test.py 文件源码

def createIndex(self, client, num_docs = 100):

        assert isinstance(client, Client)
        #conn.flushdb()
        #client = Client('test', port=conn.port)
        try:
            client.create_index((TextField('play', weight=5.0), 
                                TextField('txt'), 
                                NumericField('chapter')))
        except redis.ResponseError:
            client.drop_index()
            return self.createIndex(client, num_docs=num_docs)

        chapters = {}

        with bz2.BZ2File(WILL_PLAY_TEXT) as fp:

            r = csv.reader(fp, delimiter=';')
            for n, line in enumerate(r):
                #['62816', 'Merchant of Venice', '9', '3.2.74', 'PORTIA', "I'll begin it,--Ding, dong, bell."]

                play, chapter, character, text = line[1], line[2], line[4], line[5]

                key = '{}:{}'.format(play, chapter).lower()
                d = chapters.setdefault(key, {})
                d['play'] = play
                d['txt'] = d.get('txt', '') + ' ' + text
                d['chapter'] = int(chapter or 0)

                if len(chapters) == num_docs:
                    break

        indexer = client.batch_indexer(chunk_size=50)
        self.assertIsInstance(indexer, Client.BatchIndexer)
        self.assertEqual(50, indexer.chunk_size)

        for key, doc in chapters.iteritems():

            indexer.add_document(key, **doc)
        indexer.commit()