实体标注工具

功能：文本实体标注，用于做ner nre等的训练测试集；
工具：python2
输入：
输出：

脚本：

#!/usr/bin/python
# -*- coding: utf-8 -*-
"""
@author:
@contact:
@context：对每行每个实体做清洗，其中每个实体需要从第一个字开始依次滑动遍历查找。
"""
import sys,xlrd
reload(sys)
sys.setdefaultencoding("utf-8")


#数据读入
data=xlrd.open_workbook(r'C:/Users/Administrator/Desktop/1/1.xlsx')
table = data.sheets()[0]

#数据的行列
nrows=table.nrows
ncols=table.ncols

L_final=[]
for k in range(1,nrows):
    #下面缩进模块是对某一行的处理
    text=table.row_values(k)[2]
    text_list=list(text)
    type_ner=range(0, len(text_list))

    #第k行实体对应的类型
    nrows_type=str(table.row_values(k)[1])#<type 'str'>

    #看第k行实体个数
    nrows_list_qukong=list(filter(None,table.row_values(k)))
    nrows_len=len(nrows_list_qukong)

    #for:对第k行的实体做标注
    for i in range(3,nrows_len):

        #缩进里是对第k行第i个实体做了完整遍历
        entity=table.row_values(k)[i]
        #计算实体元素的个数，unicode和非unicode（浮点）用不同的计算方式
        if isinstance(entity, unicode)!=True:
            len_a = len(list(str(int(entity))))
        else:
            len_a = len(list(entity))
            
        search = str(entity)
        start = 0
        while True:
            index = text.find(search, start)
            if index == -1:
                break
            # print(index,index+len_a-1)
            index_s = index
            index_e = index+len_a

            for j in range(0, len(text_list)):
                if j in range(index_s, index_e):
                    if j==index_s:
                        type_ner[j] = str("B-"+ str(nrows_type) +"")
                    else:
                        type_ner[j] = str("I-"+ str(nrows_type) +"")
                elif start !=0 or i>3:
                    print("非首次标注")
                else:
                    type_ner[j] = str("O")

            start = index + 1
    # print(type_ner)

    #对第k行遍历其他类型的实体
    for entity in [u'薛宝钗']:

        # 计算实体元素的个数，unicode和非unicode（浮点）用不同的计算方式
        if isinstance(entity, unicode) != True:
            len_a = len(list(str(int(entity))))
        else:
            len_a = len(list(entity))

        search = str(entity)
        start = 0
        while True:
            index = text.find(search, start)
            if index == -1:
                break
            # print(index,index+len_a-1)
            index_s = index
            index_e = index + len_a

            for j in range(0, len(text_list)):
                if j in range(index_s, index_e):
                    if j == index_s:
                        type_ner[j] = str("B-X")
                    else:
                        type_ner[j] = str("I-X")
            start = index + 1


    final_result = []
    for i in range(0, len(text_list)):
        text_list = [str(x) for x in text_list]
        type_ner = [str(x) for x in type_ner]
        final = ' '.join([text_list[i], type_ner[i]])
        final_result.append(final)
    final_result.append(' ')
    L_final=L_final+final_result
end = '\n'.join(L_final)
with open('etl_dataset.txt', 'w') as f:
    f.write(end)