.NET Core(Linux)下保留纯净<p>标签并移除其余HTML标签的方案

Got it, let's work through this problem. Your current regex strips all HTML tags entirely, but you need to keep only clean <p></p> tags (no attributes) while removing every other tag like <div style="">, <span style="">, etc. Since you're on Linux with .NET Core and can't use full-framework parsers like AngleSharp, we can adjust your regex approach to fit this exact need.

Step 1: Fix the Regex Target #

First, note that your current regex uses <.*?> which targets HTML-escaped tags (like <div>). If your input contains actual unescaped HTML tags (with < and >), you'll want to target <.*?> instead. Let's build the solution for unescaped HTML first (we'll cover the escaped case too).

Step 2: Two-Stage Regex Replacement #

We'll use two regex passes to get the desired result:

1. Convert any <p> tag with attributes (like <p style="margin:0" class="foo">) to a clean <p>.
2. Strip all remaining HTML tags except the clean <p> and </p>.

Here's the implementation:

using System.Text.RegularExpressions;

public static string StripHtmlPreserveCleanPTags(string input)
{
    // First pass: Replace any <p> with attributes to clean <p>
    // Matches <p followed by any non-> characters, then >
    string cleanPTags = Regex.Replace(input, @"<p\b[^>]*>", "<p>", RegexOptions.IgnoreCase);
    
    // Second pass: Strip all tags except clean <p> and </p>
    // Matches any tag that isn't exactly <p> or </p>
    string result = Regex.Replace(cleanPTags, @"<(?!\/?p\b)[^>]*>", string.Empty, RegexOptions.IgnoreCase);
    
    return result;
}

How It Works #

• First regex: @"<p\b[^>]*>"
  • <p\b matches the start of a <p> tag (the word boundary \b ensures we don't match tags like <pizza>)
  • [^>]* matches any characters except > (all the attributes)
  • Replaces the entire attribute-rich <p> tag with a clean <p>
• Second regex: @"<(?!\/?p\b)[^>]*>"
  • < matches the start of a tag
  • (?!\/?p\b) is a negative lookahead: it skips tags that start with p (either <p> or </p>)
  • [^>]*> matches the rest of the tag, which we replace with an empty string (strip it)

Handling HTML-Escaped Tags #

If your input has escaped tags (like <p style="...">), adjust the regex to target the escaped entities:

public static string StripEscapedHtmlPreserveCleanPTags(string input)
{
    // Clean escaped <p> tags
    string cleanPTags = Regex.Replace(input, @"&lt;p\b[^&]*&gt;", "&lt;p&gt;", RegexOptions.IgnoreCase);
    // Strip all other escaped tags
    string result = Regex.Replace(cleanPTags, @"&lt;(?!\/?p\b)[^&]*&gt;", string.Empty, RegexOptions.IgnoreCase);
    // Optional: Convert escaped <p> back to actual tags if needed
    // result = result.Replace("&lt;p&gt;", "<p>").Replace("&lt;/p&gt;", "</p>");
    return result;
}

Test Example #

Input:

<div style="color:red">Hello <p style="margin:0">World</p> <span>Test</span> <p class="bar">Another paragraph</p></div>

Output after processing:

Hello <p>World</p> Test <p>Another paragraph</p>

Limitations to Note #

Regex isn't a perfect HTML parser, so this works best for:

• Standard content tags (no nested <p> tags, which are invalid HTML anyway)
• Tags that don't contain > characters inside attribute values (e.g., <p data-value=">"> would break the regex)
  If your HTML has edge cases like these, double-check if AngleSharp is actually compatible with .NET Core (it does support .NET Core now!)—it would be more reliable for complex HTML. But if that's a hard restriction, the regex approach will handle most common scenarios.

内容的提问来源于stack exchange，提问作者Patrick