Google Summer of Code/2019

From Official Kodi Wiki
Jump to: navigation, search
Home icon grey.png   ▶ Google Summer of Code ▶ 2019
GSOC 2016 logo.png
Welcome to the Kodi Google Summer of Code ideas page.

We encourage interested students of all ethnities and genders to review some of the ideas on this page, and then feel free to provide input on any ideas you may have in the Kodi GSoC 2019 forum and chat about any project you’d love to cover. To propose a project, see #Students project proposal ideas.

From the 25th March to the 9th of April, any interested students may apply at the GSOC home page to work with Kodi. After that, Google will notify applicants whether we get to work with each other according to the GSOC schedule.

1 About Us

As there may be many students who have landed here as a result of GSOC, but are unfamiliar with the project, here are a few resources that may help explain what we are about.

Kodi (formerly known as XBMC) is an award-winning free and open source (GPL) software media player and entertainment hub for digital media, designed around the 10-foot interface (living room) environment. Created in 2003 by a group of like minded programmers, Kodi is a non-profit project run and developed by volunteers located around the world. More than 500 software developers have contributed to Kodi, and 100-plus translators have worked to expand its reach, making it available in more than 60 languages. For more information, see the page Kodi.

To get an idea of what Kodi is truly capable of, it really must be seen. Check out a few other user-created videos:
Kodi with the default Confluence skin
Kodi with Aeon Nox skin
Kodi on Raspberry Pi
Kodi's new PVR functionality

Kodi is written primarily in c++ and runs on a variety of platforms including Android, iOS, Linux, OS X, and Windows. It has been ported to work on several low-power platforms including the Raspberry Pi and Android.

Kodi was a mentoring organization in 2008, 2012, 2013, 2015, 2017, 2018, and had team members involved in GSoC for other projects during 2011.

If Kodi is selected as a mentoring organization for 2019, students will need to review the Overview of a good project proposal, follow the outline for proposals when applying, and review the list of project ideas detailed below. Students are welcome to propose ideas outside the list and are encouraged to be as creative as they like.

1.1 Mentors

All mentors and backup mentors are extremely experienced in the Kodi codebase and will thus be able to assist students in getting to know the codebase and in quickly identifying projects that are both achievable for someone unfamiliar with the internal workings of Kodi and desirable to the wider Kodi community.

2 Prerequisites

C++ coding skills, basic familiarity with Git, solid understanding and interest in programming. Ability to quickly understand existing code is beneficial.

3 Project Proposals

3.1 Overview

Qualifications for a good Summer of Code proposal:

  • Discrete, well-defined, modular
  • Comprised of a series of measurable sub-goals
  • Based on open specs that are available free of charge
  • Based on complete specs

An example of a good proposal is the implementation of a new feature or function that is not yet available in Kodi.

An example of a less desirable proposal is one that's not as measurable, such as refactoring an existing API. Bad proposals tend to be ones that would require touching a lot of core code.

To re-iterate:

  • Localized/isolated code projects = good
  • Global code refactoring = bad
  • A project should have a set of subgoals, so even if the end goal turns out to be too big some of the parts will be of benefit.
  • Not too big! This is an important problem when choosing a project, while it is fun to think about solving a grand project its not always realistic. Its better to finish a smaller project than to start a grand one.

3.2 Where to submit proposals

In addition to submitting to the Google Summer of Code website, you are highly encouraged to submit your idea/proposal to the Kodi forum for discussion. Any proposal not submitted to the forum for discussion will likely not be considered.

3.3 Outline for proposals

4 Team-Kodi proposal ideas up for discussion

4.1 Explore building binary addons in rust

Description: While Kodi has python addons, it can also handle binary addons. But these need to be build one by one for each platform and also might fail on errors. Rust might help with both of those problems and more. The idea is to use the api we have but via rust ffi and see where that leads us.

Expected outcome: Documentation and templates on how to use rust with kodi. In the best case also one or two example addons.

Skills preferred: C++, Rust, C

Possible mentors: spiff, Razze

Difficulty: Hard

Type: Binary addon development

4.2 Better profile support

Description: Profile support inside Kodi remains extremely hacked together. Improving profile projects would likely be broken down into smaller chunks to make each chunk truly achievable in a summer. Database handling for profiles needs to be reviewed and changed so that the user experience is more streamlined. It should also be possible to better guard code paths that need to check for specific profile rights.

Expected outcome: Profiles should be more user-friendly than before and not be an afterthought. Being able to easily add more profile features is the big ultimate goal, as we want to support profiles for kids that automatically hide media that's not for them. Just as an example.

Skills preferred: C++

Possible mentors: spiff

Difficulty: Hard

Type: Core development

4.3 Replacing depends with a CMake-based system

Description: Kodi has its own system for building the many libraries Kodi depends on that supports most platforms (Linux, Android, OS X, iOS) and cross-compiling called depends. At the moment, it is implemented with autotools and hand-crafted Makefiles. This has lead to a lot of code duplication, poor maintainability, and not being very user-friendly. Also, it does not currently work on Windows. This task would be about replacing the current system with a new implementation in CMake that has better maintainability.

Expected outcome: A CMake-based dependency build system that offers roughly the same features as depends (i.e. all required libraries covered, diverse platform support, crosscompilation). If the solution can also be applied to Windows by e.g. adding minimal CMake files to replace some UNIX-only build systems, that would be a big plus, but it is not necessary.

Skills preferred: CMake, shell scripting, general familiarity with UNIX/Linux

Possible mentors: wsnipex

Difficulty: Medium

Type: Infrastructure/Automation

4.4 Integrate TUF (The Update Framework)

Description: Kodi uses a self-built mechanism for installing and updating add-ons from add-on repositories. Unfortunately, it is not very secure. This project would be about replacing the add-on repository code with the usage of TUF (The Update Framework), a quite recent library that solves many common security problems for updaters. TUF does not currently have a C or C++ implementation. The reference implementation is in Python, which we do not want to integrate with C++ for this core piece of application code. Due to security concerns, we also do not want to implement TUF ourselves. That leaves using rust-tuf, an experimental Rust implementation. First step would be to write a C wrapper library so that we can use it from Kodi.

Expected outcome: Add-on repository in Kodi using TUF that builds and works on all supported platforms (beware, there are quite a lot) and is integrated into the Kodi add-on infrastructure (i.e. there are also tools to generate valid repositories). It should support all core add-on functionality (be able to list, install, and update add-ons).

Skills preferred: C++, Rust, Python (for creating the repository the TUF reference implementation could be used)

Possible mentors: yol

Difficulty: Hard

Type: Risky/Exploratory (rust-tuf is unstable and there is also a lot of work to be done on the backend side)

4.5 Add-ons

4.5.1 Automatic add-on checker for binary add-ons

Description: Kodis add-on checker checks each Kodi add-on PR on GitHub for certain problems, which ultimately makes reviewing Python add-ons easier for the team. But the other type of add-ons, binary add-ons is not being checked so far, which makes reviews and coding guidelines hard to handle. Thus it would be nice to also check those add-ons, either via the current checker written in Python or a new binary add-on specific checker. Goals would include checking the addon.xml, the C++ code and a check for potentially malicious code.

Expected outcome: Being able to run the checker against all binary repo PRs on GitHub.

Skills preferred: Python, C++

Possible mentors: Razze

Difficulty: Hard

Type: Low-hanging fruit

4.5.2 Addon submissions

Description: Write a tool or a website to easily submit addons to our addon repo. The tool will need to login at GitHub and send a PR in the name of the author to the correct repo and branch on Kodi's GitHub. The user experience needs to be very good to not keep anyone from submitting. It should also be able to offer some opinions on the code that will be committed before it actually gets sent, so that the most common errors are caught on the client side.

Expected outcome: A tool that lowers the bar for addon authors to get started with repo submissions, so we get more submissions with better quality.

Skills preferred: Depending if this will be a standalone tool or an website the choice of technology would be quite different. It would be a new tool, so that's fine. For a website Python/Django, Elixir/Phoenix or something similar would be a good choice. For a standalone tool we might do a full UI in Java or Electron or settle for a command-line interface via JavaScript.

Possible mentors: Razze

Difficulty: Easy

Type: Low-hanging fruit

4.6 Achievements in Retroplayer

Description: Retroarch has developed support for achievements in certain select libretro cores. This project would consist of porting whatever work might be portable and otherwise integrating the existing libretro achievement system into Kodi's retroplayer implemented using Kodi's UI system. Could be tackled a few different ways, including simply linking the user to the or creating a full achievement database in Kodi.

Expected Outcome: Users should be able to see their achievements in Kodi. Player manager and Player profile support may be in development concurrently, so thought should be given to those systems.

Skills preferred: C++, python, possibly php

Possible mentor: garbear

Difficulty: medium

Type: Fun/Peripheral

Notes: Garbear has broken down the Retroplayer project into manageable pieces including input, core, game add-ons, peripheral add-ons, netplay, a game library implementation, and shaders to do things like emulate a CRT monitor. This is one example, but many pieces are currently being worked on or are near completion, and many more could use help from a student familiar with the area. Visit the retroplayer section of the Kodi forum to see all the major projects still to do.

4.7 Add a picture library

Description: At the moment Kodi has got the concept of libraries for movies, TV shows and music. Libraries are far nicer than directories for users to handle as they keep their data and basically act upon the data that got collected without having to get it on each directory traversal. It also enables filtering, sorting and searches. The project would focus on adding a picture database and getting it filled via file scrapping. When that's done we would progress to write basic library views for pictures.

Expected outcome: A functional picture library for Kodi with the possibility to scan files and having them added to your library. It should also be possible to look at your files via a library frontend, just like it's possible for TV shows, movies or music.

Skills preferred: C++, SQLite, MySQL/MariaDB

Possible mentors: Razze

Difficulty: Easy

Type: Core development

4.8 Use machine learning to improve suggestions

Description: Kodi currently can't suggest similar media to watch/listen to. It would be possible to just show media that's similar on the base of a tag, but that would be quite rudimentary. Better would be to create something driven by the user or profile that's currently logged in. This would mean that watched/listened and collected data be added to the model and we would serve the best possible suggestion for a user. Machine learning might be the best approach as we would be able to create a model and just ship that model and act on it when Kodi tries to serve suggestions.

Expected outcome: Far better suggestions than before. Some places might even just get suggestions due to this project, as it did not make much sense to do suggestions before this.

Skills preferred: Python, TensorFlow, C++, SQLite, MariaDB/MySQL

Possible mentors: Razze

Difficulty: Hard

Type: Risky/Exploratory

4.9 Using fuzz testing to fuzz test Kodi

Description: Adding fuzz testing to Kodi would be very valuable to find problems in the current code base. Especially security problems. First step would be choosing a framework either American fuzzy lop, libFuzzer or other similar projects. After we decided on that it would be time to start to fuzz different methods that are widely used in Kodi and try to expose flaws in those. If this leads to success, it should be documented and made repeatable.

Expected outcome: Being able to easily add more fuzz tests and having the process to do this documented.

Skills preferred: C++, CMake

Possible mentors: Razze

Difficulty: Medium

Type: Risky/Exploratory

4.10 inputstream.adaptive binary addon

Description: inputstream.adaptive is a binary kodi inputstream addon written in C++ and handles multi bitrate streams provided in DASH / Smoothstream manifest or HLS multi bitrate playlists. The addon provides demuxed / encoded packets to kodi and kodi renders the data with its internal videoplayer. Digital Rights Management (DRM) is implemented in inputstream.adaptive to allow playback of protected media. inputstream.adaptive is used by many (> 100) addons to play (legal) encrypted media.
The addon sources are still hosted in my own git repository ( but I'll hand over the sources to kodi repository if someone is interested on working at one of these 2 most urgent topics:

4.10.1 Bitrate switch

Description: In the current inputstream.adaptive implementation a stream resolution / bitrate is selected at start time. Because of network fluctations / manual window / screen changes or performance stats of rendered frames it should be possible to switch seamless to lower / higher stream representations.

Expected outcome: Automagically select / change seamless stream depending on external factors.

Skills preferred: C++, CMake

Possible mentors: peak3d

Difficulty: Medium

Type: Media playback

4.10.2 Read ahead

Description: Currently stream segments are downloaded on request (that is kodi is requesting more data to display). Because the kodi videoplayer internal buffer is small (8 seconds), network fluctuation can lead to bad user experience (stream interruption). What users want is that inputstream.adaptive buffers a configurable amount of time ahead to bypass this issue.
Buffering ahead in multi bitstream is not only a technical task, there are numerous resolutions / bitrates for the same movie time segment in the manifest from which the "best" one should be buffered. The concept of what to read will be an not trivial engeneering task.
Bitrate switch (see previous topic) will be a prerequasite to "Read ahead" because on low network times lower bitrate streams are read compared to good network times.

Expected outcome: Implementation of read ahead logic for multi bitrate streams, user configurable.

Skills preferred: C++, CMake

Possible mentors: peak3d

Difficulty: Medium

Type: Media playback

4.11 Other ideas

These ideas still need expanding and/or assigned a potential mentor. If one of these ideas interests you, feel free to ask for more info in the GSOC 2019 forum area. We may be able to assign a mentor if enough interest is shown

  • Visual Studio Code addon to validate/syntax highlight Kodi Addons
  • Using fruit or similar to use DI in Kodi and write tests
  • High quality scalers for OpenGL(GLSL). Scale Y and UV plane separately
  • Support downloading of of media items (the actual file not just the metadata) from another Kodi instance (e.g. through UPnP) into the local library. This could be done in a "send to" way and in a "download" way and it could also be combined with transcoding depending on the target device.
  • UPnP device profiles. there's already a PR that goes in that direction and the problem is that right now we can only really provide device specific MIME type hacks but if we have transcoding we'll definitely need this
  • Implementing a performance critical element of kodi in rust

4.12 More

We feel it is important to note that, while we are interested in a focus on the listed three areas, we would like to stress passion, expertise, and creativity above all else. If you would like to do something completely different, definitely send in that proposal. The ideas listed above are, as always, merely suggestions. We will be interested in any idea, so long as you can communicate your interest, your background, and your solution the problem.

5 Students project proposal ideas

Submit your own proposals on the Kodi GSoC forum. They can be as big or as small as you feel you can comfortably accomplish between PROJECT START DATE and PROJECT END DATE. In the end, it's better to have a smaller, completed project, rather than a larger, incomplete project.

To submit a proposal idea:

  1. Copy the text from #Outline for proposals.
  2. Create a new forum post HERE and paste the text.
  3. Fill out everything using your specific proposal.