Skip to content
jog edited this page Nov 30, 2010 · 44 revisions

#Installation Guide

  • Setup an Amazon AMI.
  • SSH in and install mysql.
  • Change the root password via mysqladmin, and login to mysql as root.
  • Create a new account called "dsadmin" and set an appropriate password:
    CREATE USER 'dsadmin'@'localhost' IDENTIFIED BY 'yourpassword';
  • Create a database called "datasphere":
    CREATE DATABASE datasphere;
  • Grant the "dsadmin" account all privileges to that database:
    GRANT ALL ON datasphere.* TO 'dsadmin'@'localhost' IDENTIFIED BY 'yourpassword';
  • Download the latest datasphere.jar.
  • SFTP the jar to the AMI (or run it on your local machine using an ssh tunnel to forward mysql 3306 to the AMI).
  • The first time you run the jar use the following arguments to setup the system tables:
    java -jar DSCatalog.jar -p 'yourpassword' --create 
  • From then on the --create argument is not required, but (for now) the password for dsadmin should be specified (n.b. this will soon be superceded by a config file). See --help for other options.

  • The create process will register a test account with the jid [email protected] for use in dataware creation.

#DSUpdate XMPP Message

The DSUpdate XMMP message is responsible for communications of activity updates from active dataware (both native and legacy). The message packets have the following common format:

"namespace" : string, //-- namespace of dataware source, as specified by the registry (max 256 chars)

"primaryTag" : string, //-- namespace of the item, defaulting to "dataware:update" (max 256 chars)

"tags" : array of strings, //-- namespaces of associated item categories (interfaces) (stored as text)

"ctime" : unix timestamp, //-- time that the update was generated by the source

"ftime" : unix timestamp, //-- time that the update was forward by the dataware

"loc" : coordinates, //-- WGS-84 coordinate attached to the update

"description" : string, //-- a short (1024 chars) textual summary of the update

"crud" : string, //-- create / read / update / delete

**"total" : long, ** //-- the total number of items represented by this item type

"meta" : array //-- other data associated with the update, specific to its type (stored as text)

#Datasphere source namespaces

All data sources must be identified using a namespace so that the catalog can reference them uniquely. The current list is currently accessible at: http://www.cs.nott.ac.uk/~jog/datasphere/namespaces.html

#Data type namespaces

All data items require a type (as specified in the DSUpdate schema), and these types must also have a referrable namespace. Data times may also be "tagged" with additional namespaces indicating their associations and broader categorizations. Currently the following are available:

##Facebook

  • ds:facebook:bio
  • ds:facebook:post
  • ds:facebook:photo
  • ds:facebook:feed
  • ds:facebook:social
  • ds:facebook:update
  • ds:facebook:like
  • ds:facebook:checkin
  • ds:facebook:avatar

##General

  • ds:microblog
  • ds:bio
  • ds:geo
  • ds:media
  • ds:photo
  • ds:unrecognized
Clone this wiki locally