Add documents source

[lecw]

Ce qui serait pratique ? Avoir les fichiers texte source des documents annotés.

À extraire / formater, soit à la main à partir du googledoc "Annotation WL+FG" (ou "Documents à annoter"), soit programmatiquement à partir de la ressource JSON.

[lecw]

e.g. this file contains all tokenized sentences, which is half good.

wire_tokenized_sentences.txt

[rali-udem]

Commit 056fd7c now provides excerpts for the original documents, as well as their source URLs. Tokenization is left to the reader, which is not an enormous task given the number of documents.