French TimeBank 1.1 Documentation
---------------------------------

License
-------
French TimeBank Corpus by André Bittar is licensed under the Lesser General Public License for Linguistic Resources. Details of the license are given in the file LICENSE.

Citation
--------
@inproceedings{bittar-etal-2011-french,
    title = "{F}rench {T}ime{B}ank: An {ISO}-{T}ime{ML} Annotated Reference Corpus",
    author = "Bittar, Andr{\'e}  and Amsili, Pascal and Denis, Pascal and Danlos, Laurence",
    booktitle = "Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies",
    month = jun,
    year = "2011",
    address = "Portland, Oregon, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/P11-2023",
    pages = "130--134",
}

Metdata
-------
Corpus name: French TimeBank
Version: 1.1
Author: André Bittar ([email protected])
Language: French (France)
Data format: ISO-TimeML compliant XML, inline annotation
Text source: Articles extracted from the Est Républicain/CNRTL corpus of news articles
             Original text corpus available for download at http://www.cnrtl.fr/corpus/estrepublicain/

Corpus structure
----------------

FR-TimeBank/Docs/
	README				This file
	LICENSE				The text of the Lesser General Public License for Linguistic Resources
	FR-Annotation-Guide.pdf		The full ISO-TimeML annotation guidelines for French used in annotation

FR-TimeBank/Data/			108 ISO-TimeML annotated documents

FR-TimeBank/Validation/			XML DTD, as well as the shell script used to validate the corpus documents

Basic corpus statistics
-----------------------

Documents: 108
Tokens: 15,423 (basic whitespace-separated)
EVENT: 2,115
TIMEX3: 533
SIGNAL: 288
ALINK: 35
SLINK: 452
TLINK: 2,392
Size of corpus: 840Kb

Version notes
-------------
2013.07.04
----------
Modifications by Véronique Moriceau ([email protected]) with help from Xavier Tannier [email protected]
** deleted file 1999-05-21-0108 that only contained normalised times "Thh:mm" instead of "YYYY-MM-DDThh:mm"

** correction of normalised values :
1999-05-19-0543.tml	14 h 	T14:00 => 1999-05-19T14:00
1999-05-30-0974.tml	début des années 60	196 => 196X
1999-07-29-0552.tml	années 60	196 => 196X
1999-07-29-0552.tml	XIXe siècle	18 => 18XX
1999-07-09-0441.tml	cette semaine	W36 => W27

** date + time tag splits <TIMEX3>jour à heure</TIMEX3> => <TIMEX3>jour</TIMEX3> à <TIMEX3>heure</TIMEX3>
1999-05-26-1342.tml	samedi à 8 h
1999-07-23-0052.tml	samedi 24 juillet à 15 h 45
2002-05-11-0410.tml	jeudi 23 mai, à 20 h 30
2003-01-04-1029.tml	Dimanche 5 janvier, à 9 h 15
2003-01-04-1029.tml	lundi 17 février, à 14 h 30

** transformation of type=DATE to type=TIME when the value is of the type YYY-MM-DDT...