This dataset is a collection of 326 Tatar language texts with extensive metadata, curated by TatarNLPWorld. Each record includes the full text and a structured metadata object containing fields such as title, author, year, source, genre, category, and more. The dataset is designed for NLP research on the Tatar language, including text classification, language modeling, and digital humanities studies.

Curated by: TatarNLPWorld Community
Language(s) (NLP): Tatar (tt)
License: other – see Licensing & Legal Notice below.

Licensing & Legal Notice

This dataset follows the practice established by large web-crawled corpora such as HPLT and OSCAR:

Original source texts (the text field) remain the property of their respective authors and publishers. They are not owned by the TatarNLPWorld team and are not covered by any open license applied to the metadata.
The metadata, structure, and compilation created by the TatarNLPWorld community are released under the CC BY‑SA 4.0 license.
Users are solely responsible for ensuring their use of the underlying texts complies with applicable copyright law. For commercial use of verbatim excerpts, permission from the original copyright holders may be required.
A notice‑and‑takedown procedure is in place: rights holders can request removal of specific texts by contacting the dataset maintainers (see Dataset Card Contact). We commit to responding within 14 business days and removing disputed content in the next release.

Dataset Sources

Repository: https://huggingface.co/datasets/TatarNLPWorld/tatar-text-corpus (предполагаемый URL; уточните)
Paper: Not applicable; this dataset is a standalone resource.

Uses

Direct Use

The dataset is suitable for:

Text classification and topic modeling using the category and genre metadata.
Language modeling and text generation.
Stylometric and authorship analysis.
Digital humanities research on Tatar literature and media.
Training embeddings for low-resource Tatar NLP.

Out-of-Scope Use

Redistribution of the full texts without permission from original copyright holders.
Any use that violates the rights of the original authors or publishers.
Automated decision-making without human oversight where such decisions could have legal or ethical consequences.

Dataset Structure

Data Fields

Field	Type	Description
`metadata`	struct	Nested object with the following subfields:
`metadata.title`	string	Title of the text
`metadata.lang`	string	Language code (e.g., `tt`)
`metadata.year`	string	Publication year
`metadata.source_original`	string	Original source identifier
`metadata.source_container`	string	Container (e.g., book, website)
`metadata.type`	string	Type of document
`metadata.author`	string	Author name(s)
`metadata.source_url`	string	URL of the source
`metadata.source_site`	string	Site name
`metadata.category`	string	Topic category
`metadata.hash`	string	Hash of the content
`metadata.source_file`	string	Source file name
`metadata.genre`	string	Genre (e.g., news, fiction)
`metadata.source`	string	Combined source label
`text`	string	Full text content in Tatar

Data Splits

Split	Size	Description
`full`	326	Complete dataset
`sample`	50	Small random sample for testing

Dataset Creation

Curation Rationale

This dataset was created to provide a small but richly annotated collection of Tatar texts for experiments in low-resource NLP. The detailed metadata enables research that requires not only text but also contextual information about its origin, authorship, and genre.

Source Data

Data Collection and Processing

Texts were collected from various Tatar-language sources (likely including news sites, literary publications, and digital archives). The metadata was extracted automatically from source HTML or provided by the original publishers. No manual annotation was performed beyond metadata extraction.

Who are the source data producers?

The original texts were authored by journalists, writers, and other content creators. The metadata was compiled by TatarNLPWorld from publicly available information.

Annotations

No additional manual annotations were added. The metadata fields are derived from the source pages or provided by the data providers.

Personal and Sensitive Information

The dataset contains only publicly available texts and metadata. No personal information about individuals (other than author names, which are typically public) is included. If any sensitive data is found, users are encouraged to report it.

Bias, Risks, and Limitations

Small size: With only 326 documents, the dataset is not suitable for training large models from scratch.
Domain bias: The sources and genres may be unevenly represented.
Metadata quality: Some metadata fields may be incomplete or contain errors due to automatic extraction.
Copyright constraints: The full texts are protected; use beyond fair dealing requires permission.

Recommendations

Use this dataset for exploration, prototyping, or as a seed for larger collections.
Combine with other Tatar corpora for training.
Respect original copyrights; cite the original sources when quoting.

Citation

BibTeX (dataset):

@dataset{tatar_text_corpus_2026,
    title = {Tatar Text Corpus with Rich Metadata},
    author = {TatarNLPWorld Community},
    year = {2026},
    publisher = {Hugging Face},
    url = {https://huggingface.co/datasets/TatarNLPWorld/tatar-text-corpus}
}

APA (dataset): TatarNLPWorld Community. (2026). Tatar Text Corpus with Rich Metadata [Data set]. Hugging Face. https://huggingface.co/datasets/TatarNLPWorld/tatar-text-corpus

Glossary

Metadata – structured information describing a text (title, author, date, source, etc.).
Low-resource language – a language with limited digital resources for NLP.

More Information

For questions, contributions, or feedback, please open an issue on the Hugging Face repository or contact the TatarNLPWorld community.

Dataset Card Authors

TatarNLPWorld Community

Dataset Card Contact

Hugging Face: https://huggingface.co/TatarNLPWorld
GitHub: https://github.com/TatarNLPWorld
Email: marabov@kpfu.ru

Downloads last month: 13

Total file size:

3.69 MB