BaseDataMaintenance/BaseDataMaintenance/maintenance/product/productUtils.py

from BaseDataMaintenance.maintenance.product.product_setting import *

import re
# 判断是不是入参字符串为全中文
def judge_pur_chinese(keyword):
    """
    中文字符的编码范围为： u'\u4e00' -- u'\u9fff：只要在此范围内就可以判断为中文字符串
    @param keyword:
    @return:
    """
    # 定义一个需要删除的标点符号字符串列表
    remove_chars = '[·’!"\#$%&\'()＃！（）*+,-./:;<=>?\@，：?￥★、…．＞【】［］《》？“”‘’\[\\]^_`{|}~]+'
    # 利用re.sub来删除中文字符串中的标点符号
    strings = re.sub(remove_chars, "", keyword)  # 将keyword中文字符串中remove_chars中包含的标点符号替换为空字符串
    for ch in strings:
        if u'\u4e00' <= ch <= u'\u9fff':
            pass
        else:
            return False
    return True



from fuzzywuzzy import fuzz
def is_similar(source,target):
    source = str(source).lower()
    target = str(target).lower()
    max_len = max(len(source),len(target))
    min_len = min(len(source),len(target))
    # dis_len = abs(len(source)-len(target))
    # min_dis = min(max_len*0.2,4)
    if min_len==0 and max_len>0:
        return False
    if max_len<=4:
        if source==target:
            return True
    else:
        #判断相似度
        similar = fuzz.ratio(source,target)
        if similar>90:
            return True
        # 全中文判断是否包含
        if judge_pur_chinese(source) and judge_pur_chinese(target):
            if len(source)==max_len:
                if str(source).find(target)>=0:
                    return True
            else:
                if target.find(source)>=0:
                    return True

    return False


SPECS_CHECK_SET = set([i for i in 'abcdefghijklmnopqrstuvwxyz']) | set([i for i in '0123456789']) | set([i for i in 'IⅠⅡⅢⅣⅤⅥⅦⅧⅨⅩⅪⅫ'])

def check_specs(source,target):
    '''
    check if the source specs is the same as the target
    same only if the chars in SPECS_CHECK_SET have the same counts
    :param source:
    :param target:
    :return:
    '''
    source = str(source).lower()
    target = str(target).lower()

    dict_source = {}
    dict_target = {}
    for s in source:
        if s in SPECS_CHECK_SET:
            if s not in dict_source:
                dict_source[s] = 0
            dict_source[s] += 1
    for s in target:
        if s in SPECS_CHECK_SET:
            if s not in dict_target:
                dict_target[s] = 0
            dict_target[s] += 1
    union_keys = set(list(dict_source.keys())) & set(list(dict_target.keys()))
    if len(dict_source.keys())!= len(union_keys):
        return False
    for k,v in dict_source.items():
        if v!=dict_target.get(k):
            return False
    return True

import json

import requests
session = requests.Session()
def request_embedding(sentence,retry_times=3):
    for _ in range(retry_times):
        resp = session.post(embedding_url,json={"sentence":sentence})
        if resp.status_code==200:
            content = resp.content.decode("utf-8")
            _d = json.loads(content)
            if _d.get("success"):
                return _d.get("vector")
    return None

def clean_product_name(product_name):
    '''
    clean before insert
    :param product_name:
    :return:
    '''
    return product_name

def clean_product_brand(product_brand):
    '''
    clean before insert
    :param product_brand:
    :return:
    '''
    return product_brand

SPECS_PATTERN = re.compile("[^A-Za-z0-9-\\/()（）]")
def clean_product_specs(product_specs):
    '''
    clean before insert
    :param product_specs:
    :return:
    '''
    _specs = re.sub(SPECS_PATTERN,'',product_specs)
    if len(_specs)>0:
        return _specs
    return product_specs


def clean_product_unit_price(product_unit_price):
    '''
    clean before insert
    :param product_unit_price:
    :return:
    '''
    try:
        if product_unit_price is not None and product_unit_price!="":
            _price = float(product_unit_price)
            return _price
    except Exception as e:
        return ""
    return ""


def clean_product_quantity(product_quantity):
    '''

    :param product_quantity:
    :return:
    '''
    try:
        if product_quantity is not None and product_quantity!="":
            _quantity = int(product_quantity)
            return _quantity

    except Exception as e:
        return ""
    return ""

if __name__ == '__main__':
    print(clean_product_specs("XY-K-JLJ-3A"))
    print(check_specs("佳士比F6",'佳士比”F6'))