ConlluEditor: a fully graphical editor for Universal dependencies treebank files Johannes Heinecke Orange / TGI / Data & IA / DESKIÑ Universal Dependencies Workshop 2019
Mo�va�on ConlluEditor Johannes Heinecke UDW 2019 Mo�va�on Features Edi�ng CoNLL-U files Demo readiblity (on one line, sub-trees easy to grasp, all important informa�on presented) browsing searching (for forms, lemmas, U/XPOS, deprels, sequence of all of these modifying any CoNLL-U column in graphic mode (including enhanced rela�ons) trees, not flat graphs abstrac�on from underlying format git support svg/XƎL A T EX export 2/7 . . . . . . . . . . . . . .
Features I ConlluEditor Johannes Heinecke UDW 2019 Edi�ng Mo�va�on Features full graphical editor for basic and enhanced dependency rela�ons Demo word edit (form, lemma, UPOS, XPOS, features, misc-column) autocomple�on (UPOS, XPOS, deprels, using user provided lists of valid labels) edi�ng mul�token words ( [1-2] ... ) and empty nodes ( [5.1] ... ) comment edi�ng support for right-to-le� scripts like Arabic or Hebrew split and join words (to correct bad tokeniza�on), split and join sentences (to modify sentence segmenta�on) undo/redo git support (add/commit every n edits, default: 1) export of dependency graphs as .svg -image, XƎL T EX-code (for the �kz-dependency A package or the deptree.sty XƎL T EX style), sd-parse, CoNLL-U A 3/7 . . . . . . . . . . . . . .
Features II ConlluEditor Johannes Heinecke UDW 2019 Mo�va�on Searching Features regex search func�ons (including sequences of tokens, sub-graphs and comments) Demo Valida�on valida�on: indicates undefined UPOS, XPOS, dependency rela�ons (based on lists given to the server) prohibi�on of invalid (cyclic) trees normalisa�on of token ids (first column, from 1 to n , taking into account mul�token words, empty words and heads) valida�on with external script (such like UD’s validate.py ) on the current sentence limited mul�-user support: as long as two users do not edit the same sentence 4/7 . . . . . . . . . . . . . .
Architecture ConlluEditor Johannes Heinecke UDW 2019 git Mo�va�on gitaa gitbb Features Demo CoNLL-U file I Firefox ≥ 63/ AJAX Chrome ≥ 60/ list of UPOS HTTP Edge ≥ 44 server list(s) of XPOS (java 8) conllueditor (JS) list(s) of deprels valida�on script I bootstrap jquery-gui 4.1.3 1.12.1 jquery 3.3.1 5/7 . . . . . . . . . . . . . .
demo.conllu 5555 $HOME/UD/tools/validate.py --lang cy --max-err 0 --level 5 {FILE} $CE/bin/conlluedit.sh -r \ --deprels deprel.cy,deprel.ud \ --UPOS cpos.ud \ --XPOS xpos.cy \ --validator valid.conf \ Let’s see the real thing ConlluEditor Johannes Heinecke UDW 2019 Mo�va�on configure validator (in valid.conf ): Features Demo start server open navigator — h�p:/ /localhost:5555 — demo 6/7 . . . . . . . . . . . . . .
ConlluEditor root Thank Johannes Heinecke thank 1 UDW 2019 VERB VBP Mood : Ind T ense : Pres Mo�va�on VerbForm : Fin Features obj o b p u l : f o n c r t Demo you attention ! you attention ! 2 5 6 PRON NOUN PUNCT PRP NN , Case : Acc Number : Sing SpaceAfter : No Person : 2 nmod:poss PronT ype : Prs case for your for you 3 4 ADP PRON IN PRP$ Person : 2 Poss : Yes PronT ype : Prs Thank you for your attention ! 1 2 3 4 5 6 https://github.com/Orange-OpenSource/conllueditor johannes.heinecke@orange.com 7/7 . . . . . . . . . . . . . .
Recommend
More recommend