Matcha-DL is a tool for supervised ontology alignment. It takes as input two ontologies and a set of mappings between them, and learns a model that can predict mappings between the two ontologies.
Project description
Matcha DL
Description
Matcha-DL is an extension of the matching system Matcha to tackle semi-supervised tasks using machine-learning algorithms. Matcha builds upon the algorithms of the established system AgreementMakerLight with a novel broader core architecture designed to tackle long-standing challenges such as complex and holistic ontology matching. Matcha-DL uses a linear neural network that learns to rank candidate mappings proposed by Matcha by using a partial reference alignment as a training set, and using the confidence scores produced by Matcha's matching algorithms as features.
Installation
To install Matcha DL, you can use pip:
pip install matcha-dl
USAGE
CLI
Matcha DL provides a command line interface for computing the alignment between two ontologies. Here's how you can use it:
matchadl --source_ontology_file <source_file_path> --target_ontology_file <target_file_path> --output_dir <output_dir_path> [--reference_file <reference_file_path>] [--candidates_file <candidates_file_path>] [--config_file <config_file_path>]
API
AlignmentRunner Class
The AlignmentRunner
is a quick and easy class that can be programaticly imported to run an alignemnt using Matcha-DL.
Here's an example on how to use it:
from matcha_dl import AlignmentRunner
runner = AlignmentRunner(
source_ontology_file="path/to/source_ontology_file",
target_ontology_file="path/to/target_ontology_file",
output_dir="path/to/output_dir",
reference_file="path/to/reference_file",
candidates_file="path/to/candidates_file",
config_file="path/to/config_file"
)
runner.run()
Arguments
- --source_ontology_file or -s: Path to the source ontology file (required)
- --target_ontology_file or -t: Path to the target ontology file (required)
- --output_dir or -o: Path to the output directory (required)
- --reference_file or -r: Path to the reference file (optional)
- --candidates_file or -c: Path to the candidates file (optional)
- --config_file or -C: Path to the config file (optional)
Details
- The reference file should be a reference alignment, that follows the standards from the OAEI's Bio-ML track.
- The candidates file should be a list of all the candidates for each of the entities to rank, that follows the standards from the OAEI's Bio-ML track.
- The configuration file is a user-defined configuration file. An example can be found at https://github.com/liseda-lab/Matcha-DL/blob/main/matcha_dl/default_config.yaml.
Tasks
Supervised/Unsupervised settings
If reference files are provided Matcha-DL will train a model to predict an alignment, otherwise it will use the scores from matcha to compute the alignment directly.
Global Alignment/ Local Alignment
If a candidates file is provided Matcha-DL will generate a ranking for those candidates (local alignment), otherwise it will perform global pairwise alignemnt for all the entities in the source and target ontologies.
Acknowledgements
This work was supported by FCT through the fellowships 2022.10557.BD (Pedro Cotovio) and 2022.11895.BD (Marta Silva), and through the LASIGE Research Unit, ref. UIDB/00408/2020 (https://doi.org/10.54499/UIDB/00408/2020) and ref. UIDP/00408/2020 (https://doi.org/10.54499/UIDP/00408/2020). It was also partially supported by the KATY project which has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No 101017453. This work was also supported partially by project 41, HfPT: Health from Portugal, funded by the Portuguese Plano de Recuperação e Resiliência.
Contributions
Authors:
Pedro Giesteira Cotovio, [1], Lucas Ferraz, [1], Daniel Faria, [2], Laura Balbi, [1], Marta Contreiras Silva, [1], Catia Pesquita, [1]
Institutions:
- LASIGE, Faculdade de Ciências, Universidade de Lisboa, Portugal
- INESC-ID, Instituto Superior Técnico, Universidade de Lisboa, Portugal
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for matcha_dl-0.1.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 15a5bdcb44295d903915c6188cead8ab787d67279f754362d29765f80e262140 |
|
MD5 | 5f5e6185925f91f6c5826cf967a828d2 |
|
BLAKE2b-256 | 1a7fa178ad0009e81029cca2df42661d8fd150574006240772c4045b4111aab3 |