competitor_product_analysis/sdk/WinhcEsCompanyV8Sdk.py

# -*- coding: utf-8 -*-
# @Time : 2020/12/10 10:50
# @Author : XuJiakai
# @File : WinhcEsV8Sdk
# @Software: PyCharm
import json
import logging
import re

from elasticsearch.exceptions import NotFoundError
from zhconv import convert

from sdk.WinhcElasticSearchSDK import WinhcElasticSearchSDK

logging.info("Begin")
logger = logging.getLogger(__name__)
logger.setLevel(logging.INFO)
fmt = logging.Formatter('%(asctime)s - %(pathname)s[line:%(lineno)d] - %(levelname)s: %(message)s')


class winhc_es_company_v8_sdk(WinhcElasticSearchSDK):
    def __init__(self, es_name='new'):
        super().__init__(es_name)
        self.pattern = re.compile('[^\\u4e00-\\u9fa50-9a-zA-Z]')
        self.index = "winhc-company-v9"
        self.doc_type = 'company'

    def get_doc_by_company_id(self, company_id):
        try:
            res = self.es.get(index=self.index, doc_type=self.doc_type, id=company_id)['_source']
            res['company_id'] = company_id
            return res
        except NotFoundError:
            return None

    def exists_company_id(self, company_id):
        return super().exists(self.index, self.doc_type, company_id)

    def query_by_company(self, company_name):
        """
        :param company_name:
        :return: 最多反回20条数据
        """
        cname = self.clean_up(company_name)
        dsl = """{"query":{"bool":{"must":[{"bool":{"should":[{"term":{"cname.value.keyword":{"value":"%s"}}},{"term":{"history_name.value.keyword":{"value":"%s"}}}]}}]}}}""" % (
            cname, cname)
        return self.__get_doc_by_dsl(dsl)

    def query_by_credit_code_reg_number(self, credit_code, reg_number):
        """
        :param company_name:
        :return: 最多反回20条数据
        """
        dsl_pre = """{"query":{"bool":{"must":[{"bool":{"filter":[{"bool":{"must_not":[{"term":{"deleted":{"value":9}}}]}}]}},{"bool":{"should":["""
        dsl_post = """]}}]}}}"""
        dsl = []
        if credit_code is not None and credit_code.strip() != '':
            dsl.append("""{"term":{"credit_code":{"value":"%s"}}}""" % credit_code)
            pass
        if reg_number is not None and reg_number.strip() != '':
            dsl.append("""{"term":{"reg_number":{"value":"%s"}}}""" % reg_number)
            pass

        if len(dsl) == 0:
            raise AttributeError("reg_number or credit_code is empty !")
        dsl = dsl_pre + ','.join(dsl) + dsl_post
        return self.__get_doc_by_dsl(dsl)

    def clean_up(self, name):
        return self.pattern.sub("", name)

    def __get_doc_by_dsl(self, dsl):
        res = super().query(index=self.index, doc_type=self.doc_type, dsl=dsl)
        return res

    def query(self, company_name):
        """
        :param company_name:
        :return: 最多反回20条数据
        """
        logger.info("es查询:%s" % (company_name))

        def get_term(n, is_clean=False):
            if is_clean:
                k = "value"
            else:
                k = 'show.keyword'
            all_term = [{
                "term": {
                    "cname." + k: {
                        "value": n
                    }
                }
            }, {
                "term": {
                    "cname.simplified_chinese.keyword": {
                        "value": n
                    }
                }
            }, {
                "term": {
                    "history_name." + k: {
                        "value": n
                    }
                }
            }, {
                "term": {
                    "history_name.simplified_chinese.keyword": {
                        "value": n
                    }
                }
            }]
            return all_term

        all_term = get_term(company_name)

        cname = self.clean_up(company_name)
        if company_name != cname:
            all_term.extend(get_term(cname))

        simplified_chinese = convert(company_name, 'zh-cn')
        if company_name != simplified_chinese:
            all_term.extend(get_term(simplified_chinese))

        simplified_cname = self.clean_up(simplified_chinese)
        if company_name != simplified_cname:
            all_term.extend(get_term(simplified_cname))

        dsl = {
            "size": 20,
            "_source": [
                "cname.show"
            ],
            "query": {
                "bool": {
                    "must": [
                        {
                            "bool": {
                                "should": all_term
                            }
                        },
                        {
                            "term": {
                                "deleted": {
                                    "value": "0"
                                }
                            }
                        }
                    ]
                }
            }
        }
        # print(json.dumps(dsl))
        return self.__get_doc_by_dsl(dsl)


if __name__ == '__main__':
    sdk = winhc_es_company_v8_sdk('new')
    res = sdk.query('成安建筑营造有限公司')
    print(res)
    # print(sdk.query_by_company('香港澳德利集团有限公司'))
    pass