Graph Search Algorithm Based Prompt Generation for Ontology Matching.
Note: The repository consists of source codes of the paper "Exploring Prompt Generation utilizing Graph Search Algorithms for Ontology Matching" was accepted by EU Semantics 2024.
├── data
├── results
└── src
├── AlignmentFormat.py
├── accronyms.json
├── batch_loaders
│ ├── alignment.py
│ ├── ontology_parsing
│ └── random_walk.py
├── config.json
├── configMatcher.json
├── configMatcherImport.py
├── globals.py
├── llms
├── maximum_bipartite_matching
├── prompt_template_generator
├── run_matcher.py
├── track.py
├── utilsODS.py
└── verbalizer
use Python version >=3.10
$ pip install -r requirements.txt
on macOS please run
$ python3 -m nltk.downloader stopwords
update dataset paths at src/config.json
adjust the pipeline tasks and algorithm configurations at src/configMatcher.json
$ git clone https://github.com/JulianSampels/OntoMatch.git
download this file and extract it to src/verbalizer/graph2text/outputs/t5-base_13881/
$ cd src
$ python3 run_matcher.py
We conducted t-test with significance level 0.1 to evaluate the significance of the prompt types and algorithms. The results can be seen on t-testfolder.