NLP pipeline for processing Bulgarian text.

This pipeline performs several NLP tasks: preprocessing (tokenization,
sentence splitting, lemmatization, part-of-speech tagging), dependency
parsing with mate-tools parser, named entity recognition, and coreference
resolution.

usage: java -jar btbpipe.jar [--input <path-to-file>] [--modules
       <preprocess,parse,coref>] [--output <path-to-file>] [--out-format
       <naf|wl>] [--in-format <txt|xml|naf|wl>] [--replace-ext]
    --in-format <txt|xml|naf|wl>         format of the input
    --input <path-to-file>               input file or folder with
                                         multiple input files
    --modules <preprocess,parse,coref>   comma separated list of modules
                                         to be applied
    --out-format <naf|wl>                format of the output
    --output <path-to-file>              output file or folder to store
                                         multiple output files
    --replace-ext                        use <out-format> as file
                                         extension when producing multiple output files: if the input file has one
                                         of the recognized <in-format> extensions, they will be replaced by the
                                         <out-format> extension. Use with caution.