Question

目标是创建一个堆叠的条形图，以显示总共360秒（以秒为单位）的tweet（我从tweepy得到的）情绪。我有两个清单。第一个按时间顺序对推文进行情感分析，第二个按时间顺序对每秒的推文量进行分析。

list1 = ("neg", "pos", "pos", "neu", "neg", "pos", "neu", "neu",...)
list2 = (2, 1, 3, 2,...)

现在，我想创建某种嵌套循环，并使用list2来计算列表1中的项目。然后，我将拥有3个列表，其中每个可用于图形的情感具有360个值。它应该给我类似以下的输出：

lis_negative = (1, 0, 1, 0, ...)
lis_positive = (1, 1, 1, 0, ...)
lis_neutral = (0, 0, 1, 2, ...)

如何创建此循环，也许有更简单的方法吗？我宁愿不使用除matplotlib之外的任何库。

Answer 1

代码：

from itertools import islice
from collections import Counter

def categorize(clas, amounts):
    cats = {'neg': [], 'pos': [], 'neu': []}
    clas = iter(clas)

    for a in amounts:
        cs = Counter(islice(clas, a)) # take a items
        for cat in cats:
            cats[cat].append(cs[cat])
    return cats

演示：

>>> t1 = ('neg', 'pos', 'pos', 'neu', 'neg', 'pos', 'neu', 'neu')
>>> t2 =  (2, 1, 3, 2)
>>> 
>>> categorize(t1, t2)
{'neg': [1, 0, 1, 0], 'neu': [0, 0, 1, 2], 'pos': [1, 1, 1, 0]}

根据要求，提供不导入的解决方案：

def make_counter(iterable):
    c = {}
    for x in iterable:
        c[x] = c.get(x, 0) + 1
    return c

def categorize(clas, amounts):
    cats = {'neg': [], 'pos': [], 'neu': []}
    pos = 0

    for a in amounts:
        chunk = clas[pos:pos+a]
        pos += a
        cs = make_counter(chunk)
        for cat in cats:
            cats[cat].append(cs.get(cat, 0))
    return cats

编辑：更短的无进口解决方案：

def categorize(clas, amounts):
    cats = {k:[0]*len(amounts) for k in ('neg', 'pos', 'neu')}
    pos = 0

    for i, a in enumerate(amounts):
        chunk = clas[pos:pos+a]
        pos += a
        for c in chunk:
            cats[c][i] += 1

    return cats

根据第二个列表中的位数对列表项进行分组

1 个答案: