public DocHTML(HTMLDocument doc)
{
this.mDoc = doc;
this.mDoc2 = (IHTMLDocument2)this.mDoc;
this.mDoc3 = (IHTMLDocument3)this.mDoc;
this.mDoc4 = (IHTMLDocument4)this.mDoc;
this.mDoc5 = (IHTMLDocument5)this.mDoc;
}
public DocHTML(HTMLDocument doc)
{
mDoc = doc;
mDoc2 = (IHTMLDocument2)mDoc;
mDoc3 = (IHTMLDocument3)mDoc;
mDoc4 = (IHTMLDocument4)mDoc;
mDoc5 = (IHTMLDocument5)mDoc;
}
public DocHTML(HTMLDocument doc)
{
mDoc = doc;
mDoc2 = (IHTMLDocument2)mDoc;
mDoc3 = (IHTMLDocument3)mDoc;
mDoc4 = (IHTMLDocument4)mDoc;
mDoc5 = (IHTMLDocument5)mDoc;
}
public DocumentAdapter( object document )
{
//_documentUri = documentUri;
_raw = document;
_document = document as IHTMLDocument;
_document2 = document as IHTMLDocument2;
_document3 = document as IHTMLDocument3;
_document4 = document as IHTMLDocument4;
_document5 = document as IHTMLDocument5;
}
public DocumentAdapter(object document)
{
//_documentUri = documentUri;
_raw = document;
_document = document as IHTMLDocument;
_document2 = document as IHTMLDocument2;
_document3 = document as IHTMLDocument3;
_document4 = document as IHTMLDocument4;
_document5 = document as IHTMLDocument5;
}
static void DispatchChangeEvent(WebControl awc, int iIndex, string sControl, IHTMLSelectElement ihie, IHTMLOptionElement ihoe, IHTMLDocument2 oDoc2)
{
ihoe.selected = true;
object dummy = null;
IHTMLDocument4 oDoc4 = (IHTMLDocument4)oDoc2;
object eventObj = oDoc4.CreateEventObject(ref dummy);
IHTMLEventObj2 obj2 = (IHTMLEventObj2)eventObj;
IHTMLSelectElement hsec = ihie as IHTMLSelectElement;
awc.ReportNavState("Before FireEvent");
hsec.FireEvent("onchange", ref eventObj);
awc.ReportNavState("After FireEvent");
}
public void Click()
{
object refObj = null;
IHTMLDocument4 document = node.document as IHTMLDocument4;
IHTMLElement3 element = (IHTMLElement3)node;
IHTMLEventObj eventObject = document.CreateEventObject(ref refObj);
object eventRef = eventObject;
element.FireEvent("onMouseDown", ref eventRef);
element.FireEvent("onMouseUp", ref eventRef);
node.click();
parent.WaitForLoadToComplete();
}
public SpellingHighlighter(ISpellingChecker spellingChecker, IHighlightRenderingServicesRaw highlightRenderingServices,
IDisplayServicesRaw displayServices, IMarkupServicesRaw markupServices, IHTMLDocument4 htmlDocument)
{
_spellingChecker = spellingChecker;
_highlightRenderingServices = highlightRenderingServices;
_displayServices = displayServices;
_markupServicesRaw = markupServices;
_markupServices = new MshtmlMarkupServices(_markupServicesRaw);
_htmlDocument = htmlDocument;
_tracker = new HighlightSegmentTracker();
//the timer to handle interleaving of spell checking
_timer = new SpellingTimer(TIMER_INTERVAL);
_timer.Start();
_timer.Tick += new EventHandler(_timer_Tick);
_workerQueue = new Queue();
}
public SpellingHighlighter(ISpellingChecker spellingChecker, IHighlightRenderingServicesRaw highlightRenderingServices,
IDisplayServicesRaw displayServices, IMarkupServicesRaw markupServices, IHTMLDocument4 htmlDocument)
{
_spellingChecker = spellingChecker;
_highlightRenderingServices = highlightRenderingServices;
_displayServices = displayServices;
_markupServicesRaw = markupServices;
_markupServices = new MshtmlMarkupServices(_markupServicesRaw);
_htmlDocument = htmlDocument;
_tracker = new HighlightSegmentTracker();
//the timer to handle interleaving of spell ghecking
_timer = new SpellingTimer(TIMER_INTERVAL);
_timer.Start();
_timer.Tick += new EventHandler(_timer_Tick);
_workerQueue = new Queue();
}
static void DispatchChangeEventTry2(
WebControl awc,
int iIndex,
string sControl,
IHTMLSelectElement ihie,
IHTMLOptionElement ihoe,
IHTMLDocument2 oDoc2)
{
object dummy = null;
IHTMLDocument4 oDoc4 = (IHTMLDocument4) oDoc2;
object eventObj = oDoc4.CreateEventObject(ref dummy);
IHTMLEventObj2 obj2 = (IHTMLEventObj2) eventObj;
IHTMLElement3 ihe3 = (IHTMLElement3) ihie;
awc.ResetNav();
awc.ReportNavState("Before FireEvent");
// ihe3.FireEvent("onchange", ref eventObj);
HtmlElement head = awc.AxWeb.Document.GetElementsByTagName("head")[0];
HtmlElement scriptEl = awc.AxWeb.Document.CreateElement("script");
IHTMLScriptElement element = (IHTMLScriptElement) scriptEl.DomElement;
// element.text = $"function changeSelect() {{ $'({sControl}').trigger('change'); }}";
element.text = "function triggerOnChange() "
+ "{{ "
+
// "alert('im here'); " +
$"var ctl = document.getElementById('{sControl}'); "
+
// "alert(ctl); "+
"var evt = document.createEvent('HTMLEvents'); "
+ "evt.initEvent('change', false, true); "
+ "ctl.dispatchEvent(evt);"
+ "}} ";
// element.text = $"function triggerOnChange() {{ alert('{sControl}');}}";
head.AppendChild(scriptEl);
// ArbWeb.AwMainForm.DebugModelessWait();
awc.AxWeb.Document.InvokeScript("triggerOnChange");
// ArbWeb.AwMainForm.DebugModelessWait();
awc.ReportNavState("After FireEvent");
awc.WaitForBrowserReady();
awc.WaitDoLog(500);
}
private void Type(string value)
{
object refObj = null;
IHTMLDocument4 document = node.document as IHTMLDocument4;
IHTMLElement3 element = (IHTMLElement3)node;
string val = "";
foreach (char c in value)
{
IHTMLEventObj eventObject = document.CreateEventObject(ref refObj);
eventObject.keyCode = c;
object eventRef = eventObject;
val += c;
element.FireEvent("onKeyDown", ref eventRef);
element.FireEvent("onKeyPress", ref eventRef);
node.setAttribute("value", val, 0);
element.FireEvent("onKeyUp", ref eventRef);
}
}
static void DispatchEventNew(ArbWebControl awc, int iIndex, string sControl, IHTMLSelectElement ihie, IHTMLOptionElement ihoe, IHTMLDocument2 oDoc2)
{
ihoe.selected = true;
object dummy = null;
IHTMLDocument4 oDoc4 = awc.Document4;
object eventObj = oDoc4.CreateEventObject(ref dummy);
var obj = eventObj;
obj.initEvent("onchange", true, true);
IHTMLEventObj2 obj2 = (IHTMLEventObj2)eventObj;
HTMLSelectElementClass hsec = ihie as HTMLSelectElementClass;
awc.ReportNavState("Before FireEvent");
hsec.selectedIndex = iIndex;
obj2.fromElement = hsec;
obj2.srcElement = hsec;
obj2.propertyName = sControl;
bool f;
oDoc4.FireEvent("change", ref eventObj);
awc.ReportNavState("After FireEvent");
}
/// <summary>
/// Helper method which will retrieve the IHTMLDocument2 for the supplied window,
/// or return the first if none is supplied.
/// </summary>
/// <param name="browserWindow">The WindowDetails to get the IHTMLDocument2 for</param>
/// <param name="document2">Ref to the IHTMLDocument2 to return</param>
/// <returns>The WindowDetails to which the IHTMLDocument2 belongs</returns>
private static DocumentContainer CreateDocumentContainer(WindowDetails browserWindow)
{
DocumentContainer returnDocumentContainer = null;
WindowDetails returnWindow = null;
IHTMLDocument2 returnDocument2 = null;
// alternative if no match
WindowDetails alternativeReturnWindow = null;
IHTMLDocument2 alternativeReturnDocument2 = null;
// Find the IE windows
foreach (WindowDetails ieWindow in GetIEWindows())
{
LOG.DebugFormat("Processing {0} - {1}", ieWindow.ClassName, ieWindow.Text);
Accessible ieAccessible = null;
WindowDetails directUIWD = IEHelper.GetDirectUI(ieWindow);
if (directUIWD != null)
{
ieAccessible = new Accessible(directUIWD.Handle);
}
if (ieAccessible == null)
{
if (browserWindow != null)
{
LOG.InfoFormat("Active Window is {0}", browserWindow.Text);
}
if (!ieWindow.Equals(browserWindow))
{
LOG.WarnFormat("No ieAccessible for {0}", ieWindow.Text);
continue;
}
LOG.DebugFormat("No ieAccessible, but the active window is an IE window: {0}, ", ieWindow.Text);
}
try {
// Get the Document
IHTMLDocument2 document2 = getHTMLDocument(ieWindow);
if (document2 == null)
{
continue;
}
// Get the content window handle for the shellWindow.Document
IOleWindow oleWindow = (IOleWindow)document2;
IntPtr contentWindowHandle = IntPtr.Zero;
if (oleWindow != null)
{
oleWindow.GetWindow(out contentWindowHandle);
}
if (contentWindowHandle != IntPtr.Zero)
{
// Get the HTMLDocument to check the hasFocus
// See: http://social.msdn.microsoft.com/Forums/en-US/vbgeneral/thread/60c6c95d-377c-4bf4-860d-390840fce31c/
IHTMLDocument4 document4 = (IHTMLDocument4)document2;
if (document4.hasFocus())
{
LOG.DebugFormat("Matched focused document: {0}", document2.title);
// Look no further, we got what we wanted!
returnDocument2 = document2;
returnWindow = new WindowDetails(contentWindowHandle);
break;
}
try {
if (ieWindow.Equals(browserWindow))
{
returnDocument2 = document2;
returnWindow = new WindowDetails(contentWindowHandle);
break;
}
else if (ieAccessible != null && returnWindow == null && document2.title.Equals(ieAccessible.IEActiveTabCaption))
{
LOG.DebugFormat("Title: {0}", document2.title);
returnDocument2 = document2;
returnWindow = new WindowDetails(contentWindowHandle);
}
else
{
alternativeReturnDocument2 = document2;
alternativeReturnWindow = new WindowDetails(contentWindowHandle);
}
} catch (Exception) {
alternativeReturnDocument2 = document2;
alternativeReturnWindow = new WindowDetails(contentWindowHandle);
}
}
} catch (Exception e) {
LOG.ErrorFormat("Major problem: Problem retrieving Document from {0}", ieWindow.Text);
LOG.Error(e);
}
}
// check if we have something to return
if (returnWindow != null)
{
// As it doesn't have focus, make sure it's active
returnWindow.Restore();
returnWindow.GetParent();
// Create the container
try {
returnDocumentContainer = new DocumentContainer(returnDocument2, returnWindow);
} catch (Exception e) {
LOG.Error("Major problem: Problem retrieving Document.");
LOG.Error(e);
}
}
if (returnDocumentContainer == null && alternativeReturnDocument2 != null)
{
// As it doesn't have focus, make sure it's active
alternativeReturnWindow.Restore();
alternativeReturnWindow.GetParent();
// Create the container
try {
returnDocumentContainer = new DocumentContainer(alternativeReturnDocument2, alternativeReturnWindow);
} catch (Exception e) {
LOG.Error("Major problem: Problem retrieving Document.");
LOG.Error(e);
}
}
return(returnDocumentContainer);
}
private bool SetupDocumentChangeNotification(IHTMLDocument4 pDoc4)
{//SetWinEventHook(EVENT_OBJECT_CREATE). könnte für DOM Changes auch hergenommen werden
if (pDoc4 == null)
return false;
IMarkupContainer2 m_pMarkupContainer2 = pDoc4 as IMarkupContainer2;
if (m_pMarkupContainer2 == null)
return false;
//To stop receiving change notifications, call Release on the IHTMLChangeLog pointer.
//m_pChangeLog = null;
int iret2 = m_pMarkupContainer2.CreateChangeLog(this, out m_pChangeLog, true, true);
int iret = m_pMarkupContainer2.RegisterForDirtyRange(this, out m_pDirtyRangeCookie);
return (iret == Hresults.S_OK) ? true : false;
}
private bool SetupDocumentChangeNotification(IHTMLDocument4 pDoc4)
{
if (pDoc4 == null)
return false;
IMarkupContainer2 m_pMarkupContainer2 = pDoc4 as IMarkupContainer2;
if (m_pMarkupContainer2 == null)
return false;
//To stop receiving change notifications, call Release on the IHTMLChangeLog pointer.
//m_pChangeLog = null;
//int iret = m_pMarkupContainer2.CreateChangeLog(this, out m_pChangeLog, true, true);
int iret = m_pMarkupContainer2.RegisterForDirtyRange(this, out m_pDirtyRangeCookie);
return (iret == Hresults.S_OK) ? true : false;
}
/// <summary>
/// Helper method which will retrieve the IHTMLDocument2 for the supplied window,
/// or return the first if none is supplied.
/// </summary>
/// <param name="browserWindowDetails">The WindowDetails to get the IHTMLDocument2 for</param>
/// <param name="document2">Ref to the IHTMLDocument2 to return</param>
/// <returns>The WindowDetails to which the IHTMLDocument2 belongs</returns>
private static DocumentContainer GetDocument(WindowDetails activeWindow)
{
DocumentContainer returnDocumentContainer = null;
WindowDetails returnWindow = null;
IHTMLDocument2 returnDocument2 = null;
// alternative if no match
WindowDetails alternativeReturnWindow = null;
IHTMLDocument2 alternativeReturnDocument2 = null;
// Find the IE window
foreach (WindowDetails ieWindow in WindowDetails.GetAllWindows("IEFrame"))
{
LOG.DebugFormat("Processing {0} - {1}", ieWindow.ClassName, ieWindow.Text);
Accessible ieAccessible = null;
WindowDetails directUIWD = GetDirectUI(ieWindow);
if (directUIWD != null)
{
ieAccessible = new Accessible(directUIWD.Handle);
}
if (ieAccessible == null)
{
LOG.InfoFormat("Active Window is {0}", activeWindow.Text);
if (!ieWindow.Equals(activeWindow))
{
LOG.WarnFormat("No ieAccessible for {0}", ieWindow.Text);
continue;
}
LOG.DebugFormat("No ieAccessible, but the active window is an IE window: {0}, ", ieWindow.Text);
}
try {
// Get the Document
IHTMLDocument2 document2 = null;
uint windowMessage = User32.RegisterWindowMessage("WM_HTML_GETOBJECT");
if (windowMessage == 0)
{
LOG.WarnFormat("Couldn't register WM_HTML_GETOBJECT");
continue;
}
WindowDetails ieServer = ieWindow.GetChild("Internet Explorer_Server");
if (ieServer == null)
{
LOG.WarnFormat("No Internet Explorer_Server for {0}", ieWindow.Text);
continue;
}
LOG.DebugFormat("Trying WM_HTML_GETOBJECT on {0}", ieServer.ClassName);
UIntPtr response;
User32.SendMessageTimeout(ieServer.Handle, windowMessage, IntPtr.Zero, IntPtr.Zero, SendMessageTimeoutFlags.SMTO_NORMAL, 1000, out response);
if (response != UIntPtr.Zero)
{
document2 = (IHTMLDocument2)Accessible.ObjectFromLresult(response, typeof(IHTMLDocument).GUID, IntPtr.Zero);
if (document2 == null)
{
LOG.Error("No IHTMLDocument2 found");
continue;
}
}
else
{
LOG.Error("No answer on WM_HTML_GETOBJECT.");
continue;
}
// Get the content window handle for the shellWindow.Document
IOleWindow oleWindow = (IOleWindow)document2;
IntPtr contentWindowHandle = IntPtr.Zero;
if (oleWindow != null)
{
oleWindow.GetWindow(out contentWindowHandle);
}
if (contentWindowHandle != IntPtr.Zero)
{
// Get the HTMLDocument to check the hasFocus
// See: http://social.msdn.microsoft.com/Forums/en-US/vbgeneral/thread/60c6c95d-377c-4bf4-860d-390840fce31c/
IHTMLDocument4 document4 = (IHTMLDocument4)document2;
if (document4.hasFocus())
{
LOG.DebugFormat("Matched focused document: {0}", document2.title);
// Look no further, we got what we wanted!
returnDocument2 = document2;
returnWindow = new WindowDetails(contentWindowHandle);
break;
}
try {
if (ieWindow.Equals(activeWindow))
{
returnDocument2 = document2;
returnWindow = new WindowDetails(contentWindowHandle);
break;
}
else if (ieAccessible != null && returnWindow == null && document2.title.Equals(ieAccessible.IEActiveTabCaption))
{
LOG.DebugFormat("Title: {0}", document2.title);
returnDocument2 = document2;
returnWindow = new WindowDetails(contentWindowHandle);
}
else
{
alternativeReturnDocument2 = document2;
alternativeReturnWindow = new WindowDetails(contentWindowHandle);
}
} catch (Exception) {
alternativeReturnDocument2 = document2;
alternativeReturnWindow = new WindowDetails(contentWindowHandle);
}
}
} catch (Exception e) {
LOG.Error(e);
LOG.DebugFormat("Major problem: Problem retrieving Document from {0}", ieWindow.Text);
}
}
// check if we have something to return
if (returnWindow != null)
{
// As it doesn't have focus, make sure it's active
returnWindow.Restore();
returnWindow.GetParent();
// Create the container
returnDocumentContainer = new DocumentContainer(returnDocument2, returnWindow);
}
if (returnDocumentContainer == null && alternativeReturnDocument2 != null)
{
// As it doesn't have focus, make sure it's active
alternativeReturnWindow.Restore();
alternativeReturnWindow.GetParent();
// Create the container
returnDocumentContainer = new DocumentContainer(alternativeReturnDocument2, alternativeReturnWindow);
}
return(returnDocumentContainer);
}
public HtmlRenderer(IArachnodeDAO arachnodeDAO)
{
Type htmldoctype = Type.GetTypeFromCLSID(Iid_Clsids.CLSID_HTMLDocument, true);
//Using Activator inplace of CoCreateInstance, returns IUnknown
//which we cast to a IHtmlDocument2 interface
MSHTML = (IHTMLDocument4)Activator.CreateInstance(htmldoctype);
IPersistStreamInit ips = (IPersistStreamInit)MSHTML;
ips.InitNew();
IOleObject oleObject = (IOleObject)MSHTML;
//Set client site
int iret = oleObject.SetClientSite(this);
CrawlRequestTimeoutInMinutes = 1;
//Getting exceptions when trying to get change notification through IPropertyNotifySink and connectionpointcontainer.
//So, using this technique.
Thread t = new Thread(() =>
{
string previousReadyState = "";
while (true)
{
try
{
if (_parse.WaitOne())
{
if (DateTime.Now.Subtract(_parseStartTime).TotalMinutes < CrawlRequestTimeoutInMinutes)
{
string currentReadyState = ((IHTMLDocument2)MSHTML).readyState;
if (string.Compare(currentReadyState, previousReadyState, true) != 0)
{
previousReadyState = currentReadyState;
ReadyState = ReadyState.Uninitialized;
switch (currentReadyState.ToLower())
{
case "loading":
ReadyState = ReadyState.Loading;
break;
case "loaded":
ReadyState = ReadyState.Loaded;
break;
case "interactive":
ReadyState = ReadyState.Interactive;
ModifyDOM(((IHTMLDocument2)MSHTML), true);
break;
case "complete":
ReadyState = ReadyState.Complete;
break;
default:
ReadyState = ReadyState.Uninitialized;
break;
}
if (ReadyStateChange != null)
{
ReadyStateChange(this, new ReadyStateChangeEventArgs(ReadyState, ((IHTMLDocument2)MSHTML)));
}
}
if (ReadyState == ReadyState.Complete)
{
EndRendering();
previousReadyState = "";
}
}
else
{
throw new Exception("The AbsoluteUri timed out while rendering.");
}
Thread.Sleep(100);
}
}
catch (Exception exception)
{
EndRendering();
previousReadyState = "";
arachnodeDAO.InsertException(AbsoluteUri, AbsoluteUri, exception, false);
}
}
});
t.SetApartmentState(ApartmentState.STA);
t.Start();
}
/// <summary>
/// Fires event on the document object
/// </summary>
/// <param name="doc"></param>
/// <param name="eventName"></param>
/// <param name="e"></param>
/// <returns></returns>
private bool FireEvent(IHTMLDocument4 doc, string eventName, EventArgs e)
{
int x = Cursor.Position.X;
int y = Cursor.Position.Y;
if (e is MouseEventArgs)
{
x = (e as MouseEventArgs).X;
y = (e as MouseEventArgs).Y;
}
//object dummy = null;
//object oEvt = doc.CreateEventObject(ref dummy);
object oEvt = doc.CreateEventObject();
var evt = (IHTMLEventObj2)oEvt; //cast
//set various properties of the event here.
evt.clientX = x;
evt.clientY = y;
//fire
return doc.FireEvent(eventName, oEvt);
}
/// <summary>
/// Public constructor.
/// </summary>
/// <param name="html">Html string to be processed.</param>
public HtmlProcessor(string html)
{
if (string.IsNullOrEmpty(html))
{
throw new ArgumentException("HTML string can not be null or empty.", "html");
}
this.documentSource = html;
this.doc2 = new HTMLDocumentClass();
this.doc2.write(new object[] { html });
//while (this.doc2.readyState != "complete" && this.doc2.readyState != "interactive")
//{
// System.Windows.Forms.Application.DoEvents();
//}
this.doc3 = (IHTMLDocument3)this.doc2;
this.doc4 = (IHTMLDocument4)this.doc2;
this.doc5 = (IHTMLDocument5)this.doc2;
this.InnerText = (this.doc2.body == null) ? null : this.doc2.body.innerText;
this.Title = this.doc2.title;
this._elements = new List <DomElement>();
IHTMLElementCollection collection = this.doc2.all;
foreach (var item in collection)
{
var element = (IHTMLElement)item;
if (element != null)
{
DomElement domElement = new DomElement
{
Class = element.className,
InnerHtml = element.innerHTML.TrimSafe(),
OuterHtml = element.outerHTML.TrimSafe(),
InnerText = element.innerText.TrimSafe(),
Id = element.id,
TagName = element.tagName.ToLowerInvariant()
};
domElement.Attributes = new NameValueCollection();
if (element.outerHTML.IsNullOrEmpty())
{
continue;
}
//@"<INPUT onblur=google&&google.fade&&google.fade() class=lst title='Google Search' value=TESTING maxLength=2048 size=55 name=q autocomplete='off' init='true'/>";
string node = element.outerHTML.Substring(0, element.outerHTML.IndexOf(">") + 1);
string pattern = @"
(?:<)(?<Tag>[^\s/>]+) # Extract the tag name.
(?![/>]) # Stop if /> is found
# -- Extract Attributes Key Value Pairs --
((?:\s+) # One to many spaces start the attribute
(?<Key>[^=]+) # Name/key of the attribute
(?:=) # Equals sign needs to be matched, but not captured.
(?([\x22\x27]) # If quotes are found
(?:[\x22\x27])
(?<Value>[^\x22\x27]+) # Place the value into named Capture
(?:[\x22\x27])
| # Else no quotes
(?<Value>[^\s/>]*) # Place the value into named Capture
)
)+ # -- One to many attributes found!";
//http://omegacoder.com/?p=512
var attr = (from Match mt in Regex.Matches(node, pattern, RegexOptions.IgnorePatternWhitespace)
select new
{
Name = mt.Groups["Tag"],
Attrs = (from cpKey in mt.Groups["Key"].Captures.Cast <Capture>().Select((a, i) => new { a.Value, i })
join cpValue in mt.Groups["Value"].Captures.Cast <Capture>().Select((b, i) => new { b.Value, i }) on cpKey.i equals cpValue.i
select new KeyValuePair <string, string>(cpKey.Value, cpValue.Value)).ToDictionary(kvp => kvp.Key, kvp => kvp.Value)
}).FirstOrDefault();
if (attr != null)
{
var attributes = attr.Attrs;
foreach (KeyValuePair <string, string> kvp in attributes)
{
if (!string.IsNullOrEmpty(kvp.Value))
{
domElement.Attributes.Add(kvp.Key.ToLowerInvariant(), kvp.Value);
}
}
}
this._elements.Add(domElement);
}
}
}
