In the first week, I have defined an idea of how the integration of
AppRecommender and AppStream will work. I have contacted the AppStream
team with the idea, that can be seen in the following link:
https://lists.freedesktop.org/archives/appstream/2016-May/000032.html

Also the links for AppRecommender and my AppStream repository can be
seen bellow:
https://gitlab.com/AppRecommender/AppRecommender

https://github.com/lucasmoura/appstream

Furthermore, this week I have also started to package AppRecommender,
but there was a problem with one of the packages dependencies,
python-nltk. AppRecommender has been using the stopwords and stemming
libraries from python-nltk for some recommendation strategies.
However, they are not present when the package is installed, requiring
them to be downloaded separately. In order to solve that, I have
removed python-nltk from AppRecommender and replaced it by
python-stemmer, which has the same stemming algorithm that
AppRecommender was using with python-nltk.

For the next week, My mentors and I have defined the following tasks:

* Create a recommendation class on
AppStream:https://gitlab.com/AppRecommender/AppRecommender/issues/12

* Create parser for recommendation XML
file:https://gitlab.com/AppRecommender/AppRecommender/issues/13

* Continue AppRecommender
package:https://gitlab.com/AppRecommender/AppRecommender/issues/14

It must also be said that depending on the feedback from the AppStream
community on my idea, new issues for this week will arise.

--
Lucas Moura

Reply via email to