This Interface to Google is very different from other Google interfaces in TWiki, the search results are formatted in XML. The result content can be filtered and combined. The links provided from Google can even be accessed and used in the page formation process. The QueryString formed to pass to Google can specify to look for XML/RDF content. Other Web Services can be accessed in the same way.
Attached files below represent 'executable stylesheets' in the form of XSL.
XslSample is the next version (0.2) of search, its goal is to make search available to any page and separate source files with a core set in a base directory, and 1 or 2 files attached to the users page. It currently searches on source from TWiki.WebHome(configured in its attached somekeys.xml). Soon it will use the local page content.
Next version of search will send
WikiNames (with spaces added) to Google.
Google's
REST interface (Xoomle)is accessed from a TWiki edit session. This eliminates the need for
SOAP to access
WebService s.
It shows the mix of Twiki edit page mixed with searches on wiki content, returned to the textarea control.
but if you have TWiki tags marked with ---++++++, the link below points to a working search interface.
---+++++ The location change to the Sandbox has caused some failures. If you cannot access, come back later and retry.
Page lock is a problem, after attempting a search cancel edit window.
RDFWeb access tool, which is a simple blend with
TWikiThe The TWiki edit form is valid
XHTML, therefore it is a valid document to process with the XSL document function. When the edit form is processed with XSL,
other sources can be mixed to form a edit form with
REST type interface. This interface allows requests to
WebService to be encode into the URL, and
XML to be returned. in the example here.
Search content format needs more adjustment.
Mozilla 1.3a/ NOT SUPPORTED There is one MS shortcut that is used. (This is fixable), but first effort is to define scope of XSL/Wiki options. -- JC
This version is safer, because it writes to a different page than it reads from.(
GoogleResultXslWiki) and makes the current page its
TopicParent. --
JC
XslREST will display the search result.
Here is the textarea from the edit form.
>
UDDI can be accessed as a REST data source this demo has UDDI data at the bottom of the edit form that was fetched at the time of the edit form page creation.
*
UDDIAccess.xml: trigger a
UDDI search(and Amazon) on topic of current page results are formated with count of results and 2 links for each topic.
*
Visual TWiki Navigator
Should work on all browsers, but needs a
SVG plugin.
* Some Green dots are not all working yet.
* node position will be made to not overlap and produce tree structures.
*
Fitzpatrick1789.rdf: Family structure of a man born in 1789.
Genealogy UDDI SurName index
*
BookSearchAmazon is a book information access from a
REST data source,(NOT YET WORKING,data hase been accessed, but Wiki data access is still a problem) this demo will get image and author at the time of the edit form creation.
please send comment on formatting results to:
johncoe42@yahooSTOPSPAMPLEASENOSPAM.com
GoogleML:
Email:
johncoe42@yahooPLEASENOSPAM.com
Company Name: Forth Party Software
--
JohnCoe - 11 Mar 2004
--
JohnCoe - 12 Mar 2004
going to
https://twiki.org/p/pub/Codev/XslWiki/GoogleML.xml crashes Mozilla 1.3a/Win.
--
MattWilkie - 14 Mar 2004
There could be a autoexec.xsl to your page. --
JC
These notes are to map perl syntax terms to Xsl syntax.
<xsl:templete match="token"> is roughly the same as if(/m/token
<xsl:templete match="token"> is roughly the same as if(/s/token/replace
<replace>
This looks very interesting, but I couldn't get any of the examples to work in IE 5.5 / Win 2000, or Mozilla Firefox 0.8. Can you post details of the browser you have used for testing? Also, screen shots would be useful so people can see a working example even if their browser doesn't support the demos.
I like the idea of auxiliary searches when editing a page, and using
REST in innovative ways - could be very handy for public TWikis or intranet TWikis that have a good intranet search engine with a
REST interface. I've recently been trying out the Amazon
WebServices REST API using Perl (with
CPAN:XML::Simple
and
CPAN:LWP::Simple
), and it's really easy to do powerful searches without much code.
--
RichardDonkin - 17 Mar 2004
Most of the problems described by Richard have been fixed, the search still will not work in Mozilla, but the
Visual TWiki Navigator will.
JC
Fails for me with Security exception IE6/XP.
--
MartinCleaver - 23 Mar 2004
This last logged failure was on the FOAF
SVG utility. The security model for
SVG is very different than the
REST access above. I will create a seperate Codev area for it soon. There is a sandbox project
GenealogyRDF that is extending the FAOF viewer to
RDF files. The sample files in this case is for genealogy. The
RDF files are TWiki file attachments.
some firewalls will block access for this interface. There are 2 levels of security that are in play here.
- 1 This code uses JavaScript to parse the textarea content. System Security will block this from a non-trusted source. If you copy the files to your system, it should be within your trusted area. With some pathing changes,
most of the files could be left in TWiki, and referenced from the local xsl file inside your trust area.
If TWiki content were available in
XML(or even
XHTML) the script would not be needed.
OpenWiki as
XML source would work better. Maybe TWiki need a
XmlSkin ??
- 2 XmlResolver object will limit trusted sources for the Xsl document("http...") access.
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnexxml/html/xml07212003.asp
more on this later.
--
JohnCoe - 24 Mar 2004
Hi John, I don't understand what you are working on here, but it looks
very interesting. Hopefully with time my thick-headedness can be alleviated.
I just wanted to let you know that lack of response doesn't equate to lack of interest.
--
MattWilkie - 29 Mar 2004