Installation#

There are different methods of installation, depending on your situation. If you aren’t sure which to choose, if you’re a programmer, go for the Pre-built packages. If you aren’t a programmer, go for the BlenderBIM Add-on.

  1. Pre-built packages is recommended for users wanting to use the latest IfcOpenShell builds.

  2. PyPI is recommended for developers using Pip.

  3. Conda is recommended for developers using Anaconda.

  4. Docker is recommended for developers using Docker.

  5. AWS Lambda is recommended for developers using AWS Lambda functions.

  6. Google Colab is recommended for developers using Google Colab.

  7. Web Assembly is recommended for developers experimenting with IfcOpenShell on the web.

  8. Using the BlenderBIM Add-on is recommended for non-developers wanting a graphical interface.

  9. From source with precompiled binaries is recommended for developers actively working with the Python code.

  10. Compiling from source is recommended for developers actively working with the C++ core.

Pre-built packages#

Pre-built packages are prepared sporadically depending on whether there are changes in the IfcOpenShell C++ core. This will give you the latest available C++ bindings, but may potentially contain outdated pure Python modules, such as the API.

  1. Choose which version to download based on your operating system, Python version, and computer architecture.

    Linux 64bit

    Windows 32bit

    Windows 64bit

    MacOS Intel 64bit

    MacOS Silicon 64bit

    Python 3.9

    py39-linux64

    py39-win32

    py39-win64

    py39-macos64

    py39-macosm164

    Python 3.10

    py310-linux64

    py310-win32

    py310-win64

    py310-macos64

    py310-macosm164

    Python 3.11

    py311-linux64

    py311-win32

    py311-win64

    py311-macos64

    py311-macosm164

    Python 3.12

    py312-linux64

    py312-win32

    py312-win64

    py312-macos64

    py312-macosm164

  1. Unzip the downloaded file and copy the ifcopenshell directory into your Python path. If you’re not sure where your Python path is, run the following code in Python:

    import sys
print(sys.path)

    This will give you a list of possible directories that you can install the IfcOpenShell module into. Most commonly, you will want to copy the ifcopenshell directory into one of these called site-packages.

  2. Test importing the module in a Python session or script to make sure it works.

    import ifcopenshell
print(ifcopenshell.version)
model = ifcopenshell.file()

PyPI#

PyPI releases are automatically performed once a month and contain the latest Python code that point in time.

Releases on PyPI may potentially ship slightly outdated precompiled binaries of the C++ core. This is because the binaries typically go through a period of manual testing prior in case of high-risk changes.

pip install ifcopenshell

Conda#

# To install the latest daily build of IfcOpenShell (recommended)
conda install -c ifcopenshell -c conda-forge ifcopenshell
# Alteratively, to install an older, stable version
conda install -c conda-forge ifcopenshell
# If you want to install additional packages that rely on occt, you should install them at the same time into a fresh environment
conda create -n testenv -c conda-forge ifcopenshell pythonocc-core

Note

Installing IfcOpenShell from Conda will also install IfcConvert. Conda also supports Linux AArch64.

Note

If you plan on using any other packages using occt (OpenCascade) you will need to ensure that the occt version installed by ifcopenshell is compatible with the other packages. A common error when using two packages compiled with different occt versions are ImportError: cannot import name ‘V3d’ from ‘OCC’. To avoid this, you should install both packages at the same time and let conda resolve the common occt version. Here is an example where ifcopenshell and pythonocc-core are installed in the same environment testenv conda create -n testenv -c conda-forge ifcopenshell pythonocc-core

Warning

Conda builds are not yet available for Mac ARM devices (M1 chip). Instead, please follow the instructions for the Pre-built packages or PyPI sections above.

Docker#

$ docker run -it aecgeeks/ifcopenshell python3 -c 'import ifcopenshell; print(ifcopenshell.version)'

Note

Installing IfcOpenShell from Docker will also install IfcConvert.

AWS Lambda#

This guide provides a Dockerfile and sample code to help you run an AWS Lambda function written in Python and using the IfcOpenShell library.

This is only a superficial guide, in order to make this work for your usecase you will need a good understanding of other AWS services like SQS, S3, API Gateways, etc.

See also

For more information on building lambda containers refer to the AWS guide on working with Lambda container images

  1. Clone or download the IfcOpenShell repository.

  2. Replace the sample Lambda function code in the example_handler directory with your own code.

  3. Update the import path in the Dockerfile’s CMD instruction to match your Lambda function’s handler function.

  4. Edit the requirements.txt file and add any additional dependencies required by your Lambda function.

  5. Build the Docker image:

    $ docker build -t lambda-ifcopenshell .

  6. Run the Docker container:

    $ docker run lambda-ifcopenshell

  7. Test lambda locally by following the AWS guide on testing Lambda container images locally.

  8. Deploy to lambda. This is beyond the scope of this guide. Please refer to the AWS documentation. Some tools that could be useful are AWS CloudFormaton, AWS CDK, pulumi or terraform.

Google Colab#

The Google Colab environment is based on the distribution from PyPI, but lets you run it in an online notebook without any local setup required. This is great for educators and those wanting to try it out without control on their local system.

Click here to launch a simple notebook.

Web Assembly#

IfcOpenShell is available as technology preview to be run using WASM. This allows you to run IfcOpenShell in a browser using pyodide. This implementation is incredibly heavy and will incur a long load time, but once loaded, will give you full access to the entire IfcOpenShell API.

Click here to learn how to use WASM.

Using the BlenderBIM Add-on#

The BlenderBIM Add-on is a Blender based graphical interface to IfcOpenShell. Other than providing a graphical IFC authoring platform, it also comes with IfcOpenShell, its utilities, and a Python shell built-in. This means you don’t need to install Python first, and you also can compare your IfcOpenShell scripting to what you see with a visual model viewer, or use a graphical interface to access the IfcOpenShell utilities.

The BlenderBIM Add-on is available either as a stable build or a daily build.

  1. Install the BlenderBIM Add-on by following the BlenderBIM Add-on installation documentation.

  2. Launch Blender. On the top left of the Viewport panel, click the Editor Type icon to change the viewport into a Python Console.

    ../_images/blenderbim-python-console-1.png

  3. Make sure you can import IfcOpenShell successfully with the following script.

    ../_images/blenderbim-python-console-2.png

Tip

Before changing the Editor Type to a Python Console, you can click on the View > Area > Vertical Split menu which will divide your viewport. This allows you to write scripts next to the 3D view of a model.

Blender also comes with a text editor so you can write longer scripts. Instead of choosing the Python Console, choose the Text Editor.

../_images/blenderbim-text-editor-1.png

You can now create a new text file for your script by clicking Text > New, and run your script using the Text > Run Script menu or by clicking on the Play Icon.

../_images/blenderbim-text-editor-2.png

See also

You may be interested in learning how to graphically explore an IFC model in Blender. This can help when learning how to write scripts as you can double check the results of your scripts with what you see in the graphical interface. Read more.

From source with precompiled binaries#

  1. Clone or download the IfcOpenShell repository.

  2. Place src/ifcopenshell-python/ifcopenshell/ in your Python path.

  3. Download the relevant pre-built package, and copy over the two files with ifcopenshell_wrapper in their filename to src/ifcopenshell-python/ifcopenshell/.

Compiling from source#

Advanced developers may want to compile IfcOpenShell. Refer to the IfcOpenShell installation guide for instructions.

Legacy versions#

It is not recommended to use older versions of IfcOpenShell, but for historic reasons you may find them here: