-- PatrickEhlen - 16 Aug 2006

Meeting Browser Magic URLs

The following are "magic URLs" that can be called to the Meeting Browser Server that will produce the underlying XML used by the browser. So if something isn't working right, you can mess around with these and enter them in a web browser address bar and make sure you're getting the correct XML information to the browser in the first place.

The following parameters should be subsituted with your own information:

[MeetingServerHost] = Your CALO Meeting Browser server host

[participantName] = The login name of a particular meeting participant (like "Patrick_Ehlen")

[meetingName] = The "short" meeting name of a particular meeting

[queryWords] = A list of query words you want to have associated with a topic (multiple words are separated by a "+")

[numTopics] = The number of topics you want returned when you do a topic query

General Browser Access

Typically, you would access the CALO Meeting Browser at:

http://[MeetingServerHost]:8080/mabrowser/index.html

Login

http://[MeetingServerHost]:8080/whatever.html?type=login&personId=[participantName]

Logs in a user. (Actually, does nothing!)

Get Users

http://[MeetingServerHost]:8080/whatever.html?type=requestUsers

Returns a list of every user who has participated in a meeting in the Annotation Manager knowledgebase, and can therefore log in to view meetings and submit feedback.

The XML looks like:

<participants>
  <participant id="http://godel.stanford.edu/corpora_2_0#52998118f82480111973363e1498154f">Patrick_Ehlen</participant> 
  <participant id="http://godel.stanford.edu/corpora_2_0#2672db244c51dbfef5e01046611f27b8">John_Niekrasz</participant> 
...
</participants>

The "id" attribute of each of the <participant> nodes is the object ID used by the Annotation Manager to signify a meeting participant (or a Person Object in the meeting ontology), and the text of that node is the username of that Person object.

Get Meetings

http://[MeetingServerHost]:8080/whatever.html?type=requestMeetings&personId=[participantName]

Returns a list of meetings that can be accessed from the Annotation Manager.

The XML looks like:

<meetings>
  <meeting>CSLI-N04-24-04-2006</meeting> 
  <meeting>CSLI-N02-07-04-2006</meeting> 
  <meeting>CSLI-July-10</meeting> 
...
</meetings>

The text of each <meeting> node is the "short" meeting name for each meeting that is available in the Annotation Manager.

Get Notes & Hypotheses

http://[MeetingServerHost]:8080/whatever.html?type=requestHyps&meetingId=[meetingName]&personId=[participantName]

Returns the SmartNotes-style notes XML along with any automatically-generated hypotheses for a particular meeting in the Annotation Manager.

The XML is too complicated to give an example of here, but follows the CALO Mercury/SmartNotes format.

Topic Query

http://[MeetingServerHost]:8080/whatever.html?type=topicQuery&q=[queryWords]&n=[numTopics]

Returns the top numTopics from the Annotation Manager Topic Pool that are associated with the query words.

The XML looks like:

<topics>
 <topic>
  <topic_name>ontolog-creat-object-knowledg-feedback</topic_name> 
  <word_vector>[feedback, service, object, knowledge, p, c, o, publish, ontology, create, bean, servlet, server, method, store]</word_vector> 
  <word_counter>{o=0.35034560564067724, intern=0.1677335989691516, c=0.4292732678495738, upset=0.07869239785955466, 
      knowledge=0.6358409011919057, p=0.5598263724574661, bean=0.2744075408414073, feedback=2.0254734318633427, publish=0.3111521161305771, 
      object=0.8428800087993131, patrick=0.07899422499788179, service=0.9548283993170744, servlet=0.26302037657379523, 
      ontology=0.3089976296324714, store=0.17822817886524206, create=0.30834605893125355, recent=0.16997211383135322, 
      replace=0.07828658672474052, server=0.2329723045361258, method=0.19688661751684605}</word_counter> 
  </topic>
...
</topics>

For each <topic> node, there is a <topic_name> node that holds the name of the topic, a <word_vector> node that lists the top 15 words for that topic, and a <word_counter> node, which provides a score for each word in the topic vector as determined by its position in the word distribution for that topic.