python - 当 PyPDF2 正在解析的 PDF 损坏时，我可以让 PyPDF2 优雅地失败吗？

Question

我有一个Python应用程序，可以从公共(public)网站上抓取数百个PDF文件，并使用python库PyPDF2来解析它们。

在成功解析的数百个此类文件中，有一个文件让我感到胃痛。它有 18 页长。文件名是“bad.pdf”。可以看到here .

这是我的代码，它将解析文档:

$ virtualenv my_env
$ source my_env/bin/activate
(my_env) $ pip install PyPDF2==1.26.0
(my_env) $ python
>>> import PyPDF2
>>> def parse_pdf_doc():
>>>     pdfFileObj = open('bad.pdf', 'rb')
>>>     pdfReader = PyPDF2.PdfFileReader(pdfFileObj)
>>>     for curr_page_num in range(pdfReader.numPages):
>>>         print 'curr_page_num = {}'.format(curr_page_num)
>>>         pageObj = pdfReader.getPage(curr_page_num)
>>>         print '\tPage Retrieved successfully'
>>>         page_text = pageObj.extractText()
>>>         print '\tText extracted successfully'

当我运行此代码时，它成功解析了前九页。但到了第十页，它就挂了。永远:

>>> parse_pdf_doc()
curr_page_num = 0
    Page Retrieved successfully
    Text extracted successfully
curr_page_num = 1
    Page Retrieved successfully
    Text extracted successfully
curr_page_num = 2
    Page Retrieved successfully
    Text extracted successfully
curr_page_num = 3
    Page Retrieved successfully
    Text extracted successfully
curr_page_num = 4
    Page Retrieved successfully
    Text extracted successfully
curr_page_num = 5
    Page Retrieved successfully
    Text extracted successfully
curr_page_num = 6
    Page Retrieved successfully
    Text extracted successfully
curr_page_num = 7
    Page Retrieved successfully
    Text extracted successfully
curr_page_num = 8
    Page Retrieved successfully
    Text extracted successfully
curr_page_num = 9
    Page Retrieved successfully
<... hung here forever ...>

第 10 页有什么问题？让我们在查看器中打开它。哇哦:即使是 Google 文档也无法解析第 10 页。所以该页面肯定有一些损坏的地方:

但是，我仍然需要 PyPDF 抛出异常或以其他方式失败，而不仅仅是进入无限循环。它破坏了我的工作流程。我该如何解决 PDF 文件中这个损坏的页面？

Answer 1

下面的模板将让您了解如何实现这一目标。

from multiprocessing import Process
pdfFileObj = open('bad.pdf', 'rb')
for page in PDFPage.get_pages(pdfFileObj):
                    processTimeout = 20
                    extractTextProcess = Process(target=Function_to_extract_text, args=(pdfObject,page)

还可以通过with关键字打开文件(以防止内存泄漏)