[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

GSoC Week 9



What was done
=============

Create privacy text to use popcon data
  - This text explains how AppRecommender wants to use the popularity-contest
  data. On text its informed what is the popularity-contest and his data,
  how to transform this data into clusters to improve the AppRecommender
  recommendation, the privacy risks that are involved, and the measures
  taken to minimize those risks

  - The text its on line 111 of pad https://pad.riseup.net/p/apprec-privacy


Create text for the Debian planet
  - This text explains my work on Google Summer of Code

  - Pad link for this text: https://pad.riseup.net/p/luciano_planet_gsoc


Create another strategy for makes recommendation with a referenced package
  - This strategy its on branch 'package_reference_strategy' on
  repository: https://github.com/TCC-AppRecommender/AppRecommender/

  - It has not been added on AppRecommender master because this code need
  be improve

  - This strategy implements the following steps:
    - Given a list of referenced packages, for example the packages installed
    with 'apt-get install vim python', the list of referenced packages
    contains 'vim' and 'python'.

    - Add on content profile the name of referenced packages, for this
    example its added 'vim' and 'python', this content profile its used as a
    query to xapian database return the recommendations.

    - By default the xapian databases used on recommendation its on
    '~/.apprecommender/axi_desktopapps', but the dependencies of uninstalled
    packages are not indexed on this database, therefore its used the
    '/var/lib/apt-xapian-index/index' as xapian databases, because this
    databases contains all debian packages.

    - Run recommendation and filter the packages, to recommends only the
    packages that have one of packages on list of referenced packages as
    dependency. Example: for vim can recommends vim-gtk and vim-athena, and
    for python can recommends founded python-numpy.


Milestone: https://gitlab.com/AppRecommender/AppRecommender/milestones/13


To the next week
================

- Send the popcon patch and the text explaining how the data is used by email

- Fix text about privacy text to use popcon data
- Fix text for the Debian planet
- Add text on Debian planet

Milestone: https://gitlab.com/AppRecommender/AppRecommender/milestones/14

Reply to: