Python docx查找单词并将其替换为斜体版本

Question

我想了几种方法来实现这一点，但每种方法都比下一种更难看。我正在想办法在word文档中搜索一个单词的所有实例并将其斜体化

我无法上传word文档，但以下是我的想法：

一个工作示例将查找billybob的所有实例，包括表中的实例，并将其斜体化。问题是运行经常对齐的方式意味着一个运行可能有billy，下一个运行可能有bob，因此没有简单的方法可以找到所有运行

Answer 1

我将不讨论这个问题，因为我提出的方法并不完美，但它在绝大多数情况下都有效。代码如下：

document = Document(<YOUR_DOC>)

# Data will be a list of rows represented as dictionaries
# containing each row's data.
characters = {}

for paragraph in <YOUR_PARAGRAPHS>:

    run_string = ""
    run_index = {}
    i = 0

    for x, run in enumerate(paragraph.runs):

        # Create a string consisting of all the runs' text. Theoretically this
        # should always be the same as parapgrah.text, but I didn't check
        run_string = run_string + run.text

        # The index i represents the starting position of the run in question
        # within the string. We are creating a dictionary of form
        # {<run_start_location>: <pointer_to_run>}
        run_index[i] = x

        # This will be the start of the next run
        i = i + len(run.text)

    word_you_wanted_to_find = re.findall("some_regex", paragraph.text)

    for word in word_you_wanted_to_find:

        # [m.start() for m in re.finditer(word, run_string)] returns the starting
        # positions of each word that was found
        for word_start in [m.start() for m in re.finditer(word, run_string)]:
            word_end = word_start + len(word)

            # This will be a list of the indices of the runs which have part
            # of the word we want to include
            included_runs = []

            for key in run_index.keys():

                # Remember, the key is the location in the string of the start of
                # the run. In this case, the start of the word start should be less than
                # the key+len(run) and the end of the word should be greater
                # than the key (the start of the run)
                if word_start <= (key + len(paragraph.runs[run_index[key]].text)) and key < word_end:
                    included_runs.append(key)

                # If the key is larger than or equal to the end of the word,
                # this means we have found all relevant keys. We don't need
                # to loop over the rest (we could, it just wouldn't be efficient)
                if key >= word_end:
                    break

            # At this point, included_runs is a full list of indices to the relevant
            # runs so we can modify each one in turn.
            for run_key in included_runs:
                paragraph.runs[run_index[run_key]].italic = True

document.save(<MODIFIED_DOC>)

问题1

这种方法的问题是，虽然不常见（至少在我的文档中是这样），但一次运行可能包含的不仅仅是目标单词。因此，你可能最终会将整个跑步记录（包括你的跑步记录和一些跑步记录）斜体化。对于我的用例来说，在这里解决这个问题是没有意义的

解决方案

如果要完善我在上面所做的工作，您必须更改此代码块：

if word_start <= (key + len(paragraph.runs[run_index[key]].text)) and key < word_end:
  included_runs.append(key)

在这里，您已经确定了符合您承诺的跑步。您需要扩展代码，将单词分隔到自己的运行中，并将其从当前运行中删除。然后你可以分别用斜体字标出那一段

问题2

上面显示的代码不能同时处理表格和普通文本。对于我的用例，我不需要这样做，但在一般情况下，您必须同时检查这两个选项