一种提议的解决方案
哇,这真是我个人的一个“诅咒”!显然对于一个问题,我不能不花费包括但不限于非常不合理的时间来解决它。
我考虑过这个。我想到了 HTML Tidy,也许它会起作用,但我很难理解它。
所以,我写了自己的解决方案。
我测试了你的输入和我自己编写的其他输入。它似乎表现得相当出色。当然,其中可能存在漏洞,但它可能为您提供一个起点。
无论如何,我的方法如下:
- 使用一个类将 HTML 文档中的单个单词的概念封装起来,包括该单词在给定“顶部”的 HTML 文档层次结构中的位置信息。我已经在下面的
HtmlWord
类中实现了这一点。 - 创建一个类,能够编写由上述 HTML 单词组成的单行,使得开始元素和结束元素标记添加到适当的位置。我已经在下面的
HtmlLine
类中实现了这一点。 - 编写一些扩展方法,使这些类可以直接从
HtmlAgilityPack.HtmlNode
对象中立即和直观地访问。我已经在下面的 HtmlHelper
类中实现了这一点。
我为做这一切而疯狂吗?可能是,是的。但是,你知道,如果你找不到任何其他的方法,你可以试试这个。
以下是它如何处理你的示例输入:
var document = new HtmlDocument();
document.LoadHtml("<p><strong>Lorem ipsum dolor sit amet, <em>consectetur adipiscing</em></strong> elit.</p>");
var nodeToSplit = document.DocumentNode.SelectSingleNode("p");
var lines = nodeToSplit.SplitIntoLines(3);
foreach (var line in lines)
Console.WriteLine(line.ToString());
输出:
<p><strong>Lorem ipsum dolor </strong></p>
<p><strong>sit amet, <em>consectetur </em></strong></p>
<p><strong><em>adipiscing </em></strong>elit. </p>
现在开始介绍代码:
HtmlWord类
using System;
using System.Collections.Generic;
using System.Linq;
using HtmlAgilityPack;
public class HtmlWord {
public string Text { get; private set; }
public HtmlNode[] NodeStack { get; private set; }
public string NodeList {
get { return string.Join(", ", NodeStack.Select(n => n.Name).ToArray()); }
}
internal HtmlWord(string text, HtmlNode node, HtmlNode top) {
Text = text;
NodeStack = GetNodeStack(node, top);
}
private static HtmlNode[] GetNodeStack(HtmlNode node, HtmlNode top) {
var nodes = new Stack<HtmlNode>();
while (node != null && !node.Equals(top)) {
nodes.Push(node);
node = node.ParentNode;
};
return nodes.ToArray();
}
}
HtmlLine类
using System;
using System.Collections.Generic;
using System.IO;
using System.Linq;
using System.Text;
using System.Xml;
using HtmlAgilityPack;
[Flags()]
public enum NodeChange {
None = 0,
Dropped = 1,
Added = 2
}
public class HtmlLine {
private List<HtmlWord> _words;
public IList<HtmlWord> Words {
get { return _words.AsReadOnly(); }
}
public int WordCount {
get { return _words.Count; }
}
public HtmlLine(IEnumerable<HtmlWord> words) {
_words = new List<HtmlWord>(words);
}
private static NodeChange CompareNodeStacks(HtmlWord x, HtmlWord y, out HtmlNode[] droppedNodes, out HtmlNode[] addedNodes) {
var droppedList = new List<HtmlNode>();
var addedList = new List<HtmlNode>();
foreach (var node in x.NodeStack.Reverse()) {
if (!Array.Exists(y.NodeStack, n => n.Equals(node)))
droppedList.Add(node);
}
foreach (var node in y.NodeStack) {
if (!Array.Exists(x.NodeStack, n => n.Equals(node)))
addedList.Add(node);
}
droppedNodes = droppedList.ToArray();
addedNodes = addedList.ToArray();
NodeChange change = NodeChange.None;
if (droppedNodes.Length > 0)
change &= NodeChange.Dropped;
if (addedNodes.Length > 0)
change &= NodeChange.Added;
return change;
}
public override string ToString() {
if (WordCount < 1)
return string.Empty;
var lineBuilder = new StringBuilder();
using (var lineWriter = new StringWriter(lineBuilder))
using (var xmlWriter = new XmlTextWriter(lineWriter)) {
var firstWord = _words[0];
foreach (var node in firstWord.NodeStack) {
xmlWriter.WriteStartElement(node.Name);
foreach (var attr in node.Attributes)
xmlWriter.WriteAttributeString(attr.Name, attr.Value);
}
xmlWriter.WriteString(firstWord.Text + " ");
for (int i = 1; i < WordCount; ++i) {
var previousWord = _words[i - 1];
var word = _words[i];
HtmlNode[] droppedNodes;
HtmlNode[] addedNodes;
CompareNodeStacks(
previousWord,
word,
out droppedNodes,
out addedNodes
);
foreach (var dropped in droppedNodes)
xmlWriter.WriteEndElement();
foreach (var added in addedNodes) {
xmlWriter.WriteStartElement(added.Name);
foreach (var attr in added.Attributes)
xmlWriter.WriteAttributeString(attr.Name, attr.Value);
}
xmlWriter.WriteString(word.Text + " ");
if (i == _words.Count - 1) {
foreach (var node in word.NodeStack)
xmlWriter.WriteEndElement();
}
}
}
return lineBuilder.ToString();
}
}
HtmlHelper静态类
using System;
using System.Collections.Generic;
using System.Linq;
using HtmlAgilityPack;
public static class HtmlHelper {
public static IList<HtmlLine> SplitIntoLines(this HtmlNode node, int wordsPerLine) {
var lines = new List<HtmlLine>();
var words = node.GetWords(node.ParentNode);
for (int i = 0; i < words.Count; i += wordsPerLine) {
lines.Add(new HtmlLine(words.Skip(i).Take(wordsPerLine)));
}
return lines.AsReadOnly();
}
public static IList<HtmlWord> GetWords(this HtmlNode node, HtmlNode top) {
var words = new List<HtmlWord>();
if (node.HasChildNodes) {
foreach (var child in node.ChildNodes)
words.AddRange(child.GetWords(top));
} else {
var textNode = node as HtmlTextNode;
if (textNode != null && !string.IsNullOrEmpty(textNode.Text)) {
string[] singleWords = textNode.Text.Split(
new string[] {" "},
StringSplitOptions.RemoveEmptyEntries
);
words.AddRange(
singleWords
.Select(w => new HtmlWord(w, node.ParentNode, top)
)
);
}
}
return words.AsReadOnly();
}
}
结论
再次强调:这只是一个临时解决方案;我相信它存在问题。我提供它只是为了让您考虑 -- 如果您无法通过其他方式获得所需的行为。