Django和python-docx

Question

我试图通过python-docx和django创建一个docx文件。

使用python直接运行示例脚本(example-makedocument.py)。一份“工作副本”。它按照预期生成示例docx文件。

然后，我尝试将example-makedocument.py代码移动到Django视图中。这篇文章底部的代码是该视图的代码。你会看到我在前面做了一些小的修改，但总体上没有改变。

在运行此视图时，我会发现无法找到图像文件的错误。

[Errno 2] No such file or directory: 'image1.png'

我有：

• 确保文件夹/文件结构与工作副本相同(它需要其子文件夹才能正常工作)
• 我尝试过将docx.py template_dir硬编码到保存文件夹/文件的位置(我检查了这些结果是否指向与工作副本相同的最终路径)。
• 我已经从示例中删除了图像，但是它导致了错误local variable 'contenttypes' referenced before assignment。

我认为这与道路有关，但我不知道为什么/如何？有人能帮忙吗？

完整的python代码可用这里。

我的看法是：

from myapp.libs.docx.docx import *
# ** commented out below line, replaced with above **
#from docx import *

def export_docx(request):

# ** commented out below line **
#if __name__ == '__main__':
    # Default set of relationshipships - the minimum components of a document
    relationships = relationshiplist()

    # Make a new document tree - this is the main part of a Word document
    document = newdocument()

    # This xpath location is where most interesting content lives
    body = document.xpath('/w:document/w:body', namespaces=nsprefixes)[0]

    # Append two headings and a paragraph
    body.append(heading("Welcome to Python's docx module", 1))
    body.append(heading('Make and edit docx in 200 lines of pure Python', 2))
    body.append(paragraph('The module was created when I was looking for a '
        'Python support for MS Word .doc files on PyPI and Stackoverflow. '
        'Unfortunately, the only solutions I could find used:'))

    # Add a numbered list
    points = [ 'COM automation'
             , '.net or Java'
             , 'Automating OpenOffice or MS Office'
             ]
    for point in points:
        body.append(paragraph(point, style='ListNumber'))
    body.append(paragraph('For those of us who prefer something simpler, I '
                          'made docx.'))
    body.append(heading('Making documents', 2))
    body.append(paragraph('The docx module has the following features:'))

    # Add some bullets
    points = ['Paragraphs', 'Bullets', 'Numbered lists',
              'Multiple levels of headings', 'Tables', 'Document Properties']
    for point in points:
        body.append(paragraph(point, style='ListBullet'))

    body.append(paragraph('Tables are just lists of lists, like this:'))
    # Append a table
    tbl_rows = [ ['A1', 'A2', 'A3']
               , ['B1', 'B2', 'B3']
               , ['C1', 'C2', 'C3']
               ]
    body.append(table(tbl_rows))

    body.append(heading('Editing documents', 2))
    body.append(paragraph('Thanks to the awesomeness of the lxml module, '
                          'we can:'))
    points = [ 'Search and replace'
             , 'Extract plain text of document'
             , 'Add and delete items anywhere within the document'
             ]
    for point in points:
        body.append(paragraph(point, style='ListBullet'))

    # Add an image
    relationships, picpara = picture(relationships, 'image1.png',
                                     'This is a test description')
    body.append(picpara)

    # Search and replace
    print 'Searching for something in a paragraph ...',
    if search(body, 'the awesomeness'):
        print 'found it!'
    else:
        print 'nope.'

    print 'Searching for something in a heading ...',
    if search(body, '200 lines'):
        print 'found it!'
    else:
        print 'nope.'

    print 'Replacing ...',
    body = replace(body, 'the awesomeness', 'the goshdarned awesomeness')
    print 'done.'

    # Add a pagebreak
    body.append(pagebreak(type='page', orient='portrait'))

    body.append(heading('Ideas? Questions? Want to contribute?', 2))
    body.append(paragraph('Email <python.docx@librelist.com>'))

    # Create our properties, contenttypes, and other support files
    title    = 'Python docx demo'
    subject  = 'A practical example of making docx from Python'
    creator  = 'Mike MacCana'
    keywords = ['python', 'Office Open XML', 'Word']

    coreprops = coreproperties(title=title, subject=subject, creator=creator,
                               keywords=keywords)
    appprops = appproperties()
    contenttypes = contenttypes()
    websettings = websettings()
    wordrelationships = wordrelationships(relationships)

    # Save our document
    savedocx(document, coreprops, appprops, contenttypes, websettings,
             wordrelationships, 'Welcome to the Python docx module.docx')

========

已更新。下面提供了我找到的解决方案的细节。

docx.py代码假设路径与实际路径不同。这导致image1.png文件“丢失”。完成下面的步骤为我解决了这个问题。

1. 编辑了docx.py文件并添加了一个变量： docx_dir =‘/path/to/docx/文件夹’
2. 编辑下面的行(docx.py)以使用上述变量(大约显示行号，将与原始源略有不同)

1. shutil.copyfile(join(docx\_dir, picname), join(media\_dir, picname))
2. pixelwidth, pixelheight = Image.open(join(docx\_dir, picname)).size[0:2]

​

1. 以上所述消除了路径问题。但是现在我遇到了错误local variable 'contenttypes' referenced before assignment。我发现这个错误不是通过与路径有关的方法解决的，而是通过重复的名称解决的(我不完全理解为什么仍然如此)。

注意:下面的代码位于我的视图or..if中，与example-makedocument.py中的原始源代码its相比较

在此之前：

contenttypes = contenttypes()
websettings = websettings()
wordrelationships = wordrelationships(relationships)

之后：

contentt = contenttypes()
webs = websettings()
wordr = wordrelationships(relationships)

# Save our document
savedocx(document, coreprops, appprops, contentt, webs,
         wordr, 'Welcome to the Python docx module.docx')

Answer 1

image1.png必须位于正在运行的Python程序的当前目录中，或者作为相对于正在运行的程序的当前目录的路径提供，或者更少的可移植性，但可能更容易工作，但它可以作为绝对路径提供。

特别是，我认为您可能会发现在这篇文章中给出的答案最有帮助：使用WSGI和Apache运行Python的工作目录 (用于动态地确定image.png的绝对路径)。

Answer 2

自上面编写的代码以来，python模块发生了一些变化。现在，在django中添加一个图像如下所示：

from docx import *
from docx.shared import Inches
document = Document()
document.add_picture((r'%s/static/images/my-header.png' % (settings.PROJECT_PATH)), width=Inches(4))