databat-kgc-revit

Knowledge Graph construction from the Revit 3D model of a building

The scripts in this repository are used to generate a knowledge graph with entry point https://ci.mines-stetienne.fr/emse

This knowledge publicly documents the model of Espace Fauriel Building at Mines Saint-Étienne, using the Building Topology Ontology.

How to reproduce

1. Export the IFC model from the Revit 2021 model rvt/MINES_TCE_DOE_2021.rvt, using the IFC for Revit add-in exporter and export setup rvt/IFCExportConfiguration_IFC4 Design Transfer View.json.
2. Run the IFC-to-LBD converter, with options as shown in rvt/IFCtoLBD-Desktop_Java_15_options_used.png
3. Execute the Dynamo script rvt/RoomAdjacenciesByLevel.dyn to generate file schedules/adjacency_guid.csv that encodes the adjacencies between rooms.
4. Export the schedules whose name begin with "Ifc" into the schedules folder. With only option Export column header checked, and option Field delimiter set to ,.
5. Run the scripts in sequence:
   1. scripts/1_repareCharacters.py restores unicode characters obfuscated by the IFC-to-LBD converter
   2. scripts/2_pruneObjects.py only keeps objects that have the following string in their IRI: site, building, storey, space, wall, window, ifcowl_ifcfurniture, electricappliance.
   3. scripts/3_transformObjects.py explicitly types the objects, keeps only a few relevant triples from the ones generated by IFC-to-LBD, selects appropriate properties, and converts the objects.
   4. scripts/4_renameObjects.py changes the IRIs of objects to IRIs that better reflects the topology of the building.
   5. scripts/5_readSchedules.py used to extract important information that have been lost somewhere in the Revit-to-IFC-to-LBD pipeline, including the location of furnitures and devices, and which spaces are adjacent to windows and doors.
   6. scripts/6_addProperties.py used to define some SOSA/SSN properties for objects: temperature, CO2, humidity of spaces, and open/close status of doors and windows.
   7. scripts/7_storeTurtleFiles.py used to generate the set of turtle files that can then be pushed to the server

Step 4. can be run with the following sequence of commands.

python3 -m pip install rdflib
mkdir temp
python3 scripts/1_repareCharacters.py lbd/MINES_TCE_DOE_2021_LBD.ttl temp/0_MINES_TCE_DOE_2021_LBD.ttl
python3 scripts/2_pruneObjects.py temp/0_MINES_TCE_DOE_2021_LBD.ttl temp/1_objects_pruned.ttl https://ci.mines-stetienne.fr/
python3 scripts/3_transformObjects.py temp/1_objects_pruned.ttl temp/2_objects_transformed.ttl https://ci.mines-stetienne.fr/
python3 scripts/4_renameObjects.py temp/2_objects_transformed.ttl temp/3_objects_renamed.ttl https://ci.mines-stetienne.fr/
python3 scripts/5_readSchedules.py temp/3_objects_renamed.ttl temp/4_schedules.ttl https://ci.mines-stetienne.fr/
python3 scripts/6_addProperties.py temp/4_schedules.ttl temp/5_properties.ttl https://ci.mines-stetienne.fr/
python3 scripts/7_storeTurtleFiles.py temp/5_properties.ttl https://ci.mines-stetienne.fr/