import jieba.posseg as pseg
import codecs
import pickle
import warnings
warnings.filterwarnings(action='ignore', category=UserWarning, module='gensim')
from gensim import corpora, models, similarities
from gensim.summarization import bm25
import os
from django.conf import settings 


basedir = settings.BASE_DIR +'/duibiao/'
#basedir = 'F:/project/safeyun/duibiao/'
#初始化部分数据
#设定停用词
stop_words = basedir +  'chineseStopWords.txt'
stopwords = codecs.open(stop_words,'r').readlines()
stopwords = [ w.strip() for w in stopwords ]
stop_flag = ['x', 'c', 'u','d', 'p', 't', 'uj', 'm', 'f', 'r']

#标准文件列表
dirname = basedir + 'biao/'
filename = [] #文件名列表
#词袋列表
corpus = []

#分词函数
def tokenization(filename):
    result = []
    with open(dirname + filename, 'r', encoding = 'utf-8') as f:
        text = f.read()
    words = pseg.cut(text)
    for word, flag in words:
        if flag not in stop_flag and word not in stopwords:
            result.append(word)
    return result

def tokenizationFromStr(text):
    result = []
    words = pseg.cut(text)
    for word, flag in words:
        if flag not in stop_flag and word not in stopwords:
            result.append(word)
    return result

def genmodel():
    #获取标准全部分词
    for root,dirs,files in os.walk(dirname):
        for f in files:
            if f.endswith(".txt"):
                #f = dirname + '/' + f
                corpus.append(tokenization(f))
                filename.append(f)
                
    #建立词袋模型并制作语料库      
    dictionary = corpora.Dictionary(corpus)
    doc_vectors = [dictionary.doc2bow(text) for text in corpus]

    #针对语料库建立tfidf模型
    tfidf = models.TfidfModel(doc_vectors)
    tfidf_vectors = tfidf[doc_vectors]

    #将语料库和字典序列化保存
    pickle.dump(dictionary, open(basedir + 'biaom/dictionary.pkl','wb'))
    pickle.dump(tfidf, open(basedir + 'biaom/tfidf.m', 'wb'))
    pickle.dump(tfidf_vectors, open(basedir + 'biaom/tfidfv.m', 'wb'))
    pickle.dump(filename, open(basedir + 'biaom/filename', 'wb'))

    
def calsim(txtpath):
    #取得序列化的模型
    dictionary = pickle.load(open(basedir + 'biaom/dictionary.pkl','rb'))
    tfidf = pickle.load(open(basedir + 'biaom/tfidf.m','rb'))
    tfidf_vectors = pickle.load(open(basedir + 'biaom/tfidfv.m','rb'))
    filename = pickle.load(open(basedir + 'biaom/filename','rb'))    

    #获取所要查询的文本并计算tfidf值
    query = tokenization(txtpath)
    query_bow = dictionary.doc2bow(query)
    query_tfidf = tfidf[query_bow]

    #计算相似度并排序
    index = similarities.MatrixSimilarity(tfidf_vectors)
    sims = index[query_tfidf]
    paixu = sorted(enumerate(sims), key=lambda item: -item[1])

    #打印出最相似的几条标准
    thestr = '最相似的安全标准是:' + '<br/>'
    for i in range(5):
        if paixu[i][1] > 0:
            filepath = dirname  + filename[paixu[i][0]]
            with open(filepath,'r', encoding = 'utf-8') as f:
                thestr = thestr + f.read()+'<br/>------------<br/>'

    return thestr


def calsimTojson(text):
    #取得序列化的模型
    dictionary = pickle.load(open(basedir + 'biaom/dictionary.pkl','rb'))
    tfidf = pickle.load(open(basedir + 'biaom/tfidf.m','rb'))
    tfidf_vectors = pickle.load(open(basedir + 'biaom/tfidfv.m','rb'))
    filename = pickle.load(open(basedir + 'biaom/filename','rb'))    

    #获取所要查询的文本并计算tfidf值
    query = tokenizationFromStr(text)
    query_bow = dictionary.doc2bow(query)
    query_tfidf = tfidf[query_bow]

    #计算相似度并排序
    index = similarities.MatrixSimilarity(tfidf_vectors)
    sims = index[query_tfidf]
    paixu = sorted(enumerate(sims), key=lambda item: -item[1])

    #返回最相似的五条标准
    thelist = []
    for i in range(5):
        if paixu[i][1] > 0:
            filepath = dirname  + filename[paixu[i][0]]
            with open(filepath,'r', encoding = 'utf-8') as f:
                thelist.append(f.read())

    return thelist
    
if __name__ == '__main__':
    genmodel()
    #thestr = calsim(basedir + 'onetrouble.txt')
    #print(thestr)