On this week I was created the working process to use popularity-contest data to make recommendations with the AppRecommender.
This is working, but still not faster and not optimized, because this feature is not on the master branch, the work of this week is on the knn_strategy branch. The issues of this week is on a milestone in AppRecommender repository, being: - Script to generate fake popularity-contest data for tests with AppRecommender - Separate the popcon submissions into clusters based on the packages of each submission - Load the clusters on AppRecommender - Index this clusters on xapian. Xapian is the database used by AppRecommender - Make the collaboratives strategies of AppRecommender works with the popcon clusters - Add progress bar when the data in indexing on xapian Link for the milestone of this week: https://gitlab.com/AppRecommender/AppRecommender/milestones/1 To the next week I work on patch to popularity-contest, to add the script that separate the popcon submissions into clusters, this script already exists, but needs some changes, like the format that script save the files, and the time to run this script. Another task is add a command that permit runs the AppRecommender after the user install a package with apt, and to begin a new strategy on AppRecommender to make a recommendation using the installed package with apt and the data of another users by the popcon submissions Link for the milestone of the next week: https://gitlab.com/AppRecommender/AppRecommender/milestones/2
