kmeans_cluster.py 文件源码

def cluster_process(filenames,key_part,s,n_clusters=2):
    documents=[]
    texts=[]
    for fname in filenames:
        # key_part: # ?? ???# ?? ???# ???# ???# ???# ???
        # # ?? ???# ?? ???# ???# ???# ??
        # = = = ?? = = =   = = = ?? = = =   = = = ?? = = =
        # key_part=['# ??']
        # ??key_part??
        d=extract(fname,key_part=key_part)
        documents.append(d)

    # documents,words=tfidf(documents)
    # print len(documents),len(words)
    # docs=create_format_mat(documents,words)

    docs=get_tfidf(documents)  # ??gensim??tfidf

    # ??
    # labels [0,1,0,1,1,...]
    labels,score=clustering(docs,n_clusters)

    print 'key_part:','_'.join(key_part).decode('utf-8')
    item_parts=[]
    filename_parts=[]
    for i in range(n_clusters):
        # item=[filenames[j] for j in range(len(labels)) if labels[j]==i]
        item=[documents[j] for j in range(len(labels)) if labels[j]==i]
        # print cchardet.detect(s)
        # ?????
        filename_parts.append(([filenames[j] for j in range(len(labels)) if labels[j]==i],u'%s_%s_%d' %(s,'_'.join(key_part),i)))
        item_parts.append((item,u'%s_%s_%d' %(s,'_'.join(key_part),i)))
        print 'class_%d:%d' %(i,len(item))
    # ?????
    print 'score:',score
    print '-'*20
    write2file(item_parts)  # ????
    return filename_parts