Open Source Projects
This page collects the open-source software, datasets, and other research artifacts published by our chair. It provides a central place to explore and access our publicly available projects and resources.
Software
| github.com/ZijianWang-ZW/IFC2StructuredData | License: MIT |
Convert IFC (Industry Foundation Classes) building models into structured, portable formats: CSV for attributes and relationships, OBJ/MTL for per-element 3D geometry, and optionally GLB for visualization.
| https://github.com/Finradon/ReqREATE | License: MIT |
Semi-automatically synthesize modular AEC models using a Top-Down approach with SysML requirements, double-pushout graph-rewriting, and Grasshopper.
| github.com/tumcms/MomenTUM | License: custom |
The Java software MomenTUM is an agent-based pedestrian simulation framework that was developed under the lead of Dr. Peter M. Kielar at the Chair of Computational Modeling and Simulation at the Technische Universität München.
| github.com/NepomukWolf/IFC-Language-Server | License: MIT |
A language server implementation for IFC STEP files, following the Language Server Protocol (LSP).
| github.com/NepomukWolf/tree-sitter-ifc | License: MIT |
Tree-sitter grammar for IFC STEP files (.ifc).
Tree-sitter is an incremental parser that builds syntax trees efficiently, including during live edits. It is commonly used for syntax highlighting, code navigation, structural queries, and language-server features.
| github.com/tumcms/Open-Infra-Platform | License: GPLv3 |
TUM Open Infra Platform (OIP in short) is a software for checking and viewing IFC data. It comes with its own EXPRESS parser and C++ classes generator.
| github.com/tumcms/TUM.CMS.VPLControl | License: MIT |
TUM.CMS.VplControl is a WPF based Visual Programming Language Control for .Net. A visual programming language can aid non-programmers to write simple programs or processes, without the knowledge of how they are internally built. The control allows simple visual programming (assignments, boolean expressions, math expressions, and scripting) from scratch. Further nodes for your purposes can be added easily. The development of this VPL framework was mainly influenced by the well known VPL tools Dynamo and Grasshopper.
| github.com/tumcms/Blue-Framework | License: Apache 2.0 |
BlueFramework is divided into different parts e.g. Core, ImageProcessing, Rasterizer, and Engine.
The Core contains basic functionality for logging, vector and matrix algebra, string handling and some basic diagnostic functionality.
The ImageProcessing module offers basic functionality for image manipulation and loading and storing image files.
The Rasterizer is a thin abstract layer for graphic APIs such as Direct3D 11, Direct3D 12, OpenGL 3.x and up and Vulkan. Currently, there is a feature complete Direct3D 11 and Direct3D 12 implementation available (BlueFramework.D3D11RenderSystem, BlueFramework.D3D12RenderSystem). The OpenGL backend currently supports only a subset of the complete feature set of the Rasterizer module. The Vulkan implementation is still work in progress. As a user of the BlueFramework library, or more specifically as a user of the BlueFramework.Rasterizer module you program to an abstract interface that is internally mapped to the different render systems. This gives your application the advantage to use different graphic APIs by only writing code once.
The Engine module over functionality to download resources such as textures from URLs, a basic camera for first person and model view orientation, a view cube and basic resource management.
Datasets
| github.com/ZijianWang-ZW/GNI-BIM-Dataset | License: CC-BY-4.0 |
Open-source IFC models under CC-BY-4.0 from BIM students projects at TUM Georg Nemetschek Institute (GNI)
| huggingface.co/datasets/sylvainHellin/ifc-bench | License: CC BY 4.0 |
A benchmark dataset for evaluating BIM (Building Information Modeling) comprehension and reasoning capabilities in AI systems. Provides curated IFC models with question-answer pairs across 4 complexity categories for testing BIM-related AI implementations.
Ontologies
| dtc-ontology.cms.ed.tum.de/ontology/index.html | License: CC BY 4.0 |
The Digital Twin Construction Ontology enables representing the most important concepts that are relevant for a digital twin of a construction site. It explicitly defines the project intent and project status. While the project intent describes the project aim in the form of, e.g., the construction schedule and a 3D building design, the project status describes the situation on the construction site as it was actually observed.