#!/bin/bash
# coding=utf-8
import re
import sys
import os
import util as u
'''
功能说明:先groupby某一列，然后统计每个类的数量,输出一份文件
'''
##############参数说明###############
SOURCE_FILE = 'hebing.csv' # 输入文件
RESULT_FILE = 'result.csv' # 输出文件
COLUMN = 1 # 要groupby和统计的列的列
####################################
count_dict = {}
if __name__ == '__main__':
source_file_head = u.create_file_head(SOURCE_FILE, 'right', ['次数'])
source_file_body = u.create_file_body(SOURCE_FILE)
for line in source_file_body:
content = u.create_content(line, COLUMN)
count_dict[content] = count_dict.get(content, 0) + 1
result_file = file(RESULT_FILE, 'w+')
result_file.write(source_file_head)
for key, value in count_dict.items():
result_file.write(key + ',' + str(value) + '\n')
result_file.close()
result = '.*' + str.replace(',', '.*|.*') + '.*'
return result
if __name__ == '__main__':
try:
LABEL_FILE, LABEL_PATH = u.getFirstFile('txt')
SOURCE_FILE, SOURCE_PATH = u.getFirstFile('csv')
print u.utf8_2_gbk('打标签文件:' + LABEL_FILE)
print u.utf8_2_gbk('数据源文件:' + SOURCE_FILE)
source_file_body = u.create_file_body(SOURCE_PATH)
for num, line in enumerate(source_file_body):
source_file_body[num] = line.strip().lower() + ',' + '\n'
labelType, labelNum = createMoreMatch(LABEL_PATH)
matchHead = u.create_file_head(SOURCE_PATH, 'right', [LABEL_FILE.split('.')[0]])
print u.utf8_2_gbk('标签个数:' + str(labelNum) + '个')
for key, value in labelType.items():
count += 1
print u.utf8_2_gbk('当前执行到第' + str(count) + '个')
words = value.strip().split('|')
if len(words) == 1:
c = createPattern(words[0], ACCURATE)
p = re.compile(c)
for num, line in enumerate(source_file_body):
content = u.create_content(line, COLUNM)
if p.match(content):
source_file_body[num] = source_file_body[num].strip() + key + '|' + '\n'
keyWordCount[key] = keyWordCount.get(key, 0) + 1
if len(words) == 2:
c = createPattern(words[0], ACCURATE)
head = linecache.getline(u.utf8_2_gbk(SOURCE_FILE), 1).strip()
TOTALCOLUNM = len(head.split(','))
print u.utf8_2_gbk('标签词个数:') + u.printDictLen(labelWordp)
source_file_body = u.create_file_body(SOURCE_FILE)
for key, value in labelWordp.items():
i += 1
print u.utf8_2_gbk('当前执行到{0}个'.format(i))
for num, line in enumerate(source_file_body):
data = line.strip().split(',')
if len(data) == TOTALCOLUNM + 1:
continue
content = data[COLUNM - 1]
p = re.compile(value)
if p.match(content):
source_file_body[num] = source_file_body[num].strip() + ',' + key + '\n'
keyWordCount[key] = keyWordCount.get(key, 0) + 1
# 补全格式
for num, line in enumerate(source_file_body):
data = line.strip().split(',')
if len(data) == TOTALCOLUNM + 1:
continue
source_file_body[num] = source_file_body[num].strip() + ',' + '' + '\n'
result_file_head = u.create_file_head(SOURCE_FILE, 'right', [u.gbk_2_utf8(columnName)])
u.create_result_file(RESULT_FILE, result_file_head, source_file_body)
KEYWORD_FILE = LABELWORD.split('.')[0] + '统计.txt'
u.writeDictFile(KEYWORD_FILE, keyWordCount) # 输出统计结果
def createPattern(match_words):
pattern = ''
cmatch_words = map(lambda line: '.*' + line.strip() + '.*', match_words[0])
ematch_words = map(lambda line: '.*' + line.strip() + '.*', match_words[1])
if len(ematch_words) == 0:
pattern = '|'.join(cmatch_words)
else:
pattern = '|'.join(cmatch_words) + '|' + '|'.join(ematch_words)
return pattern
if __name__ == '__main__':
source_file_body = u.create_file_body(SOURCE_FILE)
source_file_head = u.create_file_head(SOURCE_FILE)
match_words = u.create_match_words(MATCH_FILE)
pattern = createPattern(match_words)
p = re.compile(pattern)
for line in source_file_body:
content = u.create_content(line, COLUMN)
if p.match(content):
resultFile.append(line)
else:
removeFile.append(line)
resultFileName = getFileName(SOURCE_FILE, '-含关键词.csv')
removeFileName = getFileName(SOURCE_FILE, '-不含关键词.csv')
for key, value in res.items():
rowNum_list = value[2:].split(",")
if len(rowNum_list) >= NUMBER:
for num in rowNum_list:
remove_list.append(num)
else:
for num in rowNum_list:
save_list.append(num)
if __name__ == "__main__":
try:
print u.utf8_2_gbk('开始执行')
result_file_head = u.create_file_head(SOURCE_FILE) # 文件标题
result_file_body = u.create_file_body(SOURCE_FILE) # 文件内容
factory(result_file_body) # 构造输出文件
for num in save_list:
save_file_list.append(result_file_body[int(num) - 1])
for num in remove_list:
remove_file_list.append(result_file_body[int(num) - 1])
print u.utf8_2_gbk(SAVE_FILE + '行数:' + str(len(save_file_list)))
print u.utf8_2_gbk(REMOVE_FILE + '行数:' + str(len(remove_file_list)))
u.create_result_file(REMOVE_FILE, result_file_head,
remove_file_list) # 符合条件的输出文件（大于等于101次）
u.create_result_file(SAVE_FILE, result_file_head,
save_file_list) # 不符合条件的输出文件
except:
# coding=utf-8
import sys
import re
import util as u
import os
'''
功能说明:将指定路径的文件合并成一个文件(文件格式为csv)
'''
##################参数说明#################
FILE_PATH = r'G:\merge_n_file\data' # 文件路径
RESULT_FILE = 'result.csv' # 合并之后的文件
RESULT_HEAD = r'G:\merge_n_file\data\3779_利鑫－999道私房菜.csv' # 指定合并后的文件标题基准
##########################################
file_dict = {}
if __name__ == '__main__':
file_list = u.GetFileList(u.utf8_2_gbk(FILE_PATH), [])
for f in file_list:
file_dict[f.encode('gbk')] = u.create_file_body(f.encode('utf-8'))
result_file = file(RESULT_FILE, 'w+')
result_file_head = u.create_file_head(RESULT_HEAD)
result_file.write(result_file_head)
for key, value in file_dict.items():
result_file.writelines(value)
result_file.close()
行数----1000----10000----100000----1000000--
耗时-----2s------11s------129s------1432s---
内存----0.3mb----3mb------33mb------400mb---
"""
__author__ = "liangzhicheng"
SOURCENAME, SOURCEPATH = u.getFirstFile('csv')
cluster = {}
result_file_body = []
pattern = re.compile("\w|[/.,/#@$%^& ]")
count_file_dict = {}
if __name__ == '__main__':
try:
source_file_head = u.create_file_head(SOURCEPATH, 'left', ['类型'])
source_file_body = u.create_file_body(SOURCEPATH)
print u.utf8_2_gbk('开始执行聚类')
for num, line in enumerate(source_file_body):
content = re.sub(pattern, '', u.create_content(line, COLUMN))
if len(content) <= 20:
keywords = jieba.analyse.extract_tags(content, topK=2)
else:
keywords = jieba.analyse.extract_tags(content, topK=TOPKET)
keywords.sort()
key = ','.join(keywords)
cluster[key] = str(cluster.get(key, 0)) + "," + str(num + 1)
print u.utf8_2_gbk('聚类完成,生成输出文件')
for num, value in enumerate(cluster.itervalues()):
cluster_list = value[2:].split(',')
count_file_dict[num] = len(cluster_list)
content = ''
f = open(u.utf8_2_gbk(fileName))
for line in f:
if line[:-1].strip():
if ACCURATE:
content += '|' + '^' + line.strip() + '$'
else:
content += '|' + '.*' + line.strip() + '.*'
f.close()
return content[1:].lower()
if __name__ == '__main__':
try:
source_file_body = u.create_file_body(SOURCEPATH)
source_file_head = u.create_file_head(SOURCEPATH)
m = createPattern(LABELPATH)
print m
print '===============>>' + u.utf8_2_gbk(
'若乱码,匹配词文件请使用gbk编码') + '<<==================='
p = re.compile(m)
print u.utf8_2_gbk('数据源文件行数:') + str(len(source_file_body))
for line in source_file_body:
content = u.create_content(line, COLUMN).lower()
if p.match(content):
result_list.append(line)
else:
remove_list.append(line)
print u.utf8_2_gbk('不包含关键词行数:') + str(len(remove_list))
print u.utf8_2_gbk('包含关键词行数:') + str(len(result_list))
u.create_result_file(u.changeFileName(SOURCE, '-含关键词.csv'),
将结果按照每个cluster的大小排序，大的在前面
性能:
行数----1000----10000----100000----1000000--
耗时-----2s------11s------129s------1432s---
内存----0.3mb----3mb------33mb------400mb---
"""
__author__ = "liangzhicheng"
cluster = {}
result_file_body = []
pattern = re.compile("\w|[/.,/#@$%^& ]")
count_file_list = []
if __name__ == '__main__':
source_file_head = u.create_file_head(SOURCE_FILE, 'left', ['类型'])
source_file_body = u.create_file_body(SOURCE_FILE)
for num, line in enumerate(source_file_body):
content = re.sub(pattern, '', u.create_content(line, COLUMN))
if len(content) <= 20:
keywords = jieba.analyse.extract_tags(content, topK=2)
else:
keywords = jieba.analyse.extract_tags(content, topK=TOPKET)
keywords.sort()
key = ','.join(keywords)
cluster[key] = str(cluster.get(key, 0)) + "," + str(num + 1)
for num, value in enumerate(cluster.itervalues()):
cluster_list = value[2:].split(',')
count_file_list.append(str(num) + '\t' + str(len(cluster_list)) + '\n')