example.py.

  • By default, the build process uses all cores available on the system to parallelize. On some virtual machines, this might result in running out of memory. You can limit parallelism via:

    pip install . --no-build-isolation --config-settings=cmake.define.CMAKE_BUILD_PARALLEL_LEVEL=1

  • Build is tested on Tested with Ubuntu 18.04 with gcc 7.4.0 and MacOS 10.13.6 with Xcode 10 and clang-1000.10.25.5. If you experience compilation issues, please open an issue with the details of your OS and compiler versions.

    We also have a dev slack channel, please follow this link to get added to the channel.