/// <summary>
/// Attempts to parse an element tag at the current location. If the tag is parsed,
/// the parser position is advanced to the end of the tag name and true is returned.
/// Otherwise, false is returned and the current parser position does not change.
/// </summary>
/// <param name="parser"></param>
/// <param name="tag"></param>
private bool ParseTag(TextParser parser, out string tag)
{
tag = null;
int pos = 0;
Debug.Assert(parser.Peek() == HtmlRules.TagStart);
char c = parser.Peek(++pos);
if (c == '!' || c == '?')
{
c = parser.Peek(++pos);
}
if (HtmlRules.IsTagCharacter(c))
{
while (HtmlRules.IsTagCharacter(parser.Peek(++pos)))
{
;
}
// Move past '<'
parser.MoveAhead();
// Extract tag name
int length = pos - 1;
tag = parser.Text.Substring(parser.Index, length);
parser.MoveAhead(length);
return(true);
}
// No tag found at this position
return(false);
}
private bool ParseToClosingTag(string tag, out string?content)
#endif
{
string endTag = $"</{tag}";
int start = Parser.Index;
// Position assumed to just after open tag
Debug.Assert(Parser.Index > 0 && Parser.Peek(-1) == HtmlRules.TagEnd);
while (!Parser.EndOfText)
{
Parser.SkipTo(endTag, StringComparison.OrdinalIgnoreCase);
// Check that we didn't just match the first part of a longer tag
if (!HtmlRules.IsTagCharacter(Parser.Peek(endTag.Length)))
{
content = Parser.Extract(start, Parser.Index);
Parser.Index += endTag.Length;
Parser.SkipTo(HtmlRules.TagEnd);
Parser.Next();
return(true);
}
Parser.Next();
}
content = null;
return(false);
}
private bool ParseTag(out string tag)
#endif
{
tag = null;
int pos = 0;
Debug.Assert(Parser.Peek() == HtmlRules.TagStart);
char c = Parser.Peek(++pos);
if (c == '!' || c == '?')
{
c = Parser.Peek(++pos);
}
if (HtmlRules.IsTagCharacter(c))
{
while (HtmlRules.IsTagCharacter(Parser.Peek(++pos)))
{
;
}
// Move past '<'
Parser.Next();
// Extract tag name
int length = pos - 1;
tag = Parser.Text.Substring(Parser.Index, length);
Parser.Index += length;
return(true);
}
// No tag found at this position
return(false);
}
/// <summary>
/// Parses the attributes of an element tag. When finished, the parser
/// position is at the next non-space character that follows the attributes.
/// </summary>
private HtmlAttributeCollection ParseAttributes(TextParser parser)
{
HtmlAttributeCollection attributes = new HtmlAttributeCollection();
// Parse tag attributes
parser.MovePastWhitespace();
char ch = parser.Peek();
while (HtmlRules.IsAttributeNameCharacter(ch) || HtmlRules.IsQuoteChar(ch))
{
// Parse attribute name
HtmlAttribute attribute = new HtmlAttribute();
if (HtmlRules.IsQuoteChar(ch))
{
attribute.Name = $"\"{parser.ParseQuotedText()}\"";
}
else
{
attribute.Name = parser.ParseWhile(c => HtmlRules.IsAttributeNameCharacter(c));
}
Debug.Assert(attribute.Name.Length > 0);
// Parse attribute value
parser.MovePastWhitespace();
if (parser.Peek() == '=')
{
parser.MoveAhead(); // Skip '='
parser.MovePastWhitespace();
if (HtmlRules.IsQuoteChar(parser.Peek()))
{
// Quoted attribute value
attribute.Value = parser.ParseQuotedText();
}
else
{
// Unquoted attribute value
attribute.Value = parser.ParseWhile(c => HtmlRules.IsAttributeValueCharacter(c));
Debug.Assert(attribute.Value.Length > 0);
}
}
else
{
// Null attribute value indicates no equals sign
attribute.Value = null;
}
// Add attribute to tag
attributes.Add(attribute.Name, attribute);
// Continue
parser.MovePastWhitespace();
ch = parser.Peek();
}
return(attributes);
}
/// <summary>
/// Parses the attributes of an element tag. When finished, the parser
/// position is at the next non-space character that follows the attributes.
/// </summary>
private HtmlAttributeCollection ParseAttributes()
{
HtmlAttributeCollection attributes = new();
// Parse tag attributes
Parser.SkipWhiteSpace();
char ch = Parser.Peek();
while (HtmlRules.IsAttributeNameCharacter(ch) || HtmlRules.IsQuoteChar(ch))
{
// Parse attribute name
HtmlAttribute attribute = new();
if (HtmlRules.IsQuoteChar(ch))
{
attribute.Name = $"\"{Parser.ParseQuotedText()}\"";
}
else
{
attribute.Name = Parser.ParseWhile(HtmlRules.IsAttributeNameCharacter);
}
Debug.Assert(attribute.Name.Length > 0);
// Parse attribute value
Parser.SkipWhiteSpace();
if (Parser.Peek() == '=')
{
Parser.Next(); // Skip '='
Parser.SkipWhiteSpace();
if (HtmlRules.IsQuoteChar(Parser.Peek()))
{
// Quoted attribute value
attribute.Value = Parser.ParseQuotedText();
}
else
{
// Unquoted attribute value
attribute.Value = Parser.ParseWhile(HtmlRules.IsAttributeValueCharacter);
Debug.Assert(attribute.Value.Length > 0);
}
}
else
{
// Null attribute value indicates no equals sign
attribute.Value = null;
}
// Add attribute to tag
attributes.Add(attribute);
// Continue
Parser.SkipWhiteSpace();
ch = Parser.Peek();
}
return(attributes);
}
/// <summary>
/// Moves the parser position to the closing tag for the given tag name.
/// If the closing tag is not found, the parser position is set to the end
/// of the text and false is returned.
/// </summary>
/// <param name="parser">Parser object.</param>
/// <param name="tag">Tag name for which the closing tag is being searched.</param>
/// <param name="content">Returns the content before the closing tag</param>
/// <returns></returns>
private bool ParseToClosingTag(TextParser parser, string tag, out string content)
{
string endTag = $"</{tag}";
int start = parser.Index;
// Position assumed to just after open tag
Debug.Assert(parser.Index > 0 && parser.Peek(-1) == HtmlRules.TagEnd);
while (!parser.EndOfText)
{
parser.MoveTo(endTag, true);
// Check that we didn't just match the first part of a longer tag
if (!HtmlRules.IsTagCharacter(parser.Peek(endTag.Length)))
{
content = parser.Extract(start, parser.Index);
parser.MoveAhead(endTag.Length);
parser.MoveTo(HtmlRules.TagEnd);
parser.MoveAhead();
return(true);
}
parser.MoveAhead();
}
content = null;
return(false);
}
/// <summary>
/// Parses the given HTML string into a number of root nodes.
/// </summary>
/// <param name="html">The HTML text to parse.</param>
public IEnumerable <HtmlNode> ParseChildren(string html)
{
TextParser parser = new TextParser(html);
HtmlElementNode rootNode = new HtmlElementNode("[Root]");
HtmlElementNode parentNode = rootNode;
string tag;
bool selfClosing;
// Loop until end of input
while (!parser.EndOfText)
{
if (parser.Peek() == HtmlRules.TagStart)
{
// Test for CDATA segments, which we store but do not parse. This includes comments.
CDataDefinition definition = HtmlRules.CDataDefinitions.FirstOrDefault(dd => parser.MatchesCurrentPosition(dd.StartText, dd.IgnoreCase));
if (definition != null)
{
parentNode.Children.Add(ParseCDataNode(parser, definition));
continue;
}
// Closing tag
if (parser.Peek(1) == HtmlRules.ForwardSlash)
{
parser.MoveAhead(2);
tag = parser.ParseWhile(c => HtmlRules.IsTagCharacter(c));
if (tag.Length > 0)
{
if (parentNode.TagName.Equals(tag, HtmlRules.TagStringComparison))
{
// Should never have matched parent if the top-level node
Debug.Assert(!parentNode.IsTopLevelNode);
parentNode = parentNode.ParentNode;
}
else
{
// Handle mismatched closing tag
int tagPriority = HtmlRules.GetTagPriority(tag);
while (!parentNode.IsTopLevelNode && tagPriority > HtmlRules.GetTagPriority(parentNode.TagName))
{
parentNode = parentNode.ParentNode;
}
if (parentNode.TagName.Equals(tag, HtmlRules.TagStringComparison))
{
Debug.Assert(!parentNode.IsTopLevelNode);
parentNode = parentNode.ParentNode;
}
}
}
parser.MoveTo(HtmlRules.TagEnd);
parser.MoveAhead();
continue;
}
// Open tag
if (ParseTag(parser, out tag))
{
HtmlTagFlag flags = HtmlRules.GetTagFlags(tag);
if (flags.HasFlag(HtmlTagFlag.HtmlHeader))
{
parentNode.Children.Add(ParseHtmlHeader(parser));
}
else if (flags.HasFlag(HtmlTagFlag.XmlHeader))
{
parentNode.Children.Add(ParseXmlHeader(parser));
}
else
{
// Parse attributes
HtmlAttributeCollection attributes = ParseAttributes(parser);
// Parse rest of tag
if (parser.Peek() == HtmlRules.ForwardSlash)
{
parser.MoveAhead();
parser.MovePastWhitespace();
selfClosing = true;
}
else
{
selfClosing = false;
}
parser.MoveTo(HtmlRules.TagEnd);
parser.MoveAhead();
// Add node
HtmlElementNode node = new HtmlElementNode(tag, attributes);
while (!HtmlRules.TagMayContain(parentNode.TagName, tag) && !parentNode.IsTopLevelNode)
{
Debug.Assert(parentNode.ParentNode != null);
parentNode = parentNode.ParentNode;
}
parentNode.Children.Add(node);
if (flags.HasFlag(HtmlTagFlag.CData))
{
// CDATA tags are treated as elements but we store and do not parse the inner content
if (!selfClosing)
{
if (ParseToClosingTag(parser, tag, out string content) && content.Length > 0)
{
node.Children.Add(new HtmlCDataNode(string.Empty, string.Empty, content));
}
}
}
else
{
if (selfClosing && flags.HasFlag(HtmlTagFlag.NoSelfClosing))
{
selfClosing = false;
}
if (!selfClosing && !flags.HasFlag(HtmlTagFlag.NoChildren))
{
parentNode = node; // Node becomes new parent
}
}
}
continue;
}
}
// Text node
int start = parser.Index;
// Text must be at least 1 character (handle '<' that is not part of a tag)
parser.MoveAhead();
parser.MoveTo(HtmlRules.TagStart);
Debug.Assert(parser.Index > start);
parentNode.Children.Add(new HtmlTextNode(parser.Extract(start, parser.Index)));
}
//
return(rootNode.Children);
}
/// <summary>
/// Parses the given selector text and returns the corresponding data structures.
/// </summary>
/// <param name="selectorText">The selector text to be parsed.</param>
/// <remarks>
/// Returns multiple <see cref="Selector"/>s when the selector contains commas.
/// </remarks>
/// <returns>The parsed selector data structures.</returns>
public static SelectorCollection ParseSelector(string selectorText)
{
SelectorCollection selectors = new SelectorCollection();
if (!string.IsNullOrWhiteSpace(selectorText))
{
TextParser parser = new TextParser(selectorText);
parser.MovePastWhitespace();
while (!parser.EndOfText)
{
// Test next character
char ch = parser.Peek();
if (IsNameCharacter(ch) || ch == '*')
{
// Parse tag name
Selector selector = selectors.GetLast(true);
if (ch == '*')
{
selector.Tag = null; // Match all tags
}
else
{
selector.Tag = parser.ParseWhile(c => IsNameCharacter(c));
}
}
else if (SpecialCharacters.TryGetValue(ch, out string name))
{
// Parse special attributes
parser.MoveAhead();
string value = parser.ParseWhile(c => IsValueCharacter(c));
if (value.Length > 0)
{
SelectorAttribute attribute = new SelectorAttribute
{
Name = name,
Value = value,
Mode = SelectorAttributeMode.Contains
};
Selector selector = selectors.GetLast(true);
selector.Attributes.Add(attribute);
}
}
else if (ch == '[')
{
// Parse attribute selector
parser.MoveAhead();
parser.MovePastWhitespace();
name = parser.ParseWhile(c => IsNameCharacter(c));
if (name.Length > 0)
{
SelectorAttribute attribute = new SelectorAttribute
{
Name = name
};
// Parse attribute assignment operator
parser.MovePastWhitespace();
if (parser.Peek() == '=')
{
attribute.Mode = SelectorAttributeMode.Match;
parser.MoveAhead();
}
else if (parser.Peek() == ':' && parser.Peek(1) == '=')
{
attribute.Mode = SelectorAttributeMode.RegEx;
parser.MoveAhead(2);
}
else
{
attribute.Mode = SelectorAttributeMode.ExistsOnly;
}
// Parse attribute value
if (attribute.Mode != SelectorAttributeMode.ExistsOnly)
{
parser.MovePastWhitespace();
if (HtmlRules.IsQuoteChar(parser.Peek()))
{
attribute.Value = parser.ParseQuotedText();
}
else
{
attribute.Value = parser.ParseWhile(c => IsValueCharacter(c));
}
}
Selector selector = selectors.GetLast(true);
selector.Attributes.Add(attribute);
}
// Close out attribute selector
parser.MovePastWhitespace();
Debug.Assert(parser.Peek() == ']');
if (parser.Peek() == ']')
{
parser.MoveAhead();
}
}
else if (ch == ',')
{
// Multiple selectors
parser.MoveAhead();
parser.MovePastWhitespace();
selectors.Add(new Selector());
}
else if (ch == '>')
{
// Whitespace indicates child selector
parser.MoveAhead();
parser.MovePastWhitespace();
Debug.Assert(selectors.Any());
Selector selector = selectors.AddChildSelector();
selector.ImmediateChildOnly = true;
}
else if (char.IsWhiteSpace(ch))
{
// Handle whitespace
parser.MovePastWhitespace();
// ',' and '>' change meaning of whitespace
if (parser.Peek() != ',' && parser.Peek() != '>')
{
selectors.AddChildSelector();
}
}
else
{
// Unknown syntax
Debug.Assert(false);
parser.MoveAhead();
}
}
}
selectors.RemoveEmpty();
return(selectors);
}
/// <summary>
/// Parses the given HTML string into a collection of root nodes and their
/// children.
/// </summary>
/// <param name="html">The HTML text to parse.</param>
public IEnumerable <HtmlNode> ParseChildren(string?html, bool ignoreHtmlRules = false)
{
HtmlElementNode rootNode = new("[TempContainer]");
HtmlElementNode parentNode = rootNode;
Parser.Reset(html);
bool selfClosing;
string?tag;
// Loop until end of input
while (!Parser.EndOfText)
{
if (Parser.Peek() == HtmlRules.TagStart)
{
// CDATA segments (blocks we store but don't parse--includes comments)
CDataDefinition?definition = HtmlRules.CDataDefinitions.FirstOrDefault(dd => Parser.MatchesCurrentPosition(dd.StartText, dd.StartComparison));
if (definition != null)
{
parentNode.Children.Add(ParseCDataNode(definition));
continue;
}
// Closing tag
if (Parser.Peek(1) == HtmlRules.ForwardSlash)
{
Parser.Index += 2;
tag = Parser.ParseWhile(HtmlRules.IsTagCharacter);
if (tag.Length > 0)
{
if (parentNode.TagName.Equals(tag, HtmlRules.TagStringComparison))
{
// Should never have matched parent if the top-level node
if (!parentNode.IsTopLevelNode)
{
parentNode = parentNode.ParentNode;
}
}
else
{
// Handle mismatched closing tag
int tagPriority = HtmlRules.GetTagNestLevel(tag);
while (!parentNode.IsTopLevelNode && tagPriority > HtmlRules.GetTagNestLevel(parentNode.TagName))
{
parentNode = parentNode.ParentNode;
}
if (parentNode.TagName.Equals(tag, HtmlRules.TagStringComparison))
{
if (!parentNode.IsTopLevelNode)
{
parentNode = parentNode.ParentNode;
}
}
}
}
Parser.SkipTo(HtmlRules.TagEnd);
Parser.Next();
continue;
}
// Open tag
if (ParseTag(out tag))
{
HtmlTagFlag flags = ignoreHtmlRules ? HtmlTagFlag.None : HtmlRules.GetTagFlags(tag);
if (flags.HasFlag(HtmlTagFlag.HtmlHeader))
{
parentNode.Children.Add(ParseHtmlHeader());
}
else if (flags.HasFlag(HtmlTagFlag.XmlHeader))
{
parentNode.Children.Add(ParseXmlHeader());
}
else
{
// Parse attributes
HtmlAttributeCollection attributes = ParseAttributes();
// Parse rest of tag
if (Parser.Peek() == HtmlRules.ForwardSlash)
{
Parser.Next();
Parser.SkipWhiteSpace();
selfClosing = true;
}
else
{
selfClosing = false;
}
Parser.SkipTo(HtmlRules.TagEnd);
Parser.Next();
// Add node
HtmlElementNode node = new(tag, attributes);
while (!HtmlRules.TagMayContain(parentNode.TagName, tag) && !parentNode.IsTopLevelNode)
{
parentNode = parentNode.ParentNode;
}
parentNode.Children.Add(node);
if (flags.HasFlag(HtmlTagFlag.CData))
{
// CDATA tags are treated as elements but we store and do not parse the inner content
if (!selfClosing)
{
if (ParseToClosingTag(tag, out string?content) && content.Length > 0)
{
node.Children.Add(new HtmlCDataNode(string.Empty, string.Empty, content));
}
}
}
else
{
if (selfClosing && flags.HasFlag(HtmlTagFlag.NoSelfClosing))
{
selfClosing = false;
}
if (!selfClosing && !flags.HasFlag(HtmlTagFlag.NoChildren))
{
parentNode = node; // Node becomes new parent
}
}
}
continue;
}
}
// Text node: must be at least 1 character (handles '<' that was not a tag)
string text = Parser.ParseCharacter();
text += Parser.ParseTo(HtmlRules.TagStart);
parentNode.Children.Add(new HtmlTextNode(text));
}
// Return top-level nodes from nodes just parsed
return(rootNode.Children);
}