手记实用系列文章:
3 自然语言处理手记
语料预处理封装类:
#coding=utf-8
import os
import jieba
import sys
import re
import time
import jieba.posseg as pseg
sys.path.append("../")
jieba.load_userdict("../Database/userdict.txt") # 加载自定义分词词典
'''
title:利用结巴分词进行文本语料处理:单文本处理器、批量文件处理器
1 首先对文本进行遍历查找
2 创建原始文本的保存结构
3 对原文本进行结巴分词和停用词处理
4 对预处理结果进行标准化格式,并保存原文件结构路径
author:白宁超
myblog:http://www.cnblogs.com/baiboy/
'''
'''
分词.词性标注以及去停用词
stopwordspath: 停用词路径
dealpath:中文数据预处理文件的路径
savepath:中文数据预处理结果的保存路径
'''
def cutTxtWord(dealpath,savepath,stopwordspath):
stopwords = {}.fromkeys([ line.rstrip() for line in open(stopwordspath,"r",encoding='utf-8')]) # 停用词表
with open(dealpath,"r",encoding='utf-8') as f:
txtlist=f.read() # 读取待处理的文本
words =pseg.cut(txtlist) # 带词性标注的分词结果
cutresult=""# 获取去除停用词后的分词结果
for word, flag in words:
if word not in stopwords:
cutresult += word+"/"+flag+" " #去停用词
getFlag(cutresult,savepath) #
'''
分词.词性标注以及去停用词
stopwordspath: 停用词路径
read_folder_path :中文数据预处理文件的路径
write_folder_path :中文数据预处理结果的保存路径
filescount=300 #设置文件夹下文件最多多少个
'''
def cutFileWord(read_folder_path,write_folder_path,stopwordspath):
# 停用词表
stopwords = {}.fromkeys([ line.rstrip() for line in open(stopwordspath,"r",encoding='utf-8')])
# 获取待处理根目录下的所有类别
folder_list = os.listdir(read_folder_path)
# 类间循环
for folder in folder_list:
#某类下的路径
new_folder_path = os.path.join(read_folder_path, folder)
# 创建保存文件目录
path=write_folder_path+folder #保存文件的子文件
isExists=os.path.exists(path)
if not isExists:
os.makedirs(path)
print(path+' 创建成功')
else: pass
save_folder_path = os.path.join(write_folder_path, folder)#某类下的保存路径
print('--> 请稍等,正在处理中...')
# 类内循环
files = os.listdir(new_folder_path)
j = 1
for file in files:
if j > len(files): break
dealpath = os.path.join(new_folder_path, file) #处理单个文件的路径
with open(dealpath,"r",encoding='utf-8') as f:
txtlist=f.read()
# python 过滤中文、英文标点特殊符号
# txtlist1 = re.sub("[\s+\.\!\/_,$%^*(+\"\']+|[+——!,。?、~@#¥%……&*()]+", "",txtlist)
words =pseg.cut(txtlist) # 带词性标注的分词结果
cutresult="" # 单个文本:分词后经停用词处理后的结果
for word, flag in words:
if word not in stopwords:
cutresult += word+"/"+flag+" " #去停用词
savepath = os.path.join(save_folder_path,file)
getFlag(cutresult,savepath)
j += 1
'''
做词性筛选
cutresult:str类型,初切分的结果
savepath: 保存文件路径
'''
def getFlag(cutresult,savepath):
txtlist=[] #过滤掉的词性后的结果
#词列表为自己定义要过滤掉的词性
cixing=["/x","/zg","/uj","/ul","/e","/d","/uz","/y"]
for line in cutresult.split('\n'):
line_list2=re.split('[ ]', line)
line_list2.append("\n") # 保持原段落格式存在
line_list=line_list2[:]
for segs in line_list2:
for K in cixing:
if K in segs:
line_list.remove(segs)
break
else:
pass
txtlist.extend(line_list)
# 去除词性标签
resultlist=txtlist[:]
flagresult=""
for v in txtlist:
if "/" in v:
slope=v.index("/")
letter=v[0:slope]+" "
flagresult+= letter
else:
flagresult+= v
standdata(flagresult,savepath)
'''
标准化处理,去除空行,空白字符等。
flagresult:筛选过的结果
'''
def standdata(flagresult,savepath):
f2=open(savepath,"w",encoding='utf-8')
for line in flagresult.split('\n'):
if len(line)>=2:
line_clean="/ ".join(line.split())
lines=line_clean+" "+"\n"
f2.write(lines)
else: pass
f2.close()
if __name__ == '__main__' :
t1=time.time()
# 测试单个文件
dealpath="../Database/SogouC/FileTest/1.txt"
savepath="../Database/SogouCCut/FileTest/1.txt"
stopwordspath='../Database/stopwords/CH_stopWords.txt'
stopwordspath1='../Database/stopwords/HG_stopWords.txt' # 哈工大停用词表
# 批量处理文件夹下的文件
# rfolder_path = '../Database/SogouC/Sample/'
rfolder_path = '../Database/SogouC/FileNews/'
# 分词处理后保存根路径
wfolder_path = '../Database/SogouCCut/'
# 中文语料预处理器
# cutTxtWord(dealpath,savepath,stopwordspath) # 单文本预处理器
cutFileWord(rfolder_path,wfolder_path,stopwordspath) # 多文本预处理器
t2=time.time()
print("中文语料语处理完成,耗时:"+str(t2-t1)+"秒。") #反馈结果