Python - 使用 HTML 包围给定列表中存在于另一个字符串中的字符串实例

Question

我编写了一个函数，它用具有给定属性的 HTML 元素包围搜索词。这个想法是将生成的包围字符串稍后写入日志文件，并突出显示搜索词。

def inject_html(needle, haystack, html_element="span", html_attrs={"class":"matched"}):
    # Find all occurrences of a given string in some text
    # Surround the occurrences with a HTML element and given HTML attributes
    new_str = haystack
    start_index = 0
    while True:
        try:
            # Get the bounds
            start = new_str.lower().index(needle.lower(), start_index)
            end = start + len(needle)

            # Needle is present, compose the HTML to inject
            html_open = "<" + html_element + " " + " ".join(["%s=\"%s\""%(k,html_attrs[k]) for k in html_attrs]) + ">"
            html_close = "</" + html_element + ">"

            new_str = new_str[0:start] + html_open + new_str[start:end] + html_close + new_str[end:len(new_str)]
            start_index = end + len(html_close) + len(html_open)

        except ValueError as ex:
            # String doesn't occur in text after index, break loop
            break
    return new_str

我想打开它以接受一系列针，在大海捞针中用 HTML 定位和包围它们。我可以通过用另一个循环围绕代码来轻松地做到这一点，该循环遍历针、定位和周围的搜索词实例。问题是，这并不能防止意外包围先前注入(inject)的 HTML 代码，例如

def inject_html(needles, haystack, html_element="span", html_attrs={"class":"matched"}):
    # Find all occurrences of a given string in some text
    # Surround the occurrences with a HTML element and given HTML attributes
    new_str = haystack
    for needle in needles:
        start_index = 0
        while True:
        try:
            # Get the bounds
            start = new_str.lower().index(needle.lower(), start_index)
            end = start + len(needle)

            # Needle is present, compose the HTML to inject
            html_open = "<" + html_element + " " + " ".join(["%s=\"%s\""%(k,html_attrs[k]) for k in html_attrs]) + ">"
            html_close = "</" + html_element + ">"

            new_str = new_str[0:start] + html_open + new_str[start:end] + html_close + new_str[end:len(new_str)]
            start_index = end + len(html_close) + len(html_open)

        except ValueError as ex:
            # String doesn't occur in text after index, break loop
            break
    return new_str

search_strings = ["foo", "pan", "test"]
haystack = "Foobar"
print(inject_html(search_strings,haystack))

<s<span class="matched">pan</span> class="matched">Foo</span>bar

在第二次迭代中，代码搜索并包围上一次迭代中插入的“span”中的“pan”文本。

你会如何建议我更改我的原始功能以查找针列表，而不会有将 HTML 注入(inject)不需要的位置(例如在现有标签内)的风险。

---更新---

我通过维护“免疫”范围列表(已经被 HTML 包围因此不需要再次检查的范围)来解决这个问题。

def inject_html(needles, haystack, html_element="span", html_attrs={"class":"matched"}):
    # Find all occurrences of a given string in some text
    # Surround the occurrences with a HTML element and given HTML attributes
    immune = []
    new_str = haystack
    for needle in needles:
        next_index = 0
        while True:
            try:
                # Get the bounds
                start = new_str.lower().index(needle.lower(), next_index)
                end = start + len(needle)

                if not any([(x[0] > start and x[0] < end) or (x[1] > start and x[1] < end) for x in immune]):
                    # Needle is present, compose the HTML to inject
                    html_open = "<" + html_element + " " + " ".join(["%s=\"%s\""%(k,html_attrs[k]) for k in html_attrs]) + ">"
                    html_close = "</" + html_element + ">"

                    new_str = new_str[0:start] + html_open + new_str[start:end] + html_close + new_str[end:len(new_str)]
                    next_index = end + len(html_close) + len(html_open)

                    # Add the highlighted range (and HTML code) to the list of immune ranges
                    immune.append([start, next_index])

            except ValueError as ex:
                # String doesn't occur in text after index, break loop
                break

    return new_str

虽然它不是特别 Pythonic，但我很想看看是否有人能想出更干净的东西。

Answer 1

我会使用这样的东西:

def inject_html(phrases, text_body, html_element_name="span", html_attrs={"class":"matched"}):

    new_text_body = []

    html_start_tag = "<" + html_element_name + " ".join(["%s=\"%s\""%(k,html_attrs[k]) for k in html_attrs]) + ">"
    html_end_tag = "</" + html_element_name + ">"

    text_body_lines = text_body.split("\n")

    for line in text_body_lines:
        for p in phrases:
            if line.lower() == p.lower():
                line = html_start_tag + p + html_end_tag
                break

        new_text_body.append(line)

    return "\n".join(new_text_body)

它逐行检查并替换每一行(如果该行完全匹配(不区分大小写))。

第二轮:

根据匹配需要 (1) 不区分大小写和 (2) 在每一行匹配多个单词/短语的要求，我会使用:

import re

def inject_html(phrases, text_body, html_element_name="span", html_attrs={"class": "matched"}):

    html_start_tag = "<" + html_element_name + " " + " ".join(["%s=\"%s\"" % (k, html_attrs[k]) for k in html_attrs]) + ">"
    html_end_tag = "</" + html_element_name + ">"

    for p in phrases:
        text_body = re.sub(r"({})".format(p), r"{}\1{}".format(html_start_tag, html_end_tag), text_body, flags=re.IGNORECASE)

    return text_body

对于每个提供的短语 p，这使用不区分大小写的 re.sub() 替换来替换提供的文本中该短语的所有实例。 (p) 通过正则表达式组匹配短语。 \1 是匹配找到的短语的回填运算符，将其包含在 HTML 标记中。

text = """
Somewhat more than forty years ago, Mr Baillie Fraser published a 
lively and instructive volume under the title _A Winter’s Journey  
(Tatar) from Constantinople to Teheran. Political complications 
had arisen between Russia and Turkey - an old story, of which we are 
witnessing a new version at the present time. The English government 
deemed it urgently necessary to send out instructions to our 
representatives at Constantinople and Teheran.
"""

new = inject_html(["TEHERAN", "Constantinople"], text)

print(new)

> Somewhat more than forty years ago, Mr Baillie Fraser published a lively and instructive volume under the title _A Winter’s Journey (Tatar) from <span class="matched">Constantinople</span> to <span class="matched">Teheran</span>. Political complications had arisen between Russia and Turkey - an old story, of which we are witnessing a new version at the present time. The English government deemed it urgently necessary to send out instructions to our representatives at <span class="matched">Constantinople</span> and <span class="matched">Teheran</span>.