从文件中的 n 行块中提取项目，计算每个块的项目频率，Python

Question

我有一个文本文件，其中包含 5 行制表符分隔行块：

 1 \t DESCRIPTION \t SENTENCE \t ITEMS

 1 \t DESCRIPTION \t SENTENCE \t ITEMS

 1 \t DESCRIPTION \t SENTENCE \t ITEMS

 1 \t DESCRIPTION \t SENTENCE \t ITEMS

 1 \t DESCRIPTION \t SENTENCE \t ITEMS

 2 \t DESCRIPTION \t SENTENCE \t ITEMS

 2 \t DESCRIPTION \t SENTENCE \t ITEMS

 2 \t DESCRIPTION \t SENTENCE \t ITEMS

 2 \t DESCRIPTION \t SENTENCE \t ITEMS

 2 \t DESCRIPTION \t SENTENCE \t ITEMS

等等。

在每个块中，DESCRIPTION 和 SENTENCE 列是相同的。感兴趣的数据位于 ITEMS 列中，该列对于块中的每一行都不同，并且采用以下格式：

word1, word2, word3

...等等

对于每个 5 行块，我需要计算 word1、word2 等的频率. 在项目中。例如，如果第一个 5 行块如下所示

 1 \t DESCRIPTION \t SENTENCE \t word1, word2, word3

 1 \t DESCRIPTION \t SENTENCE \t word1, word2

 1 \t DESCRIPTION \t SENTENCE \t word4

 1 \t DESCRIPTION \t SENTENCE \t word1, word2, word3

 1 \t DESCRIPTION \t SENTENCE \t word1, word2

，则该 5 行块的正确输出将是

1, SENTENCE, (word1: 4, word2: 4, word3: 2, word4: 1)

Ie，块编号后跟句子，后跟单词的频率计数。

我有一些代码来提取五行块，并在提取后计算块中单词的频率，但我坚持隔离每个块、获取单词频率、继续下一个块等任务。

from itertools import groupby 

def GetFrequencies(file):
    file_contents = open(file).readlines()  #file as list
    """use zip to get the entire file as list of 5-line chunk tuples""" 
    five_line_increments = zip(*[iter(file_contents)]*5) 
    for chunk in five_line_increments:  #for each 5-line chunk... 
        for sentence in chunk:          #...and for each sentence in that chunk
            words = sentence.split('\t')[3].split() #get the ITEMS column at index 3
            words_no_comma = [x.strip(',') for x in words]  #get rid of the commas
            words_no_ws = [x.strip(' ')for x in words_no_comma] #get rid of the whitespace resulting from the removed commas


       """STUCK HERE   The idea originally was to take the words lists for 
       each chunk and combine them to create a big list, 'collection,' and
       feed this into the for-loop below."""





    for key, group in groupby(collection): #collection is a big list containing all of the words in the ITEMS section of the chunk, e.g, ['word1', 'word2', word3', 'word1', 'word1', 'word2', etc.]
        print key,len(list(group)),    

Answer 1

使用 python 2.7

#!/usr/bin/env python

import collections

chunks={}

with open('input') as fd:
    for line in fd:
        line=line.split()
        if not line:
            continue
        if chunks.has_key(line[0]):
            for i in line[3:]:
                chunks[line[0]].append(i.replace(',',''))
        else:
            chunks[line[0]]=[line[2]]

for k,v in chunks.iteritems():
    counter=collections.Counter(v[1:])
    print k, v[0], counter

输出：

1 SENTENCE Counter({'word1': 3, 'word2': 3, 'word4': 1, 'word3': 1})

Answer 2

标准库中有一个 csv 解析器可以为您处理输入拆分

import csv
import collections

def GetFrequencies(file_in):
    sentences = dict()
    with csv.reader(open(file_in, 'rb'), delimiter='\t') as csv_file:
        for line in csv_file:
            sentence = line[0]
            if sentence not in sentences:
                sentences[sentence] = collections.Counter()
            sentences[sentence].update([x.strip(' ') for x in line[3].split(',')])

Answer 3

稍微编辑一下你的代码，我认为它做了你想要它做的事情：

file_contents = open(file).readlines()  #file as list
"""use zip to get the entire file as list of 5-line chunk tuples""" 
five_line_increments = zip(*[iter(file_contents)]*5) 
for chunk in five_line_increments:  #for each 5-line chunk...
    word_freq = {} #word frequencies for each chunk
    for sentence in chunk:          #...and for each sentence in that chunk
        words = "".join(sentence.split('\t')[3]).strip('\n').split(', ') #get the ITEMS column at index 3 and put them in list
        for word in words:
            if word not in word_freq:
                word_freq[word] = 1
            else:
                word_freq[word] += 1


    print word_freq

输出：

{'word4': 1, 'word1': 4, 'word3': 2, 'word2': 4}

Answer 4

总结一下：如果所有“单词”不是“DESCRIPTION”或“SENTENCE”，您想将它们附加到集合中吗？试试这个：

for word in words_no_ws:
    if word not in ("DESCRIPTION", "SENTENCE"):
        collection.append(word)