使用关键字作为Python 2.7中的起点从文件中检索文本块？

Question

我正在尝试根据段落标题（第一行）从文件中读取一段文字。 例如，假设文件是​​这样的：

Paragraph 1:1
This paragraph 1. This paragraph 1. This paragraph 1. This paragraph 1. This  
paragraph 1. This paragraph 1. This paragraph 1. This paragraph 1. This paragraph 
1. This paragraph 1. This paragraph 1. This paragraph 1. This paragraph 1. This 
paragraph 1. This paragraph 1. This paragraph 1. This paragraph 1. This paragraph 

Paragraph 1:2
This is paragraph 2. This is paragraph 2. This is paragraph 2. This is paragraph 
2. This is paragraph 2. This is paragraph 2. This is paragraph 2. This is paragraph 
2. This is paragraph 2. This is paragraph 2. This is paragraph 2. This is paragraph 
2. This is paragraph 2. This is paragraph 2. This is paragraph 2. This is paragraph 2  

我有自己的代码，以便可以随机选择标题：

def get_paragraph():
    f = open('file.txt','r')
    n = [str(x) for x in range(3)]
    paragraph_list = []

    for line in f:
        for numbers in n: # I know the following code isn't very pretty...
            if (number in line) and (':' in line) and (len(line) < 20):
                paragraph_list.append(line)

    selection = random.choice(paragraph_list)
    print(selection)

get_paragraph()  

现在，我希望能够读取与第一行相关的其余段落，并以空白行结尾-以便输出将是从文件中随机选择的段落。

任何建议或帮助将不胜感激！

Answer 1

我会采取不同的方式

从集合导入defaultdict

def get_paragraph(file_path)

    myString = open(file_path.read()
    paragraphList = myString.split('\n\n') # split by adjacent newlines
    paragraphList = [paragraph.split('\n') for paragraph in paragraphList]
    paragrpahDictionary = defaultdict(list)
    for paragraph in paragraphList:
        paragraphNumber = int(paragraph[0].split(:)[-1])
        paragraphBody = paragraph[1:]
        paragraphDictionary[paragraphNumber] = paragraphBody

   choices = paragraphDictionary.keys()

   theParagraph=random.choice(choices)

   for line in paragraphDictionary[theParagraph]
         print line + '\n'

Answer 2

您可以使用str.split()方法以双\\n （换行符）分割每个段落项目，然后随机选择一个。

一条线：

paragraph = random.choice(open('file.txt').read().split('\n\n'))