Skip to content

Latest commit

 

History

History
127 lines (96 loc) · 4.81 KB

README.md

File metadata and controls

127 lines (96 loc) · 4.81 KB
Raw

Summary

Russian data from the SynTagRus corpus.

Introduction

The SynTagRus dependency treebank is being developed by the Computational Linguistics Laboratory, A.A.Kharkevich Institute of Information Transmission Problems, Russian Academy of Sciences, located in Moscow.

Currently the treebank contains over 1,000,000 tokens (over 66,000 sentences) belonging to texts from a variety of genres (contemporary fiction, popular science, newspaper and journal articles dated between 1960 and 2016, texts of online news etc.)

SynTagRus is a human-corrected corpus of Russian supplied with comprehensive morphological annotation and syntactic annotation in the form of a complete dependency tree provided for every sentence. Additionally, the original version of SynTagRus contains other types of annotation, first of all lexical functional annotation in terms of lexical functions as defined in the Meaning-Text model.

It is an integral but fully autonomous part of the Russian National Corpus developed in a nationwide research project and can be freely consulted on the Web: http://www.ruscorpora.ru/instruction-syntax.html

For more details, see the recently published paper (in Russian):

Дяченко П.В., Иомдин Л.Л., Лазурский А.В., Митюшин Л.Г., Подлесская О.Ю., Сизов В.Г., Фролова Т.И., Цинман Л.Л. Современное состояние глубоко аннотированного корпуса текстов русского языка (СинТагРус) // Сборник «Национальный корпус русского языка: 10 лет проекту». Труды Института русского языка им. В.В. Виноградова. М., 2015. Вып. 6. С. 272-299.

New Data 2015–2020

The UD releases of SynTagRus (1.3 to 2.8) differed in the gradually improved conversion of annotation but they shared the same body of underlying texts, which entered the source corpus before 2015. A new set of texts, annotated from 2015 to 2020 and manually checked by the RNC (Russian National Corpus) 2.0 disambiguation team, was converted to UD and added for release 2.9. It consists of

  • 25447 sentences
  • 409387 words
  • 756 empty nodes

The new data has been distributed to train, development and test sets proportionally so that the result has approximately 80%:10%:10% of syntactic words. Whole documents are always kept in the same data part.

The new data can be distinguished by sentence ids (starting with 2015, ..., 2020). In addition, every new sentence has the following sentence-level comment:

  • Added to SynTagRus 2015–2020.

Acknowledgements

References

  • Droganova, K., Lyashevskaya, O., & Zeman, D. (2018). Data Conversion and Consistency of Monolingual Corpora: Russian UD Treebanks. In Proceedings of the 17th International Workshop on Treebanks and Linguistic Theories (TLT 2018), December 13–14, 2018, Oslo University, Norway (No. 155, pp. 52-65). Linköping University Electronic Press.

Changelog

  • 2021-11-15 v2.9

    • Fixed remaining legacy validation errors.
    • Added enhanced relations other than gapping (gapping was already there).
    • Added 409K words of data annotated between 2015 and 2020 (see above).

  • 2019-05-15 v2.4

    • enhanced representation fixed ('бы')
    • AUX for aux 'бы'

  • 2018-11-15 v2.3

    • Rules for punctuation fixed
    • True case lammes for PROPN
    • advmod/discource distinction
    • aux for бы (fixed some issues)

  • 2018-04-15 v2.2

    • Rules for punctuation implemented
    • Rules for reported speech implemented
    • Passives fixed
    • PROPN distinguishing from NOUN improved
    • MWE fixed (underscored lemmas)

  • 2017-11-15 v2.1

    • Conversion rules for syntax completely rewritten
    • PROPN distinguishing from NOUN improved
    • csubj added
    • Elliptic constructions fixed
    • MWE fixed

  • 2017-03-15 v2.0

    • Converted to UD v2 guidelines.
    • Elliptic constructions added.
    • Compounds added.

  • 2016-11-15 v1.4

    • Fixed peculiar Latin/Cyrillic encoding errors.
    • Lemmas are now lowercased as in other treebanks.
    • PROPN distinguished from NOUN, using heuristics based on upper/lowercase.
    • Added "foreign" dependencies.

=== Machine-readable metadata ================================================= Data available since: UD v1.3 License: CC BY-NC-SA 4.0 Includes text: yes Genre: news nonfiction fiction Lemmas: converted from manual UPOS: converted from manual XPOS: not available Features: converted from manual Relations: converted from manual Contributors: Droganova, Kira; Lyashevskaya, Olga; Zeman, Daniel Contributing: elsewhere Contact: zeman@ufal.mff.cuni.cz, droganova@ufal.mff.cuni.cz

Data contributors: Droganova, Kira; Lyashevskaya, Olga; Zeman, Daniel Documentation contributors: Shakurova, Lena; Mustafina, Nina