Пример #1
0
def parse3From12():
    PcfgFileName = 'Corpus1+Corpus2-12-Corrected.pcfg'
    Directors= Directors1+Directors2

    try: nltk.corpus.set_basedir(system_corpora)
    except: system_corpora=nltk.corpus.get_basedir()
    
    logger.initLogger('ParseDirections',LogDir='MarcoLogs')
    import enchant
    from Sense import Lexicon
    spellchecker = enchant.DictWithPWL('en_US', Lexicon)
    DirParser = getDirParser(Directors, Maps, usePOSTagger, POSTaggerFileName,
                             PcfgFileName, SenseTaggerFileName, collectStats,
                             spellchecker=spellchecker)
    Directions = DirectionCorpusReader(constructItemRegexp(Directors3,Maps,mapversions='[01]'))
    parseTestSet(DirParser, Directions, list(Directions.items('CleanDirs')), 1)
Пример #2
0
    import enchant
    from Sense import Lexicon
    spellchecker = enchant.DictWithPWL('en_US', Lexicon)
    DirParser = getDirParser(Directors,
                             Maps,
                             usePOSTagger,
                             POSTaggerFileName,
                             PcfgFileName,
                             SenseTaggerFileName,
                             collectStats,
                             spellchecker=spellchecker)

    if doParses == 'Profile':
        import profile
        profile.run('testParses(DirParser)', 'parse.prof')
    elif doParses == 'TestSet':
        Directions = DirectionCorpusReader(constructItemRegexp(
            Directors, Maps))
        parseTestSet(DirParser, Directions, TestSet, 1)
    elif doParses == 'CommandLine':
        if len(sys.argv) > 1 and sys.argv[1]:
            Files = sys.argv[1]
        else:
            Files = constructItemRegexp(Directors, Maps, mapversions='[01]')
        print 'Parsing', Files
        Directions = DirectionCorpusReader(Files)
        TestSet = list(Directions.items('CleanDirs'))
        #random.shuffle(TestSet);
        #timeSort(TestSet)
        parseTestSet(DirParser, Directions, TestSet, 1)
    try: nltk.corpus.set_basedir(system_corpora)
    except: system_corpora=nltk.corpus.get_basedir()
    
    logger.initLogger('ParseDirections',LogDir='MarcoLogs')
    import enchant
    from Sense import Lexicon
    spellchecker = enchant.DictWithPWL('en_US', Lexicon)

    DirParser = getDirParser(Directors, Maps, usePOSTagger, POSTaggerFileName,
                             PcfgFileName, SenseTaggerFileName, collectStats,
                             spellchecker=spellchecker)
    
    if doParses == 'Profile':
        import profile
        profile.run('testParses(DirParser)','parse.prof')
    elif doParses == 'TestSet':
        Directions = DirectionCorpusReader(constructItemRegexp(Directors,Maps))
        parseTestSet(DirParser,Directions,TestSet,1)
    elif doParses == 'CommandLine':
        if len(sys.argv)>1 and sys.argv[1]:
            Files = sys.argv[1]
        else:
            Files = constructItemRegexp(Directors,Maps,mapversions='[01]')
        print 'Parsing', Files
        Directions = DirectionCorpusReader(Files)
        TestSet=list(Directions.items('CleanDirs'))
        #random.shuffle(TestSet);
        #timeSort(TestSet)
        parseTestSet(DirParser,Directions,TestSet,1)