File size: 3,563 Bytes
a31d217 dac812f a31d217 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 | ## Dataset Card for `Text2Sparql-Clean`
### 🧾 Dataset Summary
`Text2Sparql-Clean` is a **clean, standardized, and multilingual dataset** for the task of translating natural language questions into SPARQL queries over the **DBpedia** knowledge graph. It aggregates and harmonizes four widely-used benchmarks in the text-to-SPARQL research field:
* **QALD (versions 1–9)**
* **LC-QuAD 1.0**
* **Orange/paraqa-sparqltotext**
* **julioc-p/Question-Sparql**
The dataset includes questions in **English** and **Spanish**, making it a valuable resource for both **monolingual** and **cross-lingual** question answering over knowledge graphs.
This is the **second official version** of the dataset.
---
### 🛠️ Key Features and Improvements
This dataset goes beyond simple aggregation. It includes several **important improvements** over the original datasets:
* ✅ **Standardized SPARQL queries**: All queries were revised to **include the necessary PREFIX declarations**, ensuring compatibility with DBpedia’s SPARQL endpoint.
* ✅ **Syntactic and semantic validation**: Every SPARQL query was checked for **syntactic correctness** and **semantic executability**. Faulty or malformed queries were either **fixed** or **removed**.
* ✅ **Clean, high-quality corpus**: The result is a **clean dataset** where all queries are **both executable and meaningful**, enabling reliable model training and evaluation.
---
### ✅ Supported Tasks
* **Text-to-SPARQL generation**
→ Given a natural language question, generate the corresponding SPARQL query over DBpedia.
---
### 🌍 Languages
* **English** (`en`)
* **Spanish** (`es`)
---
### 📁 Dataset Format
Each data instance contains:
* `question`: The question in natural language (English or Spanish)
* `query`: The corresponding SPARQL query
* `dataset-id`: Original dataset source (e.g., `QALD-X`, `LCQUAD`, `Orange`, `JULIOC`)
---
### 🔧 Dataset Creation Process
This dataset is the result of a **comprehensive curation pipeline**:
1. **Aggregation** of four public datasets into a unified structure.
* **QALD 1-9**: https://github.com/ag-sc/QALD
* **LC-QuAD 1.0**: https://github.com/AskNowQA/LC-QuAD
* **Paraqa-SparqlToText**: https://huggingface.co/datasets/Orange/paraqa-sparqltotext
* **Question-Sparql**: https://huggingface.co/datasets/julioc-p/Question-Sparql
2. **Normalization** of query formats and syntax, resulting in all queries with PREFIX.
3. **Insertion of missing PREFIX declarations** to ensure compliance with DBpedia standards.
4. **Execution validation**: All queries were tested against a DBpedia SPARQL endpoint. Queries that failed were either corrected (if trivial) or removed.
5. **Language consistency checks** for both English and Spanish questions.
---
### 💡 Use Cases
* Training and evaluating **text-to-SPARQL models**
* Developing **multilingual semantic parsers**
* Enabling **cross-lingual question answering** over structured knowledge graphs
* Fine-tuning large language models on **knowledge-graph-to-text** tasks
---
### 📚 Citation
If you use this dataset in your research, please cite:
```
@misc{text2sparql-clean,
author = {Marcos Gôlo, Paulo do Carmo, Edgard Marx, Ricardo Marcacini},
title = {text2sparql-clean: Natural Language Text to SPARQL for DBpedia Cleaned Dataset},
year = {2025},
howpublished = {\url{https://huggingface.co/datasets/aksw/Text2Sparql-Clean}},
}
```
---
### 🪪 Licensing
license: apache-2.0
Check each dataset’s original repository for full licensing terms.
--- |