python - PyParsing 书目引用

Question

我在使用 PyParsing 时遇到了一些问题。我需要从简历中解析一些书目信息。一个例子:

AuthorA, B., AuthorB, M. R., AuthorC, V., and B. LastAuthor. Some sciency title. Name of the confernce, City, State, December 3, 2012

我想出了一些代码来解析(主要)作者列表和日期……其他信息对我来说不是特别重要。

from pyparsing import (Word, Literal, OneOrMore, alphanums, delimitedList, printables, 
    alphas, nums)

family_name = Word(alphanums+'-')
first_init = Word(alphanums+'.')
author = (family_name("LastName") + Literal(',').suppress() + 
          OneOrMore(first_init("FirstInitials") ) )
last_author = first_init("FirstInitials") + family_name("LastName")

author_list = delimitedList(author) + Literal('and').suppress() + last_author

sentence = OneOrMore(Word(printables))
location = delimitedList(Word(printables))
date = Word(alphas) + Word(nums) + Literal(',').suppress() + Word(nums)

citation = (author_list('AuthorLst') + sentence('Title') + location('Location') 
            + date('Date'))

citation.parseString(ntext)

但是，它对“和”放屁，作为作者列表和最后一位作者之间的区别。

我收到错误信息:

---------------------------------------------------------------------------
ParseException                            Traceback (most recent call last)
<ipython-input-142-5d7946dcb775> in <module>()
     15 
     16 
---> 17 citation.parseString(ntext)

/Users/willdampier/anaconda/lib/python2.7/site-packages/pyparsing.pyc in parseString(self, instring, parseAll)
   1123             else:
   1124                 # catch and re-raise exception from here, clears out pyparsing internal stack trace
-> 1125                 raise exc
   1126         else:
   1127             return tokens

ParseException: Expected "and" (at char 40), (line:1, col:41)

有什么建议吗？

Answer 1

定义author后，添加这一行:

author.setName("author").setDebug()

跟踪 author 表达式的匹配。然后为了获得更好的整体诊断，将您的测试线更改为:

author_list.runTests(ntext)

通过这些更改，您将获得如下输出:

Match author at loc 0(1,1)
Matched author -> ['AuthorA', 'B.']
Match author at loc 12(1,13)
Matched author -> ['AuthorB', 'M.', 'R.']
Match author at loc 28(1,29)
Matched author -> ['AuthorC', 'V.']
Match author at loc 41(1,42)
Exception raised:Expected "," (at char 46), (line:1, col:47)

AuthorA, B., AuthorB, M. R., AuthorC, V., and B. LastAuthor. Some sciency title. Name of the confernce, City, State, December 3, 2012
                                        ^
FAIL: Expected "and" (at char 40), (line:1, col:41)

所以您的直接问题是您没有处理“and”之前的尾随“,”。您还需要添加尾随“。”到您对 author_list 的定义。

但是从那里开始，您的 sentence 解析器将出现问题，因为它将处理整个字符串的其余部分。由于您的主要兴趣是获取日期，因此这可能适合您:

stuff = OneOrMore(Word(printables), stopOn=date)
citation = (author_list('AuthorLst') + stuff('body') + date('Date'))

最后，关于您对结果名称的使用(“FirstInitials”、“LastName”等)。好样的，这是一个功能我对 pyparsing 特别满意。但是你需要对每个作者引用文献中的名字进行一些隔离，否则你只会得到最后一位作者的名字。为此，将每个作者包装在一个 pyparsing 组中:

author = Group(family_name("LastName") + Literal(',').suppress() + 
          OneOrMore(first_init("FirstInitials") ) )
last_author = Group(first_init("FirstInitials") + family_name("LastName"))

现在您的 author_list 应该给您一个子结构列表。如果这样做，您可以看到它们:

print(citation.parseString(ntext).dump())

通过我的更改，我得到了您的示例文本:

[['AuthorA', 'B.'], ['AuthorB', 'M.', 'R.'], ['AuthorC', 'V.'], ',', 
 ['B.', 'LastAuthor'], '.', 'Some', 'sciency', 'title.', 'Name', 'of', 
 'the', 'confernce,', 'City,', 'State,', 'December', '3', '2012']
- AuthorLst: [['AuthorA', 'B.'], ['AuthorB', 'M.', 'R.'], 
              ['AuthorC', 'V.'], ',', ['B.', 'LastAuthor'], '.']
  [0]:
    ['AuthorA', 'B.']
    - FirstInitials: 'B.'
    - LastName: 'AuthorA'
  [1]:
    ['AuthorB', 'M.', 'R.']
    - FirstInitials: 'R.'
    - LastName: 'AuthorB'
  [2]:
    ['AuthorC', 'V.']
    - FirstInitials: 'V.'
    - LastName: 'AuthorC'
  [3]:
    ,
  [4]:
    ['B.', 'LastAuthor']
    - FirstInitials: 'B.'
    - LastName: 'LastAuthor'
  [5]:
    .

仍然需要抑制 ',' 和 '.'标点符号，但这只是清理。然后你就可以轻松地遍历您的作者列表并获取每位作者的姓名。