This folder contains various transformation tools that can be used in conjunction with each other and other components of the UD-Graph to create data transformation pipelines.
These transformations are a product of the following works:
[1] D. Vinasco-Alvarez, “Leveraging Standards in Model-Centric Geospatial Knowledge Graph Creation,” in ESWC 2022 Ph.D. Symposium, Hersonissos, Greece, May 2022. [Online]. Available: https://hal.archives-ouvertes.fr/hal-03693607
[2] D. Vinasco-Alvarez, J. S. Samuel, S. Servigne, and G. Gesquière, “Towards Limiting Semantic Data Loss In 4D Urban Data Semantic Graph Generation,” ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, vol. VIII-4/W2-2021, pp. 37–44, Oct. 2021, doi: 10.5194/isprs-annals-VIII-4-W2-2021-37-2021.
[3] D. Vinasco-Alvarez, J. S. Samuel, S. Servigne, and G. Gesquière, “Towards a semantic web representation from a 3D geospatial urban data model,” in SAGEO 2021, 16ème Conférence Internationale de la Géomatique, de l’Analyse Spatiale et des Sciences de l’Information Géographique., La Rochelle [Online Event], France, May 2021, pp. 227–238. [Online]. Available: https://hal.archives-ouvertes.fr/hal-03240567
[4] D. Vinasco-Alvarez, J. S. Samuel, S. Servigne, and G. Gesquière, “From CityGML to OWL,” LIRIS UMR 5205, Technical Report, Sep. 2020. [Online]. Available: https://hal.archives-ouvertes.fr/hal-02948955
For more information on each proposed transformation tool or transformation tool configuration can be found in each transformation sub-directory:
- ShapeChange (UML-to-OWL): Contains example ShapeChange configurations for converting UML Models based on the CityGML standard to OWL ontologies. In addition, a small script is provided for ensuring that the provided ontologies are compliant to the OWL-DL OWL sublanguage (to enable reasoning over these ontologies). These configurations are a product of the work [2]
- XSD-to-OWL: Contains an XSLT-based transformation tool for converting XML Schema to OWL ontologies. This transformation is a product of the works [3, 4]
- XML-to-RDF: Contains two transformation tools
- XML2RDF.py for executing syntactic transformations between XML files and OWL/RDF graphs using OWL ontologies and namespace mappings. This transformation is also a product of [2].
- A similar XSLT-based transformation script for creating a XML to RDF XSLT stylesheet with an XML Schema for XML data conform to that schema. This transformation is also a product of the works [3, 4].
- These pipelines use utility scripts located in the utility directory.
- DifferenceGraph-to-CityGML3: Contains a script to convert a
DifferencesAsGraph.jsonfile (produced by the Change Detection stage of the cityGMLto3DTiles transformation pipeline) to conform to a Version Transition conformant to the CityGML 3.0 Version Module in OWL/RDF.
Reproducible examples of these workflows are in development on UD-Reproducibility
These workflows are illustrated below using the CityGML standard with Workspace and Document ADEs using their models, schemas, and data stored in the test-data directory. Note that the produced data is experimental and may not be fully compliant with the CityGML standard.
This proposed workflow is composed of 3 activities:
| Activity | Description | Component | Example Input Data | Example Output Data |
|---|---|---|---|---|
| 1. UML to OWL transformation | Transform a UML model to OWL/RDF. The generated ontologie(s) may be conformant to OWL-Full but not OWL-DL. An ontologyPatcher script is provided to fix identified OWL-DL inconsistencies that can result from ShapeChange UML to OWL transformations. | CityGML 3.0 Ontology Network | ||
| 2. XML to RDF conversion | Transform XML data (conformant to the UML model) to OWL/RDF using a set of ontologies. | XML2RDF | CityGML 3.0 RDF/OWL Individuals | |
| 3. Combine Model and Data | This activity is a formality to illustrate that the resulting ontology network and the generated individuals can be combined into 1 file composed of the ABox (or model) and the TBox (or data instances) of the ontology network |
This proposed workflow is composed of 3 activities:
| Activity | Description | Component | Example Input Data | Example Output Data |
|---|---|---|---|---|
| XSD Compilation | Combine XSD files into a single XSD file | XSD combination script | CityGML 2.0 XML Schema | Composite CityGML 2.0 Schema file |
| Convert CityGML Model to OWL | Convert the composite schema to RDF/OWL | XSLT Transformation | Composite CityGML 2.0 Schema file | CityGML 2.0 Ontology |
| Convert CityGML Instances to RDF | Convert XML data conform to the XML Schema to RDF/OWL. This activity is decomposed in 3 subactivities illustrated below: |
XSLT-based transformation workflow | CityGML 2.0 RDF/OWL Individuals |
Which workflow is recommended? In summary, the answer depends on what models and schema are available and the use-case for integration or transformation. In articles [2,3] we discuss XSD to OWL vs UML to OWL referencing [3,4]. Both approaches have limitations and it's unclear which is general better and if one is better for a specific use case. However, our UML-based workflow is currently more mature and is the current workflow we recommend using.
From a theoretical perspective:
To broadly summarize:
| XSLT approach [3] | ShapeChange+XML2RDF approach [2] | |
|---|---|---|
| Pros | ||
| Cons |
There are more notes about this on the project wiki
From a technical perspective:
- Our implementation of [2] is more mature than our implementation of [3]
- Since publication of [3], the work in [7] has further developed the approach
[1] D. Vinasco-Alvarez, “Leveraging Standards in Model-Centric Geospatial Knowledge Graph Creation,” in ESWC 2022 Ph.D. Symposium, Hersonissos, Greece, May 2022. [Online]. Available: https://hal.archives-ouvertes.fr/hal-03693607
[2] D. Vinasco-Alvarez, J. S. Samuel, S. Servigne, and G. Gesquière, “Towards Limiting Semantic Data Loss In 4D Urban Data Semantic Graph Generation,” ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, vol. VIII-4/W2-2021, pp. 37–44, Oct. 2021, doi: 10.5194/isprs-annals-VIII-4-W2-2021-37-2021.
[3] D. Vinasco-Alvarez, J. S. Samuel, S. Servigne, and G. Gesquière, “Towards a semantic web representation from a 3D geospatial urban data model,” in SAGEO 2021, 16ème Conférence Internationale de la Géomatique, de l’Analyse Spatiale et des Sciences de l’Information Géographique., La Rochelle [Online Event], France, May 2021, pp. 227–238. [Online]. Available: https://hal.archives-ouvertes.fr/hal-03240567
[4] D. Vinasco-Alvarez, J. S. Samuel, S. Servigne, and G. Gesquière, “From CityGML to OWL,” LIRIS UMR 5205, Technical Report, Sep. 2020. [Online]. Available: https://hal.archives-ouvertes.fr/hal-02948955
[5] S. J. D. Cox, “An explicit OWL representation of ISO/OGC Observations and Measurements,” presented at the Proceedings of the 6th International Workshop on Semantic Sensor Networks co-located with the 12th International Semantic Web Conference, Jan. 2013, pp. 1–18.
[6] L. Brink, P. Janssen, W. Quak, and J. Stoter, “Linking spatial data: automated conversion of geo-information models and GML data to RDF,” International Journal of Spatial Data Infrastructures Research, vol. 9, pp. 59–85, Oct. 2014, doi: 10.2902/1725-0463.2014.09.art3.
[7] A. U. Usmani, M. Jadidi, and G. Sohn, “Towards the Automatic Ontology Generation and Alignment of BIM and GIS Data Formats,” in ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Oct. 2021, vol. VIII-4-W2-2021, pp. 183–188. doi: 10.5194/isprs-annals-VIII-4-W2-2021-183-2021.