Skip to content

Latest commit

 

History

History
43 lines (35 loc) · 2.03 KB

README.md

File metadata and controls

43 lines (35 loc) · 2.03 KB
Raw

Abductive discource processing pipeline

DESCRIPTION

Multilingual (English, Spanish, Farsi, Russian) abductive discource processing pipeline.

  • NLPipeline_MULT_stdinout_CM.py – running tokenizer, lemmatizer, parser, logical form converter, abductive reasoner, proof graph generator, CM output
usage: NLPipeline_MULT_stdinout_CM.py [-h] [--lang LANG] [--input INPUT]
                                   [--outputdir OUTPUTDIR] [--parse] [--henry]
                                   [--kb KB] [--kbcompiled KBCOMPILED]
                                   [--graph GRAPH] [--CMoutput]
optional arguments:
  -h, --help            show this help message and exit
  --lang LANG           Input language: EN, ES, RU, FA.
  --input INPUT         Input file: plain text (possibly with text ids),
                        observation file, henry file.
  --outputdir OUTPUTDIR
                        Output directory. If input file defined, then default
                        is input file dir. Otherwise its TMP_DIR.
  --parse               Tokenize and parse text, produce logical forms,
                        convert to obeservations.
  --henry               Process observations with Henry.
  --kb KB               Path to noncompiled knowledge base.
  --kbcompiled KBCOMPILED
                        Path to compiled knowledge base.
  --graph GRAPH         ID of text/sentence to vizualize. Possible value:
                        allN, where N is number of sentences to vizualize.
  --CMoutput            Conceptual metaphor output.

COMPONENTS