WSE-research / QADO-Dataset-Pre-Processor

Web service to transform an unstructured QA benchmark dataset into a semi-structured format that can be transformed by the QADO RDFizer

QADO Dataset Pre-Processor

This repository provides a webservice to transform an unstructured QA benchmark dataset into a semi-structured format that can be transformed by the QADO RDFizer.

Installation

Per default the service will be available at http://localhost:5000. You can use Docker or Python to run the service.

Using Docker

To run the prebuild docker image execute the following commands:

docker pull bigoli98/qado-dataset-preprocessor:latest
docker run -itd -p 5000:5000 bigoli98/qado-dataset-preprocessor:latest

Using Python directly

This script requires Python 3.9 or later. To execute the Python script directly, just run the following commands:

pip install -r requirements.txt
python app.py

Endpoints

The service provides the following endpoints to process data:

• /process/compositional_wikidata (POST): This endpoint transforms a Compositional Wikidata Questions dataset. The payload has to be

{
  "fetch_url": "URL for Plain text data",
  "language": "Language of the dataset provided by fetch_url"
}

This endpoint returns the data transformed into JSON.

[
  {
    "id": 0,
    "question": "Dataset question",
    "query": "SPARQL query used to answer the question",
    "lang": "language of the question"
  }
]