19 May, 2011
1 commit
12 May, 2011
1 commit
11 May, 2011
2 commits
20 Apr, 2011
2 commits
-
before 2010-02-17 we did not track losses, so for any question with votes prior to that date we skip some tests.
14 Apr, 2011
1 commit
-
Any question with votes before 2010-02-17 didn't keep track of the loser. We need to ignore some tests for those questions.
18 Mar, 2011
1 commit
-
only generate prompts if cache is less than 90% full
04 Jan, 2011
2 commits
21 Dec, 2010
1 commit
-
add indexes to votes table to speed up task, and should speed up other queries we do
17 Nov, 2010
2 commits
-
also, add new lines to several errors
21 Oct, 2010
2 commits
18 Oct, 2010
3 commits
15 Oct, 2010
1 commit
13 Aug, 2010
3 commits
11 Aug, 2010
1 commit
29 Jul, 2010
1 commit
02 Jul, 2010
1 commit
21 Jun, 2010
2 commits
19 Jun, 2010
1 commit
10 Jun, 2010
1 commit
27 May, 2010
1 commit
20 May, 2010
1 commit
17 May, 2010
3 commits
03 May, 2010
1 commit
16 Apr, 2010
1 commit
12 Apr, 2010
1 commit
06 Apr, 2010
1 commit
31 Mar, 2010
2 commits
-
(1) We calculate the density of votes over time to see if the catchup algorithm is having an effect. For each question, there are four types of prompts (seed vs seed, seed vs nonseed, nonseed vs seed, and nonseed vs nonseed). Over time, the density of all types of prompts should converge. This model will allow us to track over time (2)Rake task will be run every night via a cron job. Simply calculates and saves the four density records for each question in the db.