- Example 1 - Parsing HTML string into a TDocument
- Example 2 - Get HTML text from a TDocument
- Example 3 - Get a string reprentation of the DOM tree
Today we see more and more email messages formatted as HTML. For me the email is plain text medium (with attachments) and I don't use WebBrowser or Mozilla object as a message browser in my email client. So I'm write quick and dirty HTML parser instead. It seems enough to parse email messages but some work is needed to shift them to more general purpose HTML parser.
TDocument, TNode, TElement, TAttr etc. implement DOM2 Core.
THtmlParser produces TDocument from HTML string.
class function Parse(const HtmlStr: TDomString): TDocument;
THtmlParser uses THtmlReader - an event-driven SAX-like interface that performs the actual processing of the html string.
To convert DOM tree to plain text or HTML use TTextFormatter or THtmlFormatter respectively.
Parser is implemented as several modules:
DomCore.pas- core DOM implementationHmlParser.pas- Parses HTML into a DocumentHtmlReader.pas- lexical analyzer for parsing HTMLEntities.pas- HTML character definitionsHtmlTags.pas- HTML tags attributesFormatter.pas- HTML DOM tree converters
Example 1 — Parse HTML into TDocument object
To convert a string containing HTML, into a DOM TDocument object, use the THTMLParser.Parse:
var
doc: TDocument;
doc := THTMLParser.Parse('<!DOCTYPE html><HTML><BODY>Hello, world!</BODY></HTML>');
Example 2 — Get full HTML text from a TDocument
To convert a TDocument into an HTML string, use THtmlFormatter.GetHTML class in Formatter.pas:
var
html: string;
html := THtmlFormatter.GetHTML(document);
returns:
<HTML>
<BODY>Hello, world!</BODY>
</HTML>
Example 3 — Get a string reprentation of the DOM tree
Use the function Formatter.DumpDOM to get a representation of the DOM tree of a Node:
var
s: string;
doc: TDocument;
//Get a sample html document
s :=
'<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd">'+#13#10+
'<html lang="en">'+#13#10+
' <head>'+#13#10+
' <title>Sample page</title>'+#13#10+
' </head>'+#13#10+
' <body>'+#13#10+
' <h1>Sample page</h1>'+#13#10+
' <p>This is a <a href="demo.html">simple</a> sample.</p>'+#13#10+
' <!-- this is a comment -->'+#13#10+
' </body>'+#13#10+
'</html>';
doc := THtmlParser.Parse(s);
s := DumpDOM(doc);
The returned DOM string will be:
#document
├── DOCTYPE: html
╰── HTML lang="en"
├── HEAD
│ ├── #text: "⏎␣␣"
│ ├── TITLE
│ │ ╰── #text: "Sample␣page"
│ ╰── #text: "⏎␣"
├── #text: "⏎␣"
╰── BODY
├── #text: "⏎␣␣"
├── H1
│ ╰── #text: "Sample␣page"
├── #text: "⏎␣␣"
├── P
│ ├── #text: "This␣is␣a␣"
│ ├── A href="demo.html"
│ │ ╰── #text: "simple"
│ ╰── #text: "␣sample."
├── #text: "⏎␣␣"
├── #comment: this is a comment
╰── #text: "⏎␣⏎"
12/22/2021
- Formatter.pas: Added DumpDOM function, which dumps a text-based tree of the document tree
- TBaseFormatter: Fixed node iteration so starts with the Document child nodes - including doctype
- TBaseFormatter: Added support to dump doctype nodes
- THtmlParser: Added static Parse(html) class function
- THtmlParser: GetMainElement now compares element tag names case insensitively
- THtmlParser: FindThisElement now compares element tag names case insensitively
- THtmlParser: ProcessDocType: DocType nodes are now correctly adds to the DOM tree itself, rather than just being a special attribute of the Document object
- THtmlParser: Removed ProcessNotation, since the event was removed from THtmlReader
- THtmlReader: Fixed ReadDocumentType method to correctly handle: - doctypes without either a publicId or systemID (e.g. HTML 5) - doctypes where only the SystemID is specified
- THtmlReader: Changed ReadDocumentType to LowerCase the doctype name
- THtmlReader: Changed ReadDocumentType so Doctype publicid and systemid can also be enclosed in APOSTROPHE (in addition to QUOTATION MARK)
- THtmlReader: Fixed ReadQuotedValue to not return the closing '"' character. (only used by doctype reading)
- THtmlReader: Removed OnNotation event; since it's never called and nobody seems to know what could possibly be
- THtmlTagList: CompareName now does a case insensitive comparison
- TURLSchemes: Schemes are now populated in the constructor (like THtmlTagList)
- DOMCore:
- TDocumentType: removed Entities, Notations, and InternalSubsets (per HTML5)
- TDocument: Removed DocType as an attribute, and is now correctly added to the DOM tree.
- TNode: Fixed getElementById to still recursively ask child nodes even if Self node is not an Element (e.g. a Document)
- TNode: Removed IsSupported (per HTML5)
- TNamedNodeMap: GetNamedItem now compares node names case insensitively
- TNamedNodeMap: getNamedItemNS now compares node names case insensitively
- TDocument: CanInsert will now accept DOCUMENT_TYPE_NODE as a node that can be inserted
- TDocument: GetDocType is now a convenience method (like GetDocumentElement) that finds the doctype child node for you
- TDocument: CreateAttribute now lowercases the attribute name
- TNode: Added TextContent property
12/20/2021
- Fixed TNode: NodeValue was not calling GetNodeValue getter
- fixed attribute lookups to be case insensitive (like everything else in html)
- fixed tag name normalization to UPPERCASE, the canonical form in the spec (https://www.w3.org/TR/DOM-Level-3-Core/core.html#ID-104682815) (https://dom.spec.whatwg.org/#dom-element-tagname)
- TNode: InsertBefore should only fail the WRONG_DOCUMENT_ERR check if the node actually has an ownerDocument assigned.
- DOMCore: Added TNodeType type
- InsertBefore will no longer reject a node if its OwnerDocument is nil
- TSearchNodeList: AcceptNode is now case insensitive to node local names
- TNamedNodeMap.GetNamedItem and GetNamedItemNS are no longer case sensitive
- TElement: constructor now canonically UPPERCASEs the element tag name
- TDocument: No longer takes a doctype in its constructor. DocType node is just another node added as a child of the Document
- DOMImplementation: Added CreateHtmlDocument; following the spec of what it should contain
- DOMImplementation: CreateDocument doctype parameter is now optional
- DOMImplementation: Changed HasFeature to return dummy True (per HTML5)
- Removed TDocumentType.Entities, Notations, and InternalSubset (per HTML5)
Original Homepage: http://htmlp.sourceforge.net/