Project Blog

Week 18: Meeting on Sep. 24

Showed off how getPostingSlice() for IndexShard returns documents in reverse order.


To-Do List:
  • Write a small script to generate posting lists for testing.
  • Work on the iterator so that it will move through generations in reverse order.


Week 17: Meeting on Sep. 17

Talked about word offsets for reaching the posting lists, and how Yioop does a lot of the bit fiddling to conserve space.


To-Do List:
  • Work on IndexShard methods so that they return results in reverse.


Week 16: Meeting on Sep. 10

Discussed further details and next steps to take in approaching this project. Talked about how Yioop's current use of the feed shard, and that by the end it should be gone.


To-Do List:
  • Work on IndexShard so that it has a flag determining direction. Direction will be used later for iterators and should also affect some methods of IndexShard too.


Week 15: Meeting on Sep. 3

Looked over the CS 298 proposal. Discussed Yioop classes to look into for inspiration.


Start of CS 298

Week 14: Meeting on May. 14

Handed in "final" draft of the CS 297 report.


Week 13: Meeting on May. 7

Handed in first draft of CS 297 report for proofreading and scrutiny. Looked over official public release of the Word Tracker tool.


To-Do List:
  • Make small changes to CS 297 report, with respect to grammatical and consistency errors.


Week 12: Meeting on Apr. 30

More detail into what might need to be done in order to create a index bundle that goes from most recent to oldest.
Talked about the requirements for the CS 297 final report.


To-Do List:
  • Make rough draft of the CS 297 report.


Week 11: Meeting on Apr. 23

Fixed broken database for Yioop, which was introduced in the more recent updates.
Showed the news suggestion form for regular users.
Discussed ways to improve on the user suggestion feature, where root users should just be able to log in and enable sources.
Some insight into how to start thinking about the FeedArchiveBundle, or the iterator for it.


To-Do List:
  • Make patch for Deliverable 2.
  • Enhance the suggestion feature by having items added to database but not enabled.


Week 10: Meeting on Apr. 16

Discussed how to add new activities for Yioop users through database


To-Do List:
  • Make patch for Deliverable 2.
  • Work on adding a news source suggest form for regular users.


Week 9: Meeting on Apr. 9

Discovered some problem with Yioop, where the local server could not reliably start
Showed Deliverable 3, which is a mock up of how non privileged users could suggest News sources for Yioop.


To-Do List:
  • Make patch for Deliverable 2.
  • Attempt to implement design from Deliverable 3 into Yioop.


Week 8: No meeting on Apr. 2 due to spring break


Week 7: Meeting on Mar. 26

Demonstrated Deliverable 2 functionality working in a general state.
.


To-Do List:
  • Make the deliverable look a bit nicer.


Week 6: Meeting on Mar. 19

Showed current progress on Deliverable 2.
Talked about how routing works generally in Yioop, and how web pages are represents as elements which are rendered by some view.


To-Do List:
  • Implement a trending page to the Yioop website.


Week 6: Meeting on Mar. 12

Showed current progress on Deliverable 2.
Discussed the UI that the word tracking could be presented in.


To-Do List:
  • Expand work on word track to compute top hourly, daily and weekly stats.


Week 5: Meeting on Mar. 5

Presented slides for Chapter 2(Clustering).


To-Do List:
  • Start work on the top word tracking job.


Week 4: Meeting on Feb. 26

Presented slides for Chapter 1 and Chapter 2(JRFL).
Talked about the possibility of combining relevance and freshness into Yioop. But likely not happening right now.


To-Do List:
  • Make slides for Chapter 2(Clustering).
  • Start looking into the FeedUpdate and maybe WikiUpdate jobs.


Week 3: Meeting on Feb. 19

Ran through IndexShard unit tests and confirmed preexisting issues, likely due to old tests.
Came up with more substantial deliverables that would be useful to the project.


To-Do List:
  • Revise proposal with new deliverables.
  • Read chapters 1 and 2 of the book "Relevance Ranking for Vertical Search Engines".


Week 2: Meeting on Feb. 12

Discussed way to find more research papers by using the book "Relevance Ranking for Vertical Search Engines".
Went through the code for the IndexShard class and broke down what individual methods do.


To-Do List:
  • Go through the unit tests for IndexShard.
  • Find out where IndexShard is actually being stored on disk.
  • Add more papers to the proposal.


Week 1: Meeting on Feb. 5

Discussed long term project goals, as well as steps needed to achieve said goal.
Plotted out more concrete deliverables required for 297.
Talked about the current method of how Yioop stores newsfeed indexes via index shards which are loaded into memory, thus causing the bottleneck.


To-Do List:
  • Make necessary changes to 297 proposal.
  • Look into the WebArchive and IndexShard classes, as well as the associated unit tests.
  • Try to find research papers possibly related to inverted indexing storage or vertical search optimization.