JAM (Just A Multialigner) est un logiciel d'alignement multilingue permettant d'aligner simultanément 2 textes ou plus. JAM (Just A Multialigner) is multilingual alignment software which allow to align more than 2 texts at the same time. [Télécharger le script Perl / Download the Perl Script] TODO Add a documentation Publications Kraif Olivier (2015) Multi-alignement vs bi-alignement : à plusieurs, c’est mieux ! , Actes de TALN 2015, 22ème Conférence sur le Traitement Automatique des Langues Naturelles, Caen, 22-25 juin 2015, pp. 255-266 [http://www.atala.org/taln_archives/TALN/TALN-2015/taln-2015-long-022.pdf ]. Synopsisperl -w JAM.pl InputFile+ [options*] Options : --outputPath the/output/path: set the output path for output files --inputFormat : txt (with one string per line) or ces --encoding : utf8|iso-8859-1|ansi etc. --outputFormat : txt (with one string per line), or ces, or tmx, or csv --languages : list of languages corresponding to the input files (useless if the input files are in the forme *.LANG.ces where LANG indicates the language code --combinations : ordered list of language combinations for language tiling (e.g. en-fr-es fr-it-en es-pt-en pt-es-it nl-en-fr sv-da-nl de-da-nl fi-en-sv el-gr-en) --pivot : if no combination is given, a language may be used as a pivot --loadPoints CSVFILE : load some anchor points to initialize aligning process --finalCompletion : use the Gale & Church algorithme to finally complete aligning process --printPairwiseAlignment : print all the alignments two by two --printMergedAlignment : print all the pairwise alignments merged into a single file --verbose : print messages while processing --printLog : print messages in log file --printTokenNum : print the token numbers for each language combination --printTokenList : print the token list for each language combination --printBestPairComb : print the best pairwize a posteriori combination
|