Final project report part 2: overview

Overview

I am very happy with the outcome of my work at GSoC. On the one hand it is true that the complete list of goals in the original application has not been accomplished. On the other hand, the focus of contributions changed once I started working. I implemented a lot of observation handling code that was not already available as planned, instead of background model application to source observations. Indeed, the majority of the items in the API proposal have been worked out and added as functionality to Gammapy. In addition, I participated in the necessary code cleanup for the release of Gammapy version 0.3.

List of main pull-requests during my GSoC (all merged in the trunk):

• Document observation tables and improve gammapy.obs [#278]
• Observation table subset selection [#295]
• Add cube background model class [#299]
• Make background cube models [#319]

Other relevant pull-requests (also merged):

• Add function to fill acceptance image from curve [#248]
• Consistent random number handling and improve sample_sphere [#283]

(There are more (smaller) pull-request dealing with cleanups, fixes or small additions that are not listed here.)

All in all, a lot of new functionality has been successfully added to Gammapy, as demonstrated by the example in the example section in the final project report part 1 post, and the examples in the documentation links listed below, making this a fruitful project.

As summary of my work produced during the GSoC, I am repeating the list of links to the documentation I produced during the GSoC, that I already posted in the progress section in the final project report part 1.
This documentation explains the most important contributions produced during my GSoC project:

• observation lists format and generation.
• dataset generation.
• observation selection.
• observation grouping.
• cube background models.
• cube background model production.

I would like to take the opportunity to thank the mentors of the project for their useful input. There was at all times at least one of them there to answer my questions and give positive feedback. I learnt a lot during this summer!

For instance I deepened my knowledge of the scientific python stack: numpy, scipy, matplotlib. I learnt of course the basic tools needed for my project: git, GitHub, Astropy, Gammapy. I also learnt how to work collaboratively in a pull-request system with code reviews.

These recently acquired knowledge, together with my previous skills in programming with object-oriented languages granted me access to the list of Gammpy core developers and maintainers with write access to the repository.

To conclude, the completion of this works has taken many hours of hard work (~ 500 h), much satisfaction, some frustration and 0 drops of coffee! :-)