Summary

UD_French-Sequoia is an automatic conversion of the Sequoia Treebank corpus French Sequoia corpus.

Introduction

UD_French-Sequoia is an automatic conversion of the Sequoia Treebank corpus French Sequoia corpus. The conversion was done with the Grew software and the Graph Rewriting System described here.

Origin

The first version of the Sequoia Corpus was presented in (Candito & Seddah, 2012)

Splitting

The whole corpus contains 70,624 tokens in 3,099 sentences.

In UD_French-Sequoia, data were randomly split into:

• fr_sequoia-ud-test.conllu: 10,050 tokens in 456 sentences
• fr_sequoia-ud-dev.conllu: 10,013 tokens in 412 sentences
• fr_sequoia-ud-train.conllu: 50,561 tokens in 2,231 sentences

Genres

The original sentences of the corpus are taken from:

• French Europarl (sent_id prefix: Europar.550)
• Wikipédia Fr (sent_id prefix: frwiki_50.1000)
• Newspaper Est Républicain (sent_id prefix: annodis.er)
• European Medicines Agency (sent_id prefix: emea-fr-dev and emea-fr-test)

Acknowledgments

The conversion has been performed by Bruno Guillaume with the Graph Rewriting System described here developed by Bruno Guillaume and Guy Perrier.

The Sequoia Corpus was presented in (Candito & Seddah, 2012) and revised later, notably during the project of deep annotation described in (Candito & al. 2014) and (Perrier & al. 2014).

References

(Candito & Seddah, 2012) Marie Candito, Djamé Seddah. Le corpus Sequoia : annotation syntaxique et exploitation pour l'adaptation d'analyseur par pont lexical. TALN 2012 - 19e conférence sur le Traitement Automatique des Langues Naturelles, Jun 2012, Grenoble, France. 2012.

(Candito & al. 2014) Marie Candito, Guy Perrier, Bruno Guillaume, Corentin Ribeyre, Karën Fort, Djamé Seddah and Éric de la Clergerie. (2014) Deep Syntax Annotation of the Sequoia French Treebank. Proc. of LREC 2014, Reykjavic, Iceland.

(Perrier & al. 2014) Guy Perrier, Marie Candito, Bruno Guillaume, Corentin Ribeyre, Karën Fort and Djamé Seddah. (2014) Un schéma d'annotation en dépendances syntaxiques profondes pour le français. Proc. of TALN 2014, Marseille, France.

Changelog

• 2018-04-15 v2.2
  • Subtyping of the obl relation with 3 relations obl:arg, obl:mod, obl:agent
  • Several corrections in the original corpus and in the conversion process
• 2017-11-15 v2.1
  • Manual corrections in the original Treebank
  • Application of an updated conversion system, taking into account new decisions taken for harmonisation of several French Treebanks (causative, copules, auxiliaries)
• 2017-03-01 v2.0
  • First release in UD

=== Machine-readable metadata (DO NOT REMOVE!) ================================ Data available since: UD v2.0 License: LGPL-LR Includes text: yes Genre: wiki medical new nonfiction Lemmas: converted from manual UPOS: converted from manual XPOS: not available Features: converted from manual Relations: converted from manual Contributors: Candito, Marie; Seddah, Djamé; Perrier, Guy; Guillaume, Bruno Contributing: elsewhere Contact: bruno.guillaume@inria.fr