jy-data-extract/udpdataclear/udpSensitiveWords/util/words.go

package util

import (
	"encoding/json"
	"log"
	"regexp"
	"strings"
	"unicode"
	"unicode/utf8"
)

func dealWithNameScoreRules(name string) (string, bool, float64 , []map[string]interface{}) {
	new_name, new_score, isok := "", float64(0), false
	old_name := escapeNew(name)
	if old_name == "" {
		return "", false, new_score ,nil
	}

	//标准分：
	standard_score := float64(4.0)
	query_name := old_name
	endstr := endWordReg.FindString(query_name)
	if endstr !="" {
		standard_score = float64(3.0)
		query_name = strings.ReplaceAll(query_name,endstr,"")
	}

	query := `{"query":{"bool":{"must":[{"query_string":{"default_field":"unique_qy.name_word","query":"` + query_name + `"}}],"must_not":[],"should":[]}},"from":"0","size":"200"}`
	tmp := make(map[string]interface{})
	json.Unmarshal([]byte(query), &tmp)
	searchResult, err := Client_Es.Search().Index(es_index).Type(es_type).Source(tmp).Do()
	if err != nil {
		log.Println("ES查询出错", name, old_name,err)
		return "", false,new_score, nil
	}
	if searchResult.Hits!= nil{
		resNum := len(searchResult.Hits.Hits)
		res := make([]map[string]interface{}, resNum)
		if searchResult.Hits != nil {
			if resNum < 1000 {
				for i, hit := range searchResult.Hits.Hits {
					data := make(map[string]interface{}, 0)
					json.Unmarshal(*hit.Source, &data)
					res[i] = map[string]interface{}{
						"name":  data["name"],
						"score": *hit.Score,
					}
				}
			} else {
				log.Println("查询结果太多，查询到：", resNum, "条")
			}
		}
		if len(res) > 0 {
			//针对特殊情况-特殊处理 公司结尾
			new_name = ObjToString(res[0]["name"])
			new_score = Float64All(res[0]["score"])
			if endComReg.MatchString(name) {
				new_name,new_score = dealWithSpecialName(name,res)
				if new_name=="" {
					new_name = ObjToString(res[0]["name"])
					new_score = Float64All(res[0]["score"])
				}
			}
		}
		if new_name != "" { //分析hit比例
			total, hit := dealWithWordsRules(name, new_name)
			proportion := float64(hit) / float64(total)
			if proportion >= 1.0 {
				isok = true
			} else {

				//前置规则-与分数无关  江苏凤凰出版社--江苏凤凰出版社有限公司
				//吉林省彩虹城市建设工程有限公司--吉林彩虹城市建设工程有限公司
				if dealWithPreRule(query_name,new_name) {
					return new_name, true,new_score, res
				}

				if float64(hit)/float64(total) >= 0.8 && new_score > standard_score {
					str1,str2:=startWordReg_1.FindString(name),startWordReg_1.FindString(new_name)
					if str1!="" && str2!="" {
						if strings.Contains(str1,str2)||strings.Contains(str2,str1) {

						}else {
							return new_name, false,new_score, res
						}
					}
					str1,str2 = startWordReg_2.FindString(name),startWordReg_2.FindString(new_name)
					if str1!="" && str2!=""{
						if str1 != str2 {
							return new_name, false,new_score, res
						}
					}
					isok = true
				}else {}
			}
		}
		return new_name, isok,new_score, res
	}
	return new_name,isok,new_score,nil
}







func dealWithPreRule(name string , new_name string) bool {
	//log.Println("规则时：",name,new_name)
	endstr := endWordReg.FindString(new_name)
	if endstr !="" {
		new_name = strings.ReplaceAll(new_name,endstr,"")
	}
	if name==new_name {
		return true
	}
	//去掉开头 全程简称
	str1,str2 := startWordReg_2.FindString(name),startWordReg_2.FindString(new_name)
	if str1!="" && str2!="" && str1==str2 {
		//在清全程简称
		start_str_all_1 := startWordReg_3.FindString(name)
		if start_str_all_1 !="" {
			name = strings.ReplaceAll(name,start_str_all_1,"")
		}else {
			start_str_sim_1 := startWordReg_2.FindString(name)
			name = strings.ReplaceAll(name,start_str_sim_1,"")
		}

		start_str_all_2 := startWordReg_3.FindString(new_name)
		if start_str_all_2 !="" {
			new_name = strings.ReplaceAll(new_name,start_str_all_2,"")
		}else {
			start_str_sim_2 := startWordReg_2.FindString(new_name)
			new_name = strings.ReplaceAll(new_name,start_str_sim_2,"")
		}
		if name==new_name {
			return true
		}
	}


	//去掉指定维文字
	name = strings.ReplaceAll(name,"科技","")
	new_name = strings.ReplaceAll(new_name,"科技","")

	name = strings.ReplaceAll(name,"建筑工程","")
	new_name = strings.ReplaceAll(new_name,"建筑工程","")

	name = strings.ReplaceAll(name,"工程","")
	new_name = strings.ReplaceAll(new_name,"工程","")

	name = strings.ReplaceAll(name,"标识","")
	new_name = strings.ReplaceAll(new_name,"标识","")

	name = strings.ReplaceAll(name,"工业","")
	new_name = strings.ReplaceAll(new_name,"工业","")

	name = strings.ReplaceAll(name,"公司","")
	new_name = strings.ReplaceAll(new_name,"公司","")

	name = strings.ReplaceAll(name,"(","")
	new_name = strings.ReplaceAll(new_name,"(","")

	name = strings.ReplaceAll(name,")","")
	new_name = strings.ReplaceAll(new_name,")","")

	name = strings.ReplaceAll(name,"（","")
	new_name = strings.ReplaceAll(new_name,"（","")

	name = strings.ReplaceAll(name,"）","")
	new_name = strings.ReplaceAll(new_name,"）","")


	name = strings.ReplaceAll(name,"信息技术","信息")
	new_name = strings.ReplaceAll(new_name,"信息技术","信息")

	name = strings.ReplaceAll(name,"电子科技","电子")
	new_name = strings.ReplaceAll(new_name,"电子科技","电子")

	name = strings.ReplaceAll(name,"电子技术","电子")
	new_name = strings.ReplaceAll(new_name,"电子技术","电子")

	name = strings.ReplaceAll(name,"建设集团","建设")
	new_name = strings.ReplaceAll(new_name,"建设集团","建设")



	//log.Println("最终清理后-",name,new_name)
	if name==new_name {
		return true
	}




	return false
}












//击中数量以及比例
func dealWithWordsRules(info_name string, source_name string) (int, int) {
	total, hit := 0, 0
	//字符串处理，替换指定字符
	source_name = strings.ReplaceAll(source_name, "（微型企业）", "")
	source_name = strings.ReplaceAll(source_name, "（有限合伙）", "")
	source_name = strings.ReplaceAll(source_name, "（普通合伙）", "")

	info_name = strings.ReplaceAll(info_name, "(", "")
	info_name = strings.ReplaceAll(info_name, ")", "")
	info_name = strings.ReplaceAll(info_name, "（", "")
	info_name = strings.ReplaceAll(info_name, "）", "")
	info_name = strings.ReplaceAll(info_name, "〉", "")

	source_name = strings.ReplaceAll(source_name, "(", "")
	source_name = strings.ReplaceAll(source_name, ")", "")
	source_name = strings.ReplaceAll(source_name, "（", "")
	source_name = strings.ReplaceAll(source_name, "）", "")
	source_name = strings.ReplaceAll(source_name, "〉", "")

	nameArr, _ := calculateWordCount(info_name)
	_, total = calculateWordCount(source_name)
	for _, v1 := range nameArr {
		if strings.Contains(source_name, v1) {
			hit++
		}
	}
	return total, hit
}

//分词结果
func calculateWordCount(name string) ([]string, int) {
	arr, space := make([]string, 0), 2
	total := utf8.RuneCountInString(name) - (space - 1)
	if name == "" || total <= 0 {
		return arr, 0
	}
	nameRune := []rune(name)
	for i := 0; i < total; i++ {
		new_str := string(nameRune[i : space+i])
		arr = append(arr, new_str)
	}
	return arr, len(arr)
}

//func escape(s string) string {
//	news := ""
//	s = strings.ReplaceAll(s," ","")
//	for _, c := range s {
//		//if unicode.Is(unicode.Han, c) || unicode.IsNumber(c) || unicode.IsLetter(c) {
//		//	news = news + string(c)
//		//}else if c == '\\' || c == '+' || c == '-' || c == '!' || c == '(' || c == ')' || c == ':' || c == '^' || c == '[' || c == ']' || c == '"' || c == '{' || c == '}' || c == '~' || c == '*' || c == '?' || c == '|' || c == '&' || c == '/' || c == '#' || c == '@' || c == '（' || c == '）' || c == '>' || c == '<' || c == '“' || c == '”' || c == '？' || c == '、' || c == '.' {
//		//	a := string([]rune{os.PathSeparator, '\\'})
//		//	news = news + a + string(c)
//		//} else {
//		//	return ""
//		//}
//		if c == '\\' || c == '+' || c == '-' || c == '!' || c == '(' || c == ')' || c == ':' || c == '^' || c == '[' || c == ']' || c == '{' || c == '}' || c == '~' || c == '*' || c == '?' || c == '|' || c == '&' || c == '/' || c == '#' || c == '@' || c == '（' || c == '）' || c == '>' || c == '<' || c == '“' || c == '”' || c == '？' || c == '、' || c == '.' {
//			a := string([]rune{os.PathSeparator,'\\'})
//			//news = news + a + `\` + string(c)
//			news = news + a  + string(c)
//		} else {
//			news = news + string(c)
//		}
//
//	}
//	return news
//}

func escapeNew(s string) string {
	news := ""
	s = strings.ReplaceAll(s, " ", "")
	for _, c := range s {
		if unicode.Is(unicode.Han, c) || unicode.IsNumber(c) || unicode.IsLetter(c) {
			news = news + string(c)
		}
	}
	return news
}




//处理特殊情况-分析最终
func dealWithSpecialName(name string,res []map[string]interface{})(string ,float64) {
	//log.Println("特殊...")
	new_name ,new_score,proportion := "",float64(0),float64(0)
	for k,v:=range res {
		tmp_name := ObjToString(v["name"])
		tmp_score := Float64All(v["score"])
		if k>=20 || tmp_score<3.0 {
			break
		}

		if endComReg.MatchString(tmp_name) {
			if endFComReg.MatchString(name)  {
				if !endFComReg.MatchString(tmp_name)  {
					continue
				}
			}else {
				if endFComReg.MatchString(tmp_name)  {
					continue
				}
			}
		}else {
			continue
		}
		total, hit := dealWithWordsRules(tmp_name, name)
		tmp_proportion := float64(hit) / float64(total)
		//log.Println(tmp_proportion,tmp_name,name)
		if tmp_proportion > proportion {
			proportion = tmp_proportion
			new_name = tmp_name
			new_score = tmp_score
		}
	}

	return new_name,float64(new_score)
}


var endComReg *regexp.Regexp = regexp.MustCompile("(公司)$")
var endFComReg *regexp.Regexp = regexp.MustCompile("(分公司)$")