Source code for chatbot.nlp.intend_analyzer

#from cluster.service.service_predict_seq2seq import PredictNetSeq2Seq
from cluster.service.service_predict_wcnn import PredictNetWcnn
from chatbot.common.chat_share_data import ShareData
import logging

class IntendAnalyzer(ShareData):

    """
    parse raw text to tageed, entity filterd sentence
    ※ Example
    input : I bought a car yesterday
    output : I bought a car [time]
    """

    def __init__(self, cb_id, nn_id):
        """
        init global variables
        """
        self.cb_id = cb_id
        self.nn_id = nn_id
        #self.seq2seq_model = PredictNetSeq2Seq()
        self.wcnn_model = PredictNetWcnn()

    def parse(self, share_data, type):
        """
        run intent analyzer
        :param context:
        :return:
        """
        if (share_data.get_intent_id() != ""):
            logging.info("■■■■■■■■■■ 의도 존재  : " + share_data.get_intent_id())
        else :
            if(type == 'Rule'):
                convert_data = share_data.get_convert_dict_data()
                logging.info("■■■■■■■■■■ Rule 의도 분석 Input Data : " + ' '.join(convert_data))
                intent_model = self.get_intent_model(' '.join(convert_data))
                logging.info("■■■■■■■■■■ Rule 의도 분석 결과(Model) : " + intent_model)
                share_data.set_pattern_intent_id([intent_model])
                share_data.set_intent_history({"P": intent_model})

            elif(type == 'NER'):
                convert_data = share_data.get_convert_data()
                logging.info("■■■■■■■■■■ NER 의도 분석 Input Data : " + ' '.join(convert_data))
                intent_model = self.get_intent_model(' '.join(convert_data))
                logging.info("■■■■■■■■■■ NER 의도 분석 결과(Model) : " + intent_model)
                share_data.set_intent_id([intent_model])
                share_data.set_intent_history({"i" : intent_model})

        return share_data

    def get_intent_model(self, convert_data):
        intent_model = str(self.wcnn_model.run(self.nn_id, {"input_data": convert_data, "num": 0, "clean_ans": False})[0])
        return intent_model