|
|
|
|
|
'''
|
|
|
|
|
|
使用 multiprocessing.Manager:
|
|
|
|
|
|
Manager 提供了一个可以在不同进程之间共享和修改的数据类型，如 list, dict, Namespace 等。
|
|
|
|
|
|
它实际上是在背后启动了一个单独的服务器进程，其他进程通过代理来访问这些共享对象。
|
|
|
|
|
|
'''
|
|
|
|
|
|
# 使用 multiprocessing.Manager 来完成统计词频
|
|
|
|
|
|
# 怎么得到最快的一个结果，是一个试错过程：X程创建数目多少、分片的大小 ...
|
|
|
|
|
|
|
|
|
|
|
|
from cppy.cp_util import *
|
|
|
|
|
|
from collections import Counter
|
|
|
|
|
|
from multiprocessing import Manager, Process
|
|
|
|
|
|
|
|
|
|
|
|
stop_words = get_stopwords()
|
|
|
|
|
|
|
|
|
|
|
|
def process_chunk(chunk,word_count):
|
|
|
|
|
|
words = [ w for w in chunk if ( not w in stop_words ) and len(w) >= 3 ]
|
|
|
|
|
|
for word in words: # 非常化时间
|
|
|
|
|
|
word_count[word] = word_count.get(word, 0) + 1
|
|
|
|
|
|
# word_count.update( Counter(words) ) # 类型不起作用
|
|
|
|
|
|
|
|
|
|
|
|
@timing_decorator
|
|
|
|
|
|
def main():
|
|
|
|
|
|
manager = Manager()
|
|
|
|
|
|
word_count = manager.dict()
|
|
|
|
|
|
|
|
|
|
|
|
chunks = get_chunks(testfilepath,2800)
|
|
|
|
|
|
print('-------------------',len(chunks))
|
|
|
|
|
|
processes = []
|
|
|
|
|
|
for chunk in chunks:
|
|
|
|
|
|
p = Process(target=process_chunk,
|
|
|
|
|
|
args=(chunk,word_count) )
|
|
|
|
|
|
processes.append(p)
|
|
|
|
|
|
p.start()
|
|
|
|
|
|
|
|
|
|
|
|
for p in processes: p.join()
|
|
|
|
|
|
|
|
|
|
|
|
word_count = dict(word_count)
|
|
|
|
|
|
word_freqs = Counter(word_count).most_common(10)
|
|
|
|
|
|
print_word_freqs(word_freqs)
|
|
|
|
|
|
|
|
|
|
|
|
if __name__ == '__main__':
|
|
|
|
|
|
main()
|
