nos_gl_CC0
Copyright-free (CC0) sentences in Galician, collected by Proxecto Nós in order to contribute to the text corpus of Mozilla Common Voice.
The rights to the original sentences were transferred to the University of Santiago de Compostela (USC) by public or private entities, to whom we are grateful for their contribution.
On this material, Proxecto Nos applied a series of transformations which resulted in the corpora herein mentioned. The original sentences were filtered for length and, where necessary, adapted to meet the Common Voice requirements (normalization of numbers, abbreviations, and acronyms, foreign word elimination, appropriateness, grammatical correctness, etc.).
Preprocessing and processing of the corpora was done using dedicated scripts available [here](https://github.com/proxectonos/nos_gl_CC0/tree/main/Scripts], as well as the Sentence Extractor tool.
The resulting corpora, published here copyright-free, are the following:
PG CORPUS
cv_pg_anon.txt
166.229 sentences extracted from transcrips of the debates of the Galician Parliament (Parlamento de Galicia) over a 7-year timespan (2015-2022). The Galician Parliament allows public access to the original material and authorizes its reproduction provided the source is aknowledged. Apart from the transformations mentioned above, the names present in the corpus were anonymised.
CRTVG CORPUS
cv_tvg.txt
166.951 sentences extracted from the 2019-2022 transcripts of the daily news broadcast of Televisión de Galicia (TVG).
NÓS DIARIO (SERMOS GALIZA) CORPUS
cv_nos_diario.txt
140.104 sentences extracted from the 2012-2022 issues of the newspaper Nós Diario.
PRAZA.GAL CORPUS
cv_praza.txt
89.688 sentences extracted from the 2012-2022 issues of the newspaper Praza.gal.
CORPUS CV_GL_2025
Corpus formed by 305.322 sentences collected by the Proxecto Nós from various sources, which is integrated by different subsets.
Corpus Mix Media
More than 190.000 sentences extracted from publications ceded to the Proxecto Nós of the University of Santiago de Compostela through data transfer agreements:
- 84.106 sentences from the Gran Enciclopedia Galega ceded by the Xunta de Galicia and from texts ceded by the Consello da Cultura Galega.
- 7.988 sentences from novels ceded by the Hugín & Muñín publisher.
- 12.944 sentences from the publication "Tempos Dixital" and the magazine "A Trabe de Ouro" ceded by the Atlántica Comunicación entity.
- 28.949 sentences from publications ceded by the Editorial Galixia.
- 12.918 sentences from the magazine "Tempo Exterior" and other web content ceded by the Instituto Galego de Documentación (IGADI).
- 692 sentences from CLUVI corpus, ceded by the Instituto da Lingua Galega (USC).
- 14.492 sentences from USC Publications, specifically from the Classics Collection.
- 6.873 sentences from USC Periodical Publications (magazines, articles, etc.).
- 22.257 sentences from various sources.
Corpus Diario Dixital Que pasa na costa
Subcorpus formed by 86.652 sentences extracted from the Costa da Morte Quepasanacosta Digital Newspaper.
Corpus DRAG
Subset of 28.051 sentences extracted from the definitions of terms published in the Dicionario da Real Academia Galega (DRAG).
Other resources
Within the framework of ILENIA, the Proxecto Nós also participated in the creation of the multilingual Escagleu dataset, which can be consulted here.
Funding and Acknowledgments
This dataset was produced within the framework of the Proxecto Nós, funded by the Ministry for Digital Transformation and Public Administration and the Recovery, Transformation, and Resilience Plan – Funded by the European Union – NextGenerationEU, as part of the Ilenia Project with reference 2022/TL22/00215336. It also received support from the Investigo program (PRTR - NextGenerationEU).
We would like to thank the Parliament of Galicia, Nós Diario, Praza.gal, and the Corporación Radio e Televisión de Galicia (CRTVG) for providing the original data.
Cite this dataset
If you use this dataset, please cite it as follows:
Adina Ioana Vladu, Noelia García Díaz, Marta Vázquez Abuín, Patricia Ramos Maceiras, María Janeiro Cerqueiro, Cristian Canosa Pérez, José Javier Francisco Marini, Elisa Fernández Rei (2022). nos_gl_CC0
URL: (https://huggingface.co/datasets/proxectonos/nos_gl_CC0)
Also on GitHub