Thanks cmdskp,
I must have missed your response to the other thread, sorry. I'd be
happy to swap notes. Here's my traversal approach. Apologies for
this long message...
package com.googlecode.wave.myrobot;
import com.google.wave.api.*;
import java.util.logging.Logger;
import java.util.Date;
@SuppressWarnings("serial")
public class MyRobotServlet extends AbstractRobotServlet
{
static String INDENT = " ";
static String ERROR_BULLET = " #";
static String RULE = "------";
private static final Logger logger =
Logger.getLogger(MyRobotServlet.class.getName());
@Override
public void processEvents(RobotMessageBundle bundle)
{
if( bundle.wasSelfAdded() )
{
log( "I've been added to " + bundle.getWavelet().getWaveId() );
actOnBundle( bundle, "self_added" );
}
}
protected void actOnBundle( RobotMessageBundle bundle, String
eventname )
{
Wavelet wavelet = bundle.getWavelet();
StringBuilder list_html = new StringBuilder();
list_html.append("I'm responding to ").append( eventname ).append
("\n\n");
traverseBlipGadgets( wavelet.getRootBlip(), list_html, 0 );
log( list_html.toString() );
}
protected void traverseBlipGadgets( Blip blip, StringBuilder
list_html, int indent_level )
{
StringBuilder indentation = generateIndentation( indent_level );
list_html.append( getBlipLabel( blip, indentation ) );
try
{
for( Blip blippie : blip.getDocument().getInlineBlips() )
{
traverseBlipGadgets( blippie, list_html, indent_level + 1 );
}
}
catch( NullPointerException ex )
{
String warning = indentation + ERROR_BULLET +
"NullPointerException while trying to getInlineBlips()";
list_html.append( warning ).append( "\n" );
logger.warning( warning );
}
try
{
for( Blip blooper : blip.getChildren() )
{
list_html.append( indentation ).append( RULE ).append( "\n" );
traverseBlipGadgets( blooper, list_html, indent_level );
}
}
catch( NullPointerException ex )
{
String warning = indentation + ERROR_BULLET +
"NullPointerException while trying to getChildren()";
list_html.append( warning ).append( "\n" );
logger.warning( warning );
}
}
protected static StringBuilder getBlipLabel( Blip blip,
StringBuilder indentation )
{
StringBuilder blip_label = new StringBuilder();
try
{
blip_label.append( indentation + "Blip Id: " + blip.getBlipId()
+ "\n" );
blip_label.append( indentation + "Parent ID: " +
blip.getParentBlipId() + "\n" );
blip_label.append( indentation + "Has Children: " +
(blip.hasChildren() ? "Y" : "N" ) + "\n" );
}
catch( NullPointerException ex )
{
blip_label.append( indentation + "(NULL Blip Exception #1)\n" );
}
try
{
long lastmod = blip.getLastModifiedTime();
Date meDate = new Date(lastmod);
blip_label.append( indentation + "Last Modified: " +
meDate.toString() + "\n" );
}
catch( NullPointerException ex )
{
blip_label.append( indentation + "(NULL Blip Exception #2)\n" );
}
try
{
blip_label.append( indentation + "Content: " + blip.getDocument
().getText() + "\n" );
}
catch( NullPointerException ex )
{
blip_label.append( indentation + "(NULL Blip Exception #3\n" );
}
return blip_label;
}
protected static StringBuilder generateIndentation( int
indent_level )
{
StringBuilder indentation = new StringBuilder();
for ( int i = 0; i < indent_level; i++ )
{
indentation.append( INDENT );
}
return indentation;
}
}
On Nov 25, 11:25 pm, cmdskp <[email protected]> wrote:
> Good to know it looks like you can combine the context values. Though
> with WAVELET_SELF_ADDED there would not be a parent - since the
> root-node is returned with that event - it is the top-most blip and
> has no parent logically...though I've not checked to see if it's null
> or points to itself, I would expect null there.
>
> I don't know if you've tested my robot cleantxt on WAVELET_SELF_ADDED
> yet as I offered in another thread - it proves it's possible. Another
> quick test on it on a new wave there and the functionality remains
> tonight. Perhaps if you show your recursive code we might spot a
> difference?
>
> However, during my testing tonight I was trying the BLIP_SUBMITTED
> event and with that one I can't get the submitted blip's children, not
> even their Id's - it always returns 0 children, same for
> DOCUMENT_CHANGED.
>
> On Nov 26, 3:50 am, Stephen George <[email protected]> wrote:
>
> > Olreich,
>
> > I'll start by saying that I haven't had much success with getting
> > both.
>
> > However, if you take a look at Google's implementation of the private
> > method AbstractRobot.processCapabilities(), it looks like it *should*
> > be possible by splitting the contexts with commas:
>
> > <w:capability name="WAVELET_SELF_ADDED"
> > context="PARENT,SIBLINGS,CHILDREN" />
>
> > I believe that is at least intention from the code links below. I
> > personally have still not had luck recursively traversing through all
> > children and inline blips from the root blip when my robot is added to
> > a dense wave.
>
> > See
> > also:http://code.google.com/p/wave-robot-java-client/source/browse/trunk/s......
>
> > On Nov 24, 5:18 am, Olreich <[email protected]> wrote:
>
> > > The only remaining question is:
>
> > > Can one specify both parent and child context, and thus get all
> > > blip_data with a single event?
>
> > > On Nov 23, 10:53 pm, cmdskp <[email protected]> wrote:
>
> > > > @qMax: As of today, it appears to delete only the blip and its inline
> > > > replies (the collapseable blips) - not standard reply children.
>
> > > > @shitu: Last night I was testing deleting of all blips on
> > > > WAVELET_SELF_ADDED. It worked fine then, I recursively went through
> > > > the children and all were deleted - it also gave me every blips text
> > > > content. (I did not test inline replies...)
> > > > However, sometimes .delete() will throw an exception in the logs - but
> > > > that issue has been raised on the wave API issue tracker.
>
> > > > There still seems some confusion about the model used by Wave when it
> > > > comes to what blips are given with an event. It's also a moving
> > > > target as bugs get fixed.
>
> > > > In the documentation and guides I've used for the 'capabilities.xml'
> > > > file there is an extra attribute which tells the wave server which
> > > > blips content to send on that event to the robot.
>
> > > > This attribute can be set to at least four values (sadly, I can't find
> > > > the blog page which gave the list), but the main two are:
>
> > > > - parent
> > > > - children
>
> > > > Here's a nice example
> > > > capabilities.xml:http://code.google.com/p/google-wave-resources/source/browse/trunk/sa...
>
> > > > To ensure your get the children for an event, specify
> > > > context="children" as well as content="true" (to get their content)
> > > > for your event. Like so:
> > > > <w:capability name="blip_submitted" content="true" context="children" /
>
> > > > If you want the parent above of the event's blip, specify
> > > > context="parent" as well as content="true" (to get the parent's
> > > > content) for your event. Like so:
> > > > <w:capability name="blip_submitted" content="true" context="parent" />
>
> > > > You always get the content of the event blip I believe - though I
> > > > could be wrong if you set content="false" - never tried that.
>
> > > > The context attribute is also mentioned in passing under the Java API
> > > > method blip.getChild(int index), getChildren(), etc... methods
> > > > here:http://wave-robot-java-client.googlecode.com/svn/trunk/doc/com/google...
>
> > > > "Returns the Nth child blip of this blip (if available). Use the
> > > > isChildAvailable() method to check for availability of the child
> > > > blip.
> > > > Note: The behavior of this method is dependent on the 'context'
> > > > settings in the Capabilities XML configuration. Child blips may not
> > > > have been sent with this event."
>
> > > > Hope that clears up some of the misunderstanding on how you choose
> > > > what data to be sent for an event and what is possible currently.
>
> > > > If you have a sandbox account I recommend you do a search for:
> > > > with:public tag:Office Hours
>
> > > > There's a lot of answers there from Google on various topics not
> > > > covered in the main tutorials - though some of it is out-of-date now
> > > > since the API development is moving forward.
>
> > > > On Nov 23, 3:22 am, Vikram Dhillon <[email protected]> wrote:
>
> > > > > The visual really helped, I wasn't thinking properly so I apologize
> > > > > for that
> > > > > and thanks for your clarification on the subject. So from what I
> > > > > understand
> > > > > now, how about this: if we can export the whole conversation, to a
> > > > > file or
> > > > > something, then the robot can stri[p off everything besides the root
> > > > > blip and
> > > > > then sync it back may be
>
> > > > > --
> > > > > Regards,
> > > > > Vikram Dhillon
>
> > > > > On Sunday 22 November 2009 08:15:13 pm Olreich wrote:
>
> > > > > > GetChildren on the root blip will get the first blip. Unfortunately,
> > > > > > because the root blip has to be the blip submitted to have that
> > > > > > function work, it's limited. In fact, it's even more limited because
> > > > > > of the fact that the root blips child's child is null in the current
> > > > > > version of the API. (an annoying quirk)
>
> > > > > > Visual:
>
> > > > > > r--->Blip 1
> > > > > > ^--->Blip 2
> > > > > > ^--->Blip A
> > > > > > ^--->Blip B
> > > > > > ^--->Blip C
> > > > > > ^--->Blip 3
> > > > > > ^--->Blip 4
> > > > > > r = root
> > > > > > ^ = child of last indentation (Blips A and 3 are both children of
> > > > > > Blip
> > > > > > 2)
>
> > > > > > So in that context, let's say the Event happened on Blip 1. Well,
> > > > > > it's
> > > > > > the root blip, so it has no parents, but it does have a child (Blip
> > > > > > 2). getChildren will return a list with Blip 2 in it. But, if one
> > > > > > get's Blip 2 (in any fashion) because it is not the Blip that was
> > > > > > acted upon, you will get null for both it's parent (Blip 1) and it's
> > > > > > children (Blips A and 3).
>
> > > > > > The reason for this is that a bot works in the same way that a human
> > > > > > participant does, in that every update has just the information
> > > > > > about
> > > > > > the one blip, instead of all the blips, to reduce bandwidth.
>
> > > > > > Something that just came to me would be, upon the bot being added,
> > > > > > it
> > > > > > should have to be able to see ALL the blips (like when you get a new
> > > > > > wave and having the entire conversation load, but only having small
> > > > > > parts change after that). Thus, it should be able to make a full
> > > > > > list
> > > > > > of the blips. Then, when a blip event happens, and the blip is not
> > > > > > on
> > > > > > the list you made, then the Bot adds it to the list. I'm not sure if
> > > > > > this is the functionality, but it should be. And if it's not when I
> > > > > > test it, I'll be adding a bug, because the Robot is not a full
> > > > > > participant unless it is able to read the entire conversation right
> > > > > > off the bat.
>
> > > > > > --
>
> > > > > > You received this message because you are subscribed to the Google
> > > > > > Groups
> > > > > > "Google Wave API" group. To post to this group, send email to
> > > > > > [email protected]. To unsubscribe from this group,
> > > > > > send
> > > > > > email to [email protected]. For more
> > > > > > options,
> > > > > > visit this group
> > > > > > athttp://groups.google.com/group/google-wave-api?hl=.
--
You received this message because you are subscribed to the Google Groups
"Google Wave API" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/google-wave-api?hl=en.
