wiki:Event/2013/GSoC/MessageParsing

Source Code

BluePrints

This project draws ideas from the two Blueprints below:

Weekly Meeting

Time : Thursday 1300 UTC

Venue : Skype (Contact - ashwyn10)

Meeting Doc : https://docs.google.com/document/d/1Mln0QSZQ_SNdQziMwhdlqIKRNXhKDqXRlAtFBZdfp2Y/edit

Milestones/Progress

Goal Remarks Status
S3ChannelModel redefined https://github.com/ashwyn/eden/commit/43887059be9d197203edd94e052a1f933e21cd8e Done.
RSS Integration https://github.com/ashwyn/eden/commit/644e9dcc752ccec368c545973ec5e3b48172fbfc Done.
Identifying Relevant Users [Geo-Typical User Identification https://docs.google.com/document/d/1MJlOICZtHIEPiw0Tp6mAMUybILJw_FiomJ8T5fLOHZY/edit]- Using KeyGraph now.Done.
Super Entity msg_message Link all inboxes/outboxes to msg_message Done for Email, Twilio and RSS.
Evaluating KeyGraph on the local system Topic detection helps filter out irrelevant information Done.
Tweet Crawler Identified and used TweetSearch(https://github.com/ashwyn/TwitterSearch) as the best available option that uses Twitter REST API 1.1 Done.
Tweaking KeyGraph code to archive into a jar file. https://github.com/ashwyn/KeyGraph Done.
Interfacing KeyGraph with the database The KeyGraph code needs to be tweaked so that it can interact with the DB using Py4J Done.
Integrated support for Py4J Gateway to interact with Java applications using python Done.
Preprocssing of Tweets to feed KeyGraph Identified tagdef. Done.
Visualisation of parsed output from KeyGraph Igraph and Networkx visualisation support Done.
Marking senders Marking senders as trusted/untrusted Done

Future Work

  • Interactive keyword highlighting - Selecting and highlighting tool for adding keyword
  • Interactive visualisations for KeyGraph results.
  • Adopt the DMPRP template for UI enhancements.
  • Fetch Facebook feeds
  • Use progress bars for searching and analysing actions.
  • API support for Py4J

Documentation

Last modified 8 years ago Last modified on 09/23/13 17:28:00
Note: See TracWiki for help on using the wiki.