当前位置: 首页>>代码示例>>Python>>正文


Python Space.get_neighbours方法代码示例

本文整理汇总了Python中composes.semantic_space.space.Space.get_neighbours方法的典型用法代码示例。如果您正苦于以下问题:Python Space.get_neighbours方法的具体用法?Python Space.get_neighbours怎么用?Python Space.get_neighbours使用的例子?那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在composes.semantic_space.space.Space的用法示例。


在下文中一共展示了Space.get_neighbours方法的1个代码示例,这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞,您的评价将有助于系统推荐出更棒的Python代码示例。

示例1: TextExtraction

# 需要导入模块: from composes.semantic_space.space import Space [as 别名]
# 或者: from composes.semantic_space.space.Space import get_neighbours [as 别名]
def TextExtraction(text,nthreads):
    pattern = re.compile('\W')   
    # Split the text into data parts
    data=[]; # Data container. This one is a list whose elements are list of 
            # tokens that will be proccessed by a thred. This token itself 
            #is a list or tokens inside a document. It is a list of list of list
    for i in xrange(0,nthreads):
        data.append([]);
    # Push data to list
    ln_count=0;
    text2 = set([]);
    for ln in set(text):
        ln = ln.replace(".", "").replace(",", "").replace("-", " ");
        ln = re.sub(pattern, ' ', ln.strip());
        ln = re.sub(' +',' ',ln);
        text2.add(ln.strip(" \n"));
#    print text2;
    for line in text2:
        ln_count+=1;
        tokens=line.strip().split(" ");
        data[ln_count%nthreads-1].append(tokens);
    print "      Splited data to ",len(data), " parts";
#   for ds in data:
#       print len(ds);
        
    # Now start the threads
    threads=[];
    try:
        for i in xrange(0,nthreads):
            new_thread = Extract_vectors("Thread-"+str(i),data[i]);
            new_thread.start();
            threads.append(new_thread);
    except:
        print "Error: unable to run multithreading";
        traceback.print_exc();
        
     # Wait for all threads to complete
    for t in threads:
        t.join()
        
    # Release data
    data=None;
    #my_space=None;
    
    # Collect results
    final_matrix = threads[0].resMatrix;
    rows = threads[0].proceedRows; 
    total_succecced = threads[nthreads-1].nsucceed;
    total_proccessed = threads[nthreads-1].size;
    print "      --> Size of result matrix of ",threads[nthreads-1].threadName\
          ,"is ",threads[nthreads-1].resMatrix.get_shape();
    for i in xrange(1,nthreads,1):
        final_matrix = final_matrix.vstack(threads[i].resMatrix); # Stack mats
        total_succecced += threads[i].nsucceed;
        total_proccessed += threads[i].size;     
        print "      --> Size of result matrix of ", threads[i].threadName ,\
              "is ",threads[i].resMatrix.get_shape();
        rows.extend(threads[i].proceedRows);  
    print "      Size of vector final matrix: ", final_matrix.get_shape(); 
    print "      Successed ",total_succecced, " in ",total_proccessed," with "\
          , total_proccessed - total_succecced,\
          " unsupported document(s) was corrected.";
    cols = [i for i in xrange(0,space_dim)];
    try:
        print "      --> Starting space build. Size to matrix " + \
                str(final_matrix.get_shape()) + \
                ". Size of rows " + str(len(rows)) +\
                ". Size of columns " + str(len(cols));
        heads_space = Space(final_matrix,rows,cols,None,None,[],None);
        print "          Toy neigbors: ",heads_space.get_neighbours(rows[0], 3\
              ,CosSimilarity());
 #       print heads_space.get_id2row();
    except:
        print "/!\ Error building semantic space!"
        traceback.print_exc();
   # Now write to file
#    space_out = open(sys.argv[3]+".space", 'w')
    return heads_space;
#    pickle.dump(heads_space, space_out);
#    space_out.close(); # Close space file
    print "      Saved the space file!"
开发者ID:nduy,项目名称:stock-trend-predict-headline,代码行数:83,代码来源:utils.py


注:本文中的composes.semantic_space.space.Space.get_neighbours方法示例由纯净天空整理自Github/MSDocs等开源代码及文档管理平台,相关代码片段筛选自各路编程大神贡献的开源项目,源码版权归原作者所有,传播和使用请参考对应项目的License;未经允许,请勿转载。