__author__?=?‘LiNing‘??
#coding:?utf-8
import?os
import?sys
import?time
from?sys?import?exit
from?os?import?listdir
from?os?import?makedirs
from?os?import?remove
from?os.path?import?isdir
from?os.path?import?exists
from?os.path?import?join
from?os.path?import?split
from?shutil?import?copy
from?shutil?import?rmtree
from?sklearn?import?metrics
import?numpy?as?np
import?pylab?as?pl
import?nltk
import?math

from?TextProcess?import?TextRmSame
from?TextProcess?import?TextRename
from?TextProcess?import?MakeStopWordsList
from?TextProcess?import?TextProcessing
from?TextProcess?import?MakeAllWordsList
from?TextProcess?import?MakeFeatureWordsDict

from?TextFeature?import?TextBool
from?TextFeature?import?ComputeTf
from?TextFeature?import?ComputeTfIdf
from?TextFeature?import?ExtractTags

from?TextClassify?import?TextClassifier
from?TextClassify?import?TextClassifier_nltk
from?TextClassify?import?Vote

from?TextEvaluation?import?calculate_result

##import?scipy.io?as?sio
##sio.savemat（‘file.mat‘?dict）
##data?=?sio.loadmat（‘file.mat‘）



if?__name__?==?‘__main__‘:
????##參數設定
????lag?=?“chs“?#chs?or?eng??
????Test?=?“False“?#True?or?False
????GenerateLogs?=?“False“?#True?or?False
????GenerateKeywords?=?“False“?#True?or?False
????GenerateFolders?=?“True“?#True?or?False?

????
????##生成results.log文件
????if?GenerateLogs?==?“True“:
	#sys.stdout?=?open（‘results.log‘?‘a‘）?#追加模式
	sys.stdout?=?open（‘results.log‘?‘wb‘）????
	pass
????
????starttime?=?time.clock（）?#開始時間
????print?“start“
????
????
????##訓練文本預處理
????##Wing?IDE沒有輸入參數，所以才會出現list?index?out?of?range
????##raw_train_container_path?=?sys.argv[1]?#datas\\train_raw
????#train_container_path?=?sys.argv[1]?#datas\\train?
????#test_container_path?=?sys.argv[2]?#datas\\test?or?datas\\unknown?
????#訓練集
????raw_train_container_path?=?“datas\\train_raw“
????train_container_path?=?“datas\\train“?
????#中斷程序條件運行
????（a?b）?=?split（train_container_path）
????error_files_path?=?join（a?“errorfiles“）	????
????if?exists（error_files_path）:
	rmtree（error_files_path）?
????else:
	TextRmSame（raw_train_container_path?train_container_path）?#remove?the?same?files
	#TextRename（raw_train_container_path?train_container_path）?#rename?the?filename
	pass
????#測試集
????if?Test?==?“True“:
	test_container_path?=?“datas\\test“
????else:
	test_container_path?=?“datas\\unknown“
????#停用詞
????stopwords_file?=?“stopwords.txt“
????stopwords_list?=?MakeStopWordsList（stopwords_file）

????
????##文本信息提取，若生成errorfiles文件夾則需重新運TextProcessing
????##（train_data?train_filenames?train_target?train_target_names）?=?TextProcessing（train_container_path?lag?‘True‘?stopwords_list）?
????（train_data?train_filenames?train_target?train_target_names）?=?TextProcessing（train_container_path?lag?‘True‘）?
????if?Test?==?“True“:
	##（test_data?test_filenames?test_target?test_target_names）?=?TextProcessing（test_container_path?lag?‘True‘?stopwords_list）?
	（test_data?te

?屬性????????????大小?????日期????時間???名稱
-----------?---------??----------?-----??----
?????文件?????5420898??2014-05-12?14:08??文本分類系統\jieba_dict.txt
?????文件???????10701??2014-06-17?10:47??文本分類系統\Main.py
?????文件????????9839??2014-06-13?15:44??文本分類系統\stopwords.txt
?????文件????????3398??2014-06-16?16:00??文本分類系統\TextClassify.pyc
?????文件?????????867??2014-06-16?16:00??文本分類系統\TextEvaluation.pyc
?????文件????????2564??2014-06-16?16:00??文本分類系統\TextFeature.pyc
?????文件????????6895??2014-06-17?10:24??文本分類系統\TextProcess.pyc