コード例 #1
0
ファイル: app.py プロジェクト: pawel717/PJN

def indexCorpus():
    indexer = Indexer(database)
    # index normal articles
    indexer.corpus_cursor = database.fetch_data("SELECT * FROM articles")
    indexer.compute_tf()
    indexer.compute_tf_idf()
    indexer.purge()
    # index lemmatized articles
    indexer.corpus_cursor = database.fetch_data("SELECT * FROM articles_lemma")
    indexer.output_catalog = "./indexes_lemmatized/"
    indexer.compute_tf()
    indexer.compute_tf_idf()
    indexer.purge()


if __name__ == "__main__":
    if (len(sys.argv) > 2):
        _usage()

    lemmatizer = Lemmatizer()
    lemmatizer.makeDictionaryMap()

    if (len(sys.argv) == 2):
        if (sys.argv[1] == 'index'):
            database = Database()
            lemmatizeCorpus(lemmatizer)
            indexCorpus()

    app.run()